JPM Fixed in

European Rates Strategy
J.P. Morgan Securities Ltd. November 24, 2011
Global Fixed Income Markets 2012 Outlook

Contents Overview Economics Euro Cash European Derivatives United Kingdom US Cross Sector US Treasuries US Interest Rate Derivatives Japan Inflation-linked Markets Australia New Zealand Interest rate forecasts Recent curve movements Recent sov cash spread movements Recent sov CDS spread movements Sov & bank redemptions Euro area sov ratings / SMP purchases / Election calendar Euro area fact sheet 3 31 38 57 77 90 107 131 158 176 204 219 226 227 228 229 230 231 232
Pavan Wadhwa
AC
(44-20) 7777-3370 pavan.wadhwa@jpmorgan.com
Fabio Bassi
(44-20) 7325-8615 fabio.bassi@jpmorgan.com
www.morganmarkets.com
J.P. Morgan Securities Ltd.
See page 233 for analyst certification and important disclosures.
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Pavan WadhwaAC (44-20) 7777-3370 pavan.wadhwa@jpmorgan.com J.P. Morgan Securities Ltd.
Key 2012 Global Fixed Income Markets Trading Themes

Overview Pavan Wadhwa, Kedran Panageas
The negative feedback loop between sovereigns, banks, and the real economy is exacerbating the peripheral debt crisis. Technicals are poor for peripheral sovereign debt; the existing investor base has been impaired and distressed credit funds are keeping their distance due to the highly politicised nature of sovereign debt restructurings. We expect further rounds of forced capital increases for European banks as the Euro zone slips into recession. The ECB is unlikely to step up its pace of buying significantly, absent a catastrophe. Position for lower German yields and a flatter curve. Underweight intra-EMU debt relative to Germany. Australian and Chilean duration is attractive. We highlight trades that are positively convex in peripheral spreads, and others that are uncorrelated to the sovereign debt crisis.
Euro Pavan Wadhwa, Fabio Bassi, Gianluca Salford

We see the refi rate at 0.50% and 10Y Bund yields at 1.25% by mid-2012; go long duration and initiate 2s/10s flatteners with a 100bp target. Underweight peripherals and position for flatter credit curves. EONIA fixings will fall to 30bp as excess liquidity stays elevated; go long 6Mx6M EONIA. Position for 1s/5s outright and conditional bull flatteners, and buy receiver structures on 6Mx5Y swaptions. Swap spreads are likely to trade wider than their Lehman peak; we target 2Y and 10Y swap spreads at 145bp and 90bp, respectively, and favour 2Y wideners. Conditional swap spread wideners offer better risk/reward than outright wideners. In EUR and GBP gamma, buy intermediate-tails and sell short-tails; we target EUR 3Mx10Y at 9.4bp/day. Favour longs in Bund volatility vs. swaption volatility.
UK Francis Diamond
We expect QE gilt purchases to be upsized to a whopping 425bn in 2012. QE, low non-domestic ownership of gilts, and a flexible currency should limit any increase in gilt yields should the UK fall under the sovereign risk spotlight. 2Y gilts should trade in a 50-60bp range. 10Y and 30Y gilt yields are likely to decline: we target 1.50% in 10Y gilts by 2Q12 and expect the 2s/10s gilt curve to flatten to 100bp. Position for wider 5Y and 10Y swap spreads in 1H12.
US Terry Belton, Srini Ramaswamy

Stay long duration in early-2012, targeting 1.70% in 10Y Treasuries. Avoid consensus trades. Synthetic Treasuries created by asset swapping cheap foreign bonds should outperform in 2012. FRA/OIS and intermediate swap spreads will widen initially before narrowing back. Initiate synthetic conditional curve trades by replacing swaptions at the front end with YCSOs. Approach 2012 with a long gamma bias, but look for 3Yx10Y to decline to 6bp/day by 1Q12 end.
Japan Takafumi Yamawaki, Yuya Yamashita

We expect 10Y JGBs to trade in a 0.8-1.1% range in 1H12 and 0.9-1.3% in 2H12, with risks biased towards lower JGB yields. We have a flattening bias on the JPY swap curve. Large negative USD/JPY cross currency basis makes it attractive to buy JGBs and swap them into USD. 3s/6s basis is likely to widen.
Inflation Jorge Garayo, Francis Diamond, Kimberly Harano

We expect significant declines in headline and core inflation across the board in 2012, although inflation expectations will remain anchored. Real yields will test new lows, turning negative in Europe and staying close to 0% in 10Y TIPS. Breakevens will be biased narrower. Expect cash breakevens to underperform relative to inflation swaps in early 2012. Euro area: Position for flatter breakeven curve and higher peripheral linker yields. UK: Be long intermediate real yields, and short inflation breakevens in the 10Y sector. US: Declining inflation and nominal yields, along with fiscal tightening, should cause breakevens to narrow.
Australia / New Zealand Sally Auld

Buy AUD duration on dips as offshore demand is likely to stay elevated, the RBA will ease policy, and domestic investors are short their benchmarks. Position for curve steepening. We are biased towards wider swap spreads. In NZD, we expect yields to make new lows in 2012 and the curve to be biased steeper.
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Pavan WadhwaAC (44-20) 7777-3370 Kedran Panageas (44-20) 7777-0326 J.P. Morgan Securities Ltd
Overview
The negative feedback loop between sovereigns, banks and the real economy is exacerbating the peripheral crisis in the Euro zone Sovereigns Fallen angels: peripheral sovereign debt has too much credit risk to remain in DM portfolios, but is too large to be easily absorbed in EM portfolios while the benchmarking benefit that typically accrues to corporate fallen angels will not accrue to distressed sovereign debt and, given the highly politicized nature of sovereign debt restructurings, distressed credit funds have been slow to invest in this space Non-domestic investors, who own nearly half of all peripheral sovereign debt, have been steadily selling and are likely to have limited appetite to add risk in the near term France is at risk of losing its AAA rating by mid2012, which would cause EFSF capacity to fall by one-third Banks Spreads on senior bank debt and covered bonds are near, if not above, Lehman highs, while Euro zone bank debt issuance has plummeted to below Lehman levels The exposure of core country banks to all peripheral assets tops 1tn, or 90% of bank capital and reserves and the EBAs 104bn estimate of bank capital shortfall for Euro zone banks is 150bn below J.P.Morgan estimates. We expect further rounds of forced capital increases for Euro zone banks Economy J.P.Morgan economists expect Euro area growth to significantly undershoot official forecasts as further fiscal consolidation by Euro area sovereigns pressures GDP growth lower and forced bank deleveraging puts 1-2%-pts of downward pressure on GDP growth
Recommended Trades Position for lower yields and a flatter German curve We recommend underweighting peripheral debt in the first half of 2012; however, we expect peripheral yields to fall in 2H12 in response to a concerted ECB/EU/IMF response to the crisis Global markets that offer high real yields, potential policy rate cuts, and high sensitivity of GDP growth to Europe are attractive; go long 10Y duration in Australia and Chile We highlight trades that are positively convex when peripheral spreads widen/narrow as well as trades that offer diversification from the peripheral debt crisis
The sovereign pandemic

If 2010 was the year that the EU sovereign crisis erupted, 2011 was the year that it morphed into a dangerous new phase. No longer did it limit itself to small nations such as Greece, Ireland, and Portugal. Rather, the contagion spread to Italy and Spain and to a certain extent, core countries such as France. Italy and Spain constitute nearly one-third of European debt markets, with a combined 2tn of marketable debt outstanding. This dangerous new phase began around mid-year, when Greece started to slip on its fiscal targets and the threat of private-sector burden-sharing became real (Exhibit 1). Despite making good progress on fiscal targets and structural reforms in 2010, Greece hit a wall in 2011. By mid-May, it was massively behind target and the IMF needed assurance of future funding in order to continue disbursing promised monies. This led to a second Greek bailout in July that stipulated a bond exchange on private holdings that would extend maturities to 1530 years and lower coupon rates modestly. This broke the taboo that sovereign debt was riskless and reinforced the notion that peripheral debt was now a credit product. This debt restructuring, although never actually consummated, as well as the anticipated need for widescale bank recapitalisation and unhelpful political rhetoric in Italy, were the three main factors that allowed the crisis to jump containment. By August Italian and Spanish 10Y bond yields had breached 6% and the ECB was forced to start buying Italian and Spanish bonds in
3
Exhibit 1: The sovereign debt crisis engulfed Italy and Spain in 2011, morphing into a dangerous new phase. This was spurred by 1) political risk in Greece and Italy, 2) Greek debt restructuring, and 3) concerns around bank capital needs
Timeline of the EU sovereign debt crisis; 10Y weighted* peripheral spreads vs. key events; bp
800 700 600 500 400 300 200 100 0 Jan-10 EBA 1st stress test Apr-10 Jul-10 Oct-10 Discussions on EFSF2 / ESM Jan-11 Portuguese bailout; discussions of Greek PSI Apr-11 Jul-11 Oct-11 EU plans mandatory PSI with ESM aid; Irish bailout Greek backsliding /renegotiation of PSI; referendum and Italian political gridlock EBA 2nd stress test; 2nd Greek bailout
Greek EFSF crisis concerns
Irish concerns Italian crisis and ECB buys IT/ES
* Spreads to Germany of Greece, Ireland, Portugal, Italy, and Spain. Weighted by proportional contribution to the J.P. Morgan EMU Bond Index. Note: For reference, Appendix-1 presents a detailed timeline of key crisis events since 2009.
their Securities Market Programme (SMP). The crisis took another turn for the worse in September and October when Greeces backsliding on reforms deepened, the EU negotiated a new and more severe Greek debt restructuring, and Greek Prime Minister George Papandreou called for a public referendum on the bailout package.1 In response, Angela Merkel and Nicolas Sarkozy raised the prospect of a Greek exit from the Euro zone which reinforced the risk of a hard default and/or a currency switch in investors minds. All five peripheral countries have now been forced to go to early elections or technocratic governments since the onset of the crisis, with Greece and Italy being the most recent victims. This indicates the difficulty in enacting needed reforms in the face of economic slowdown, as well as the high political and implementation risk of fiscal austerity. Going forward, we think the crisis will continue to escalate in 1H12. Our baseline view for 2012 includes the following developments: 1) The market faces a bimodal distribution of outcomes going forward: either the ECB will step
1 For reference, Appendix-1 presents a detailed timeline of key crisis events since 2009.
up its bond-buying pace significantly, or another country may eventually lose access to capital markets. Our basecase view is that the ECB (and Germany) will remain deeply reluctant to greatly increase the size of the its purchases at this time, which will contribute to an escalation of the crisis in 1H12. This is due to the ECBs desire to 1) maintain oversight and aid conditionality, in order to mitigate moral hazard, and 2) keep monetary and fiscal policy distinct. However, we expect a concerted ECB/EU/IMF policy response in 2H12 which will help to ultimately stabilize yields. 2) Greece may continue to fail to meet its fiscal and structural reform targets. We think Greece will eventually need a more severe debt restructuring than the 50% haircut agreed to at the 26 October EU summit, which does not go far enough to restore debt sustainability.2 3) The Euro area will dip into recession, with the biggest contraction seen in the periphery (see Economics). A recession dynamic will make it more difficult for sovereigns to meet fiscal and structural reform targets.
2 Please see Overview, Global Fixed Income Markets Weekly, 28 October 2011.
Exhibit 2: The negative feedback loop between sovereigns, banks, and the real economy is exacerbating the peripheral crisis
Transmission mechanism between the three pillars of the EU sovereign debt crisis
PROBLEM: Overleverage Loss of capital market access Solvency crisis
Sovereign
fis c c al
ow n
s on olid
/ ion t s es s rec l unre ia s oc
wr ited
bai lou
ts
n atio
PROBLEM: Exposure to peripheral assets Overleverage Loss of capital market access credit losses Banks deleveraging Economy
PROBLEM: Slow growth Recession risk Social unrest
4) Faced with recession and disinflation, the ECB will cut the refi rate to 50bp by June 2012, in successive 25bp moves occurring at its December, March, and June meetings. We expect the deposit rate to be reduced to 25bp at the December meeting (see Economics). 5) Portugal and Ireland will need to renegotiate funding packages in mid-to-late 2012, which will likely require private sector involvement (PSI).3 6) France is at risk of losing its AAA rating given contingent commitments to other sovereigns and French banks, especially if economic growth disappoints. This could reduce EFSF capacity by up to one-third (see Euro Cash).
most negative feedback loops, a circuit-breaker is needed in the form of some external factor such as outside money, central bank intervention, or a positive growth shock from abroad. In the sections below, we discuss each of the three legs of the crisis in more detail sovereigns, banks, and the economy with regard to how they will drive outcomes in the periphery in 2012. Second, and critically, the scale of the problem is much larger than anything experienced before. The Russian and Argentine defaults in 1998 and 2001, respectively, only affected between 100125bn of marketable debt each. This was considered to be large at the time, but it pales in comparison to the nearly 2.5tn of marketable debt outstanding today across Italy, Spain, Greece, Ireland, and Portugal.4 Given the scale of the problem, it is difficult to find a credible and/or willing lender of last resort. Third, EU policy response to date has been fractured and behind the curve. The financially sounder core countries and the ECB have been reluctant to support at-risk countries to the extent necessary, for fear of socialising debt and exacerbating moral hazard. This problem will
4 The marketable debt figure includes T-bills, zero-coupon notes, and bonds, but does not include official sector loans, private sector loans, commercial credits, etc. We estimate total debt outstanding of the five peripherals is between 3.03.5tn, based on current GDP and debt/GDP ratios.
Why is this crisis so pernicious?

Exiting this crisis has proven extremely difficult for several reasons. First, we are in a negative feedback loop that is proving difficult to stabilize (Exhibit 2). The banking crisis and recession of 2008/2009 damaged sovereign balance sheets, and todays sovereign crisis (and the policy response) are in turn damaging bank balance sheets and stunting economic growth. As with
3 Please see Overview, Global Fixed Income Markets Weekly, 8 July 2011.
be difficult to overcome until there is more control, more trust, and more homogenisation of fiscal policies across Europe. The fiscal integration process will take years if not decades to achieve, and will require profound EU treaty changes.5 We now discuss each of the three pillars of the EU crisis in more detail. 1. The sovereign debt and funding crisis First, since the onset of the financial crisis in 2008, sovereign balance sheets have been severely damaged: it is, in the words of German Chancellor Angela Merkel, at least a 10-year project to restore fiscal health. Debt-toGDP ratios amongst the peripherals are up 37%-pts on average (Exhibit 3). Even stronger countries now appear stretched. Second, rating agencies have been relentless in downgrading peripheral sovereign debt. The average peripheral rating has fallen from A+/Aa2 two years ago to BBB-/Ba1 today (Exhibit 4). This stands in stark contrast to the rest of the developed market (DM) sovereign universe, the vast majority of which has minimal credit risk and is rated AA or higher (more on this later).6 Third, the crisis has severely damaged market liquidity. This is evident in wide bid/offers, thin trading volumes, and skyrocketing market volatility for peripheral sovereigns (Exhibit 5). Together, the heightened credit risk and market volatility have severely impaired the investor base for peripheral sovereign debt. The natural investor base consists mainly of rates-based investors, including central banks, commercial banks, insurance companies, pension funds, etc. These investors begin to shy away at the first sign of credit risk or investment volatility. Further, these investors also face heightened regulatory scrutiny. This consists of additional disclosure requirements, capital stresses on assets held in banks held-to-maturity books, a possible move to non-zero riskweightings, and the need for deleveraging or asset sales. This treatment makes them less able to withstand the
Exhibit 3: Peripheral sovereign balance sheets have been severely impaired
Debt-to-GDP ratios across selected Euro zone countries, end-2011E* vs. end-2008; % 2011E* 2008 Change
Greece Ireland Portugal Italy Spain Av erage**, peripherals Belgium France Germany Euro area
158 112 102 120 68 112 97 85 82 88
111 44 72 106 40 75 90 68 66 70
47 68 30 14 28 37 7 17 16 18
* Official estimate as provided by Eurostat. ** Simple average. Source: Eurostat
Exhibit 4: and rating agencies have been relentless in downgrading peripheral sovereign debt
Number of notches that peripheral countries have been downgraded by Moodys and S&P over the past two years
S&P Dow ngrade 18-Nov Greece Ireland Portugal Italy Spain Av erage* CC NEG BBB+ BBB- NEG A NEG AA- NEG BBBDec 2009 BBB+ (WN) AA A+ A+ AA+ A+ (notches) 12 5 5 1 2 5 18-Nov Ca DEV Ba1 NEG Ba2 NEG A2 NEG A1 NEG Ba1 Dec 2009 A2 Aa1 Aa2 Aa2 Aaa Aa2 Moody's Dow ngrade (notches) 14 9 9 3 4 8
* Simple average across the peripherals based on a unit ratings scale (AAA=1, AA+ = 2, AA = 3, etc.). Note: NEG, POS, DEV indicates the ratings outlook is negative, positive, or developing, respectively. WN (WP) indicates the rating is on negative (positive) watch. Source: Bloomberg, Moodys, S&P
Exhibit 5: while market liquidity has been severely impaired, leading to wide bid-offers and skyrocketing volatility in peripheral yields
6M standard deviation of daily yield changes in Italian 10Y b/m vs. average monthly bid-offer spread on Italian 10Y bonds* bp/day bp of yield*
16 14 12 10 8 6 6M y ield v ol bid-offer spread 9 8 7 6 5 4 3 2 1 May 09 Nov 09 May 10 Nov 10 May 11 Nov 11
See Overview, Global Fixed Income Markets Weekly, 19 August 2011 for a discussion on the lengthy process required to change EU treaties. 6 In fact, Moodys classifies only three DM issuers (outside of the peripherals) as single-A (Israel, the Czech Republic, and South Korea). These are stronger single-A credits, all of them rated A1 at Moodys and two split-rated at S&P. Specifically, Israel is rated AA- at S&P, the Czech Republic is rated AA at S&P, and South Korea is rated A+ at S&P.
6
4 2 Nov 08
* Estimated from the average bid-offer spread reported in price terms by MTS and price and duration statistics of the J.P.Morgan Global Bond Index for Italy 7-10Y bonds. Source: MTS, J.P.Morgan
MTM volatility that they previously would have been able to weather through fair value accounting in held-tomaturity portfolios. As a result, the traditional peripheral sovereign investor base has grown significantly more risk-averse over the past year, leading to a constant stream of media reports of large institutional investors paring their peripheral debt holdings. This selling has pushed yields significantly higher: peripheral sovereign debt now has too much credit risk to remain in DM portfolios, but is too large to be easily absorbed in EM portfolios. First, outstanding debt from EU peripherals comprises over half of outstanding EM sovereign debt, making it a huge challenge to absorb (Exhibit 6). Second, downgraded sovereign debt will most likely not be added to EM sovereign indices (which have fairly low country wealth and income thresholds) or to high-yield or EM corporate credit indices. Thus, the benchmarking benefit that typically accrues to corporate fallen angels when they migrate from investment-grade to high-yield territory is unlikely to accrue to distressed sovereign debt. For instance, we estimate that the yield on the typical auto industry fallen angel fell an average of 150bp after crossing into high-yield territory, due to demand from benchmarked investors in high-yield space (Exhibit 7). Third, even though credit-focused hedge funds or distressed investors may theoretically be interested, the highly politicised nature of sovereign restructurings makes them more difficult to analyse as compared to the better-defined corporate restructuring protocol.7 For instance, GM is routinely pointed to as an example of a politicised bankruptcy in corporate space, where the governments involvement led to the typical recovery priority being discarded, to the detriment of bond investors and to the benefit of unionized employees. Although several investors explored legal options, they ultimately settled for the government-negotiated outcome as it was too difficult/expensive to fight in court. Such treatment is likely to be amplified many times over in the sovereign debt space, where there is no defined priority of payments in a debt restructuring. This is ultimately to the detriment of the EU as it is likely to dissuade new bondholder investment in peripheral sovereigns. For instance, the market reacted quite negatively when Merkel and Sarkozy first announced in late 2010 that any official loans granted out of the ESM after 2013 would
7 See Overview, Global Fixed Income Markets Weekly, 10 December 2010, for sovereign debt legal issues.
Exhibit 6: Peripheral sovereign debt has too much credit risk to remain in DM portfolios but is too large to be easily absorbed in EM portfolios
Outstanding developed and emerging markets debt universe* by Moodys rating**; % of total Rating DM ex peripherals Peripheral EM Total
Aaa Aa A Baa Ba B <= Caa Total
48% 26%^ 1%^^ 75%
7% 1% 1% 9%
4% 1% 6% 3% 2% 16%
48% 30% 9% 6% 4% 2% 1% 100%
* Sample consists of 105 countries of which 27 (78) are developed (emerging) debt markets. Total gross debt outstanding is $55tn of which $46tn ($9tn) is from developed (emerging) markets. Debt includes all outstanding bonds, loans, and other obligations. ** Local currency long-term debt rating. Note: the US is rated AAA by Moodys. ^ Consists of three countries: Belgium, Hong Kong, and Japan. ^^ Consists of three countries: Israel, South Korea, and the Czech Republic. Source: Moodys, IMF
Exhibit 7: and the benchmarking benefit that typically accrues to corporate fallen angels will not accrue to distressed sovereign debt. Moreover, given the highly politicised nature of sovereign restructurings, distressed credit funds have been slow to invest in this space
Avg. yield-to-maturity of 10Y GM, Ally Financial, Ford, and Ford Motor Credit* bonds around the 5 May 2005 downgrade to HY**; %
10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 -3 -2 -1 0 1 2 # months around dow ngrade to HY** 3
* Specific bonds used are: GM 7.125% Jul13, Ally Financial 6.5% Feb16, Ford 7.45% Jul31, and Ford Motor Credit 7% Oct13. We use the Jul31 bond for Ford Motor Co. as it did not have a large/liquid 10Y note outstanding at the time. ** S&P downgraded Ford, GM, and their subsidiaries on the same date, 5 May 2005.
have preferred creditor status over private debt investors8. Partly due to these concerns, distressed credit funds have been slow to invest in the peripheral sovereign space and will likely remain wary.
8 For instance, see Cross-country variations in capital structures: the role of bankruptcy codes, Viral Acharya et. al., Journal of Financial Intermediation, 2011, and Creditor rights and Corporate risk-taking, Viral Acharya et. al., Journal of Financial Economics, forthcoming.
Exhibit 8: Additionally, non-domestic investors own nearly half of all peripheral sovereign debt. These investors have been steadily selling and may be reluctant to return to this asset class in the near future
% of marketable debt outstanding held by non-domestic investors; % % non-domestic
Exhibit 9: Given these poor demand technicals, sovereign spreads continue to trade close to their lifetime wides, despite ECB buying
10Y spreads to Germany; 18 November 2011 vs. 2Y maximum; bp 18-Nov-11 Max 18-Nov / Max
Greece Ireland Portugal Italy Spain Belgium France Av erage*

* Simple average.
2425 625 910 494 469 282 147 765
2471 1142 1161 575 486 309 188 905
98% 55% 78% 86% 97% 91% 79% 85%
Greece Ireland Portugal Italy Spain Wtd av g., peripherals* France Germany
67% 83% 80% 45% 38% 49% 66% 81%
* Weighted by marketable debt outstanding. Source: National central banks. Data as of latest available update for each sovereign; generally 2Q11/3Q11.
Exhibit 10: Even EFSF paper has widened to 120bp over Libor, indicating an aversion to credit risk, ratings downgrade risk, and structural complexity
10Y ASW spreads for EFSF* and France; bp
Additionally, non-domestic investors are likely to have limited appetite to add risk in the near term. Non-domestic investors own nearly half of all peripheral sovereign debt; their share ranges from nearly 40% in Spain to over 80% in Ireland (Exhibit 8). These investors are typically amongst the first to sell and, once having exited the asset class, will likely be reluctant to return in the near future. We expect them to provide a steady drip of selling pressure going forward. Given these poor demand technicals, sovereign market access remains tenuous and there is no firm ceiling for spreads. Indeed, European sovereign spreads are currently close to their lifetime wides, despite the steady stream of bailouts, crisis measures, and backstop ECB buying (Exhibit 9). Even EFSF paper has widened to 120bp over Libor, indicating an aversion to credit risk, ratings downgrade risk, and structural complexity (Exhibit 10).
140 120 100 80 60 40 20 0 -20 Aug

* EFSF 3 3/8 July 2021
France
EFSF
Sep
Oct
Nov
Exposure of core-country European banks to peripheral country assets as % of bank capital and reserves; 2Q11 vs. 4Q09; bn Origin country Total exposure
Exhibit 11: Although core-country banks have been cutting their exposure to peripherals, their total exposure still tops 1tn, or almost 90% of bank capital and reserves
Claims on Greece Ireland Italy Portugal Spain Total Capital and reserves (C&R) Exposure / C&R
Austria 2 2 17 1 6 28 92 30%
Belgium 1 18 17 2 16 54 58 93%
France Germany 38 22 287 18 104 469 492 95% 15 76 111 25 122 349 391 89%
Netherlands 3 12 36 5 53 109 102 107%
2Q11 4Q09 % chg 60 129 468 50 302 1009 1135 89% 101 214 576 81 420 1392 1081 129% -41% -40% -19% -37% -28% -28% 5% --
Source: BIS, ECB. Data as-of 2Q11.

8
Exposure of core-country European banks to peripheral public sector debt (including sovereign debt) as of 2Q11; bn Origin country Total exposure
Exhibit 12: German and French banks alone hold nearly 180bn of peripheral public sector debt, which is now imperilled given the PSI precedent set in Greece
Exhibit 13: The recession and its aftermath have led to sharply higher NPLs in the periphery, even on non-sovereign lending
Non-performing loans as a % of total gross domestic loans, quarterly data for Greece* and Spain**; %
12 10 8 6 4 2 0 2006 2007 2008 2009 2010 2011 Spain Greece
Claims on Greece Ireland Italy Portugal Spain Total
Belgium France Germany 2Q11 4Q10 1 0 11 1 3 17 7 2 74 4 21 108 9 2 33 6 20 70 19 5 125 17 48 215 32 6 131 20 50 240
% chg -41% -18% -4% -16% -4% -11%
Source: BIS. Data as-of 2Q11.
2. Bank funding and equity capital crisis Sovereign fiscal issues and slowing economic growth have direct repercussions on banking sectors. Due to cross-border capital flows, losses are transmitted quickly from borrower to lender countries. Although corecountry banks have been cutting their exposure to peripherals, their total peripheral exposure still tops 1tn, or almost 90% of bank capital and reserves (Exhibit 11). A sizable portion of this total exposure is peripheral sovereign debt; German and French banks alone hold nearly 180bn, which is now imperilled given the PSI precedent set by Greece (Exhibit 12). In addition, anticipated recessions in the peripherals will increase the losses on other assets such as consumer and corporate loans. Non-performing loan ratios (NPLs) are up sharply across the periphery since the recession of 2009 and will likely continue rising as the periphery slides back into recession (discussed further below; Exhibit 13). Not surprisingly, these exposures have pummeled bank stock prices and impaired funding access. Bank equity prices are down 3540% on the year and spreads on senior bank debt and covered bonds are at multi-year highs (Exhibit 14). Additionally, debt issuance has ground to a halt; rolling quarterly long-term debt issuance is at a 5-year low, lower than even Lehman levels (Exhibit 15). Net, we estimate that core-country banks have between 0-2 quarters of funding coverage at this time.9 ECB reliance has steadily increased and the ECBs balance sheet likely will have to expand massively if it is to contain the incipient bank funding crisis (discussed further below). For instance, the Wall Street Journal recently reported on the growing volume
* For 2006-2007, quarterly data interpolated from year-end data. ** Total of loans to household and corporate sector. Source: Bank of Greece, Bank of Spain
Exhibit 14: As a result of large exposures to peripheral assets, spreads on senior bank debt and covered bonds are near, if not above, Lehman highs
Senior bank debt and covered bond ASW spreads; current as of 18 Nov 2011; bp
250 200
Current : 216bp Current : 198bp
150 100 50 0 -50 2007
Senior
Cov ered bonds
2008
2009
2010
2011
Source: J.P. Morgan Maggie Credit Index
Exhibit 15: and Euro zone bank debt issuance has plummeted to below Lehman levels
Rolling 3M sum of gross bank debt and covered bond issuance for Euro zone banks; all currencies/all domiciles; bn/quarter
300 250 200 150 100 50 Total = 65bn (53+12bn) 2007

Source: Dealogic
Av g Total = 130bn (90+40bn)
Bank debt + CB Bank debt only
See Overview, Global Fixed Income Markets Weekly, 30 September 2011.
Total = 50bn (33+17bn) 2010 2011

9
0 2008
2009
of collateral swaps designed to boost banks access to ECB facilities.10 On the equity front, European banks are overleveraged, especially given growing recession risk and marketimplied losses on sovereign debt. The recently announced capital shortfall of 104bn11 for Euro zone banks is roughly 150bn below J.P.Morgan estimates (Exhibit 16).12 To be sure, the 9% core Tier 1 hurdle that the EBA used is aggressive, but in our opinion the rapid deterioration in economic fundamentals is a more significant driver. We therefore believe that further capital increases (above and beyond the 104bn estimate) will likely be required of banks. Capital needs are likely to be higher still if other peripheral countries such as Ireland and Portugal eventually undergo debt restructuring, which is our basecase expectation.13 Bank stress feeds back into the sovereign crisis by increasing sovereign contingent liabilities. Particularly at risk are Spain, Belgium, and France. Spain has large unrecognised capital shortfalls at its cajas14, while the prospect of bank guarantee schemes and capital injections have already been cited as risk factors for Frances AAA rating. 3. Spillover to the real economy Sovereign fiscal austerity and bank deleveraging are spilling over to the real economy, stunting economic growth. Our economists expect Euro area real GDP to grow at a rate of -0.6% in 2012, with the contraction concentrated in the periphery (we forecast 2012 GDPweighted growth of -1.8% in the periphery) (Exhibit 17). This is driven in large part by the large and sustained fiscal consolidation that must be done. Based on prior episodes of fiscal consolidation, we estimate that each 1%-pt of fiscal tightening in the primary deficit lowers
Exhibit 16: The EBAs 104bn estimate* of bank capital shortfall for Euro zone banks is 150bn below J.P. Morgan estimates. This is unlikely to be the last round of forced capital increases for European banks, especially if economic growth disappoints and sovereigns such as Portugal and Ireland eventually undergo debt restructuring
Capital shortfalls announced by the European Banking Authority (EBA) for various Euro zone countries banking sectors vs. J.P. Morgan estimates**; bn J.P.Morgan
EBA target Country Total France Germany Spain Italy Greece Portugal Others capital 104 9 5 26 15 30 8 11
estimated capital shortfall 254 48 43 53 35 40 17 18
J.P.Morgan EBA estimate 151 39 38 27 20 10 9 7
* Note that the 104bn shortfall estimate is for Euro zone banks only. The total estimated shortfall of bank capital across the entire Euro area is 106bn. ** See The Great Bank Deleveraging: Banking Sector Outlook 2012, 4 November 2011. Assumptions underlying J.P. Morgan estimates are 1) 9% Tier 1 capital threshold; 2) original Basel 2.5 Tier 1 capital ratios as provided by the EBA; 3) sovereign MTM on entire holdings of sovereign debt as of end-October; 4) macro stresses as provided by the EBA in June 2011; and 5) cajas included in Spain. Source: EBA, J.P. Morgan
Exhibit 17: J.P.Morgan economists expect Euro area growth in 2012 to significantly undershoot official forecasts
2012 GDP growth forecasts; % YoY JPM Official Greece -6.6 0.8 Ireland 0.0 1.6
Portugal Spain Italy Wtd avg., peripherals* Belgium France Germany Euro area* -3.9 -1.1 -1.5 -1.8 -0.4 -0.1 0.3 -0.6 -2.8 2.3 0.6 1.1 2.3 1.0 1.0 1.1 EC -2.8 1.1 -3.0 0.7 0.1 0.0 0.9 0.6 0.8 0.5 JPM - Official JPM - EC -7.4 -1.6 -1.1 -3.4 -2.1 -2.8 -2.7 -1.1 -0.7 -1.7 -3.8 -1.1 -0.9 -1.8 -1.6 -1.8 -1.3 -0.7 -0.5 -1.1
* Weighted by 2011 GDP. Source: National treasury ministries, European Commission and J.P.Morgan forecasts See Banks face funding stress, 17 November 2011, Wall Street Journal. In these collateral swaps, a bank pays a spread to swap collateral which is not eligible for ECB repo in order to receive higherquality collateral which is eligible. This enables the bank to boost ECB access and is a sign of the increasing scarcity/cost of market funding. 11 Note that the 104bn shortfall estimate is for Euro zone banks only. The total estimated shortfall of bank capital across the entire Euro area is 106bn. 12 We estimate that the largest discrepancy stems from removing the original macroeconomic stresses applied in July; a modest portion (roughly 20bn) stems from leaving out the cajas in Spain 13 Please see Overview, Global Fixed Income Markets Weekly, 8 July 2011. 14 See EBA Stress Test Results Analysis: Must do much better, Roberto Henriques, 8 July 2011.
10
10
GDP growth by 0.7%-pts (Exhibit 18).15 Additionally, we estimate that each 1%-pt of bank deleveraging depresses economic growth by 0.60.8%-pts (Exhibit 19). Extrapolating to the Euro area suggests 12%-pts of downward pressure on GDP growth from bank deleveraging. Thus, in our opinion, economic risk is skewed to the downside.
15 See Growth in the Euro area periphery to underwhelm, 16 September 2011, Nicola Mai.
We note that J.P.Morgan economic projections are significantly more bearish than official forecasts, either by national treasury ministries or by the European Commission (EC).16 For example, the J.P.Morgan GDP forecast for the Euro area is over 1%-pt more bearish than the EC forecast, and the forecast for the periphery is nearly 2%-pts more bearish (Exhibit 17). Slower growth feeds back into the sovereign crisis by increasing the fiscal deficit, forcing the sovereign into a vicious cycle of more cutbacks which, in turn, negatively impact growth and damage political and popular support. For instance, the realised GDP contraction in Greece has been several %-pts worse than forecast by the EU/IMF, coming in close to -5.5% for 2011 vs. an estimated -3% originally. Similarly, unemployment has surged, leading to widescale social protests and political backsliding (Exhibit 20). Across the periphery, all five ruling coalitions have changed since the onset of the crisis. However, any benefits from a technocratic government will likely accrue only over the long term. In the short term, the effect can be more market-negative if it increases uncertainty or makes it more difficult to implement reforms. Net, the result of continued GDP contraction in 2012 is likely to be further slippage on fiscal targets, higher deficits, increases in debt/GDP trajectories, and heightened risk of PSI in other peripherals besides Greece (despite official declarations that Greece is a one-off).
Exhibit 18: as further fiscal consolidation by Euro area sovereigns puts significant downward pressure on GDP growth
Fiscal consolidation targets and their impact on GDP growth*; % of GDP
Fiscal consolidation target 2011 Greece Ireland Portugal Spain Italy Wtd avg., peripherals* Belgium France Germany Euro area 6.0 2.2 5.7 2.7 0.8 2.1 0.2 1.4 0.0 1.2 2012 5.1 1.6 6.1 2.6 3.5 3.4 1.2 1.6 1.0 2.1 2013 0.8 2.8 2.2 2.5 2.3 2.3 1.4 2.0 0.0 1.4 2011 -4.2 -1.5 -4.0 -1.9 -0.6 -1.5 -0.1 -1.0 0.0 -0.8 2012 -3.6 -1.1 -4.3 -1.8 -2.5 -2.3 -0.8 -1.1 -0.7 -1.5 Impact on GDP growth* 2013 -0.6 -2.0 -1.5 -1.8 -1.6 -1.6 -1.0 -1.4 0.0 -1.0
* Fiscal tightening defined as the change in the cyclically-adjusted primary balance. We estimate the follow-on change in GDP growth to be -0.7 times the annual fiscal tightening. Please see Growth in the Euro area periphery to underwhelm, 16 September 2011, Nicola Mai. Source: National treasury ministries, EC and J.P.Morgan estimates.
Exhibit 19: Further, the US experience suggests that annual GDP growth declines by 0.7%-pts for each 1%-pt increase in bank capital ratio; extrapolating to the Euro area suggests 12%-pts of downward pressure on GDP growth from bank deleveraging
Residual of US vs. global real GDP growth (YoY) regressed against US banks equity capital ratio*; January 1999March 2011; quarterly data; %
4 3 2 1 0 -1 -2 -3 8.0% 8.5% 9.0% 9.5% 10.0% 10.5% Bank equity capital ratio*
y = -70.0x + 6.7 R 2 = 31%
Analysing the risk posed by peripherals

The catalogue of risks facing the Euro zone is daunting, especially where Greece and Italy are concerned. As discussed previously, Italys size and cross-border linkages make it systemically important. If Italy were to lose market access, it would greatly increase the contingent liabilities and funding costs of other sovereigns, and would most likely result in Spain losing market access as well. We thus view Italy as the most challenging issue for 2012. Why is Italy at risk? First, Italys debt burden is large. For instance, we estimate that Italian debt/GDP will peak near 125% in 2013, and slowly decline towards 114% by 2020 (Exhibit 21). This is based on our economists projections of economic growth and takes into account recently announced austerity measures (and anticipated countermeasures when growth disappoints). These projections envision, for instance, that Italy will run a
16
11.0%
11.5%
* Bank equity capital ratio defined as the bank equity capital of FDIC-insured commercial banks and savings institutions divided by total assets. Tier-1 capital ratio has roughly a 1:1 beta to this definition of equity capital ratio over the last 20 years. Source: FDIC, J.P. Morgan
Exhibit 20: The unemployment rate has climbed over 2.5%-pts on average since the start of the peripheral debt crisis, leading to social unrest, political backsliding, and slippage on structural reforms
Unemployment rate; current vs. end-2009; % Current Dec-09
Chg. 7.4 1.3 1.2 0.0 3.5 2.7 0.1

11
Greece Ireland Portugal Italy Spain Av g., peripherals* Euro area

* Simple average. Source: Eurostat
17.6 14.2 12.5 8.3 22.6 15.0 10.2
10.2 12.9 11.3 8.3 19.1 12.4 10.1
The latter is typically more up-to-date for latest economic trends.
Exhibit 21: Fiscal sustainability in Italy and Spain is uncertain

2011 Debt/GDP (% of GDP)* Italy Nominal grow th rate (% YoY) Primary surplus (% of GDP) Wtd av g. coupon rate*** (%) Debt/GDP (% of GDP)*,** Spain Nominal grow th rate (% YoY) Primary surplus (% of GDP) Wtd av g. coupon rate*** (%) 121 2.0 0.5 4.1 70 1.9 -4.9 3.7 2012 124 -0.1 3.2 4.6 75 -0.1 -3.2 4.5 2013 125 -0.3 4.8 4.9 81 0.6 -1.5 4.9
J.P.Morgans basecase forecast of debt/GDP trajectory* and key macro assumptions underlying the trajectory for Italy and Spain**; blue selection indicates peak debt/GDP
2014 125 1.2 5.0 5.1 85 1.6 -0.6 5.2
2015 123 2.7 5.0 5.3 86 2.2 0.8 5.4
2016 121 3.0 5.0 5.4 87 2.4 2.0 5.6
...2020 114 2.8 5.0 5.9 84 3.4 3.0 6.0
...2030 100 2.8 5.0 6.3 77 3.4 3.0 6.4
* Note that J.P.Morgan estimate for 2011 debt/GDP is slightly more bearish than the latest available official forecast (shown in Exhibit 3) which is based on Eurostat data. ** Spanish debt/GDP trajectory does not include expected bank recapitalisation needs of roughly 50bn, covering both banks and cajas (J.P. Morgan estimate). 2011 Spanish GDP is estimated near 1.08tn, so this equates to roughly 5% of current GDP. Rating agency estimates of Spanish bank recap needs run from around 20bn to as high as 120bn in stressed cases. *** Based upon current weighted average coupon rate and average marginal borrowing rate of 6.50% on new issuance.
sustained primary surplus equivalent to 5% of GDP, and that real growth will climb to 0.8% per annum, with annual inflation at 2%. Second, given that the Italys debt burden is second only to Greeces, any slippage on fiscal consolidation or debt costs could quickly turn the liquidity crisis into a solvency crisis. To illustrate this, we calculate the sensitivity of debt/GDP ratios projected in 2020 and 2030 to various economic assumptions. These sensitivities are shown in Exhibit 22. We find that: A -1%-pt shock to estimated nominal GDP growth rates (vs. J.P. Morgan base case trajectory shown in Exhibit 21) will boost debt/GDP in 2020 (2030) by 13% (33%)17 A -1%-pt shock to the primary surplus each year (again vs. J.P. Morgan basecase trajectory) will boost debt/GDP in 2020 (2030) by 10% (26%) A 1%-pt higher borrowing rate on new Italian debt (vs. J.P. Morgan assumption of 6.5%) will boost debt/GDP in 2020 (2030) by 7% (22%)
Change in projected 2020 and 2030 debt/GDP ratio for a -1%-pt shock to nominal growth and primary surplus assumptions*, and for a +1%-pt shock to borrowing rate assumption*; ratios Variable Shocked
Exhibit 22: and highly sensitive to adverse shocks in GDP growth, primary surplus and funding costs
Nominal Year Italy Spain 2020 2030 2020 2030 grow th rate 13 33 9 22
Primary surplus 10 26 10 25
Marginal borrow ing rate on debt 7 22 5 16
* Measured around the J.P.Morgan baseline forecast, holding other variables constant. J.P. Morgan base case assumptions shown in Exhibit 21.
rapidly spiral out of control. To be sure, a mild outperformance will have the opposite impact, with debt/GDP ratios falling more than is reflected in our baseline scenario.18 How likely is Italy to deviate to the downside from our basecase scenario? Unfortunately, our assumptions may already by generous. On the economic front, risk is skewed to the downside. In the short term, this is driven
18 However, we note that the sensitivity of the debt/GDP trajectory to interest rate and growth rate shocks is non-linear due the compounding of financing costs over time. Therefore, the betas are somewhat larger (in absolute terms) when the situation is deteriorating, and somewhat smaller when the situation is improving. For instance, for a 2020 horizon, the beta for a positive shock (+1%pt to growth, or -1%-pt to the borrowing rate) is roughly 5-10% smaller than the beta for a negative shock (in absolute value terms). For the 2030 horizon, the beta for a positive shock is roughly 15-20% smaller than the beta for a negative shock. On the other hand, the beta to primary surplus assumption is linear since it is not multiplied by the outstanding debt stock each year but rather is added to it.
Third, while any one of these shocks on its own would be negative and leave the debt burden at a very high level, the shocks are likely to be correlated. That is, if growth surprises to the downside, it will be more difficult to run the targeted primary surplus, and borrowing costs will likely rise. Thus, even mild underperformance can
17
Note that a shock to real growth and a shock to inflation have the same impact on nominal GDP growth. Thus, they would have the same impact on debt sustainability.
12
by the possibility of European policy dysfunction, and in the long term, by Italys deep-seated structural challenges. For instance, we assume a long-term 0.8% per annum growth rate even though Italy has only managed a 0.3% annualised growth rate over the past decade. On the financing front, Italian yields have spiked past 7% in recent weeks, so that a 6.5% borrowing cost may also be ambitious. Fiscal pressures on Spain are also significant. Spain has a lower debt burden than Italy does, but is running a higher deficit. Thus, we expect Spanish debt/GDP to peak later, somewhat shy of 90% of GDP in 2016 (Exhibit 21). This assumes that the incoming administration in Spain will pursue additional fiscal tightening relative to what has been budgeted so far, eventually achieving a primary surplus of 3% per annum. However, this trajectory does not include expected bank recapitalisation costs that could add between 2-12% debt/GDP, depending on the estimate.19 The sensitivity of Spains debt/GDP trajectory to shocks in macro assumptions is similar to Italys, but somewhat smaller given a lower starting debt burden (Exhibit 22). Thus, there is significant slippage potential, especially for countries whose debt levels are already high. It may be challenging for Italy to achieve the necessary fiscal consolidation amidst a recession and a high level of social resistance. The recently appointed technocratic government may prove more effective than the previous administration, but the impact of any structural changes that it implements is not likely to be felt on the economy for a long time.20 Given these pressure points, Italy, and possibly Spain, risk losing market access in 2012.
Exhibit 23: Not counting bills, Italy and Spain have gross funding needs over the next three years of just under 900bn
Yearly Italian and Spanish funding needs; includes bond redemptions and general government budget deficits*; bn 2012 2013 2014 Total
Italy Spain Italy + Spain
233 113 347
171 111 282
134 103 237
539 327 866
* We assume T-bills continue to be rolled. Bond redemptions include both conventional and non-conventional bond redemptions. Budget deficits based on J.P. Morgan base case expectations.
Exhibit 24: while the maximum funding available from various sources is just over 700bn, or about 150bn short of their 3Y funding needs
Maximum funding available to Italy and Spain from private and official sources; bn Private sources Official sources
One-off w ealth tax es, etc.* Italy Spain Italy + Spain 135 -135
SWFs** IMF*** 40 20 60 210 110 320
EFSF^ 140 70 210
Total 525 200 725
* Includes one-off wealth taxes, privatisation, and sale of gold and FX reserves. Please see Appendix-2 for details. ** We estimate sovereign wealth funds would be willing to contribute no more than 60bn to purchases of peripheral sovereign debt. Please see Appendix-3 for details. *** Please see IMF 101, Global Fixed Income Markets Weekly, 21 October 2011 for discussion of IMF lending capacity. ^ Please see Overview, Global Fixed Income Markets Weekly, 14 October 2011 and 28 October 2011 for discussion of EFSF spare capacity and leverage prospects. We assume that 123bn of the EFSFs current unallocated capacity is allotted to Greece for its revised bailout package and that 61bn will be used for a second bailout of Ireland and Portugal in 2H12.
note that funding needs are front-loaded (350bn in 2012). Unfortunately, the official resources available to deal with this outcome are limited. Euro area sovereigns are already stretched financially, and non-Euro area sovereigns see little reason to support fundamentally wealthy countries. We estimate that in extremis, there is a maximum of roughly 725bn available to support Italy (525bn) and Spain (200bn), as shown in Exhibit 24. This arises from the following sources: 1) We believe that Italy can raise at most 125150bn from one-off measures such as wealth taxes, privatisations, and sales of gold and FX reserves over the next 12 months.21 Please see Appendix-2 for our calculations. We assume that anything Spain
21 Because we give Italy the benefit of the doubt by achieving a high rate of privatisation in the first year, we assume potential privatisation revenues in the subsequent 1-2 years are negligible.
Possible policy response if another peripheral country loses market access

Given their size, it would be challenging to fund Italy and Spain in the event that they lose market access. We estimate total funding needs for Italy and Spain of nearly 900bn over the next three years (Exhibit 23). This includes bond redemptions (conventionals and nonconventionals) and anticipated deficits, but assumes that both continue to roll T-bills in the private market. Also,
J.P. Morgan estimates roughly 50bn of Spanish bank recapitalisation needs, covering both banks and cajas. Rating agency estimates run from around 20bn to as high as 120bn in stressed cases. 20 See Euro Cash, Global Fixed Income Markets Weekly, 11 November 2011, for an analysis of Italys last experience with a technocratic government.
19
13
manages to raise from one-off measures will be small and likely absorbed by further bank recapitalisation costs.22 2) We believe that sovereign wealth funds (SWFs) will be willing to contribute no more than 60bn, or roughly one-third of the amount currently allocated to European fixed income in their portfolios. Please see Appendix-3 for our calculations. Of this, we split the contribution roughly 2:1 between Italy and Spain. 3) We believe that the IMF can allocate no more than 320bn to Italy and Spain combined. Even that would require upsizing IMF capacity through exercising New Arrangements to Borrow (NABs) and allocating a huge percentage of available capacity to the two countries.23 It would also likely be unpopular politically and might require quid pro quos such as a reshuffling of voting rights. Of this, we split the contribution roughly 2:1 between Italy and Spain. 4) Finally, we do not think that the EFSF will be able to meet its leverage targets, for reasons discussed previously.24 After allocating funds for Greeces second bailout package and assuming another 60bn is required to support Ireland and Portugal in 2H12,25 only 210bn of spare EFSF capacity would be left. This assumes France manages to keep its AAA rating (France provides roughly one-third of EFSF AAA guarantees and hence one-third of EFSF issuance capacity). We also think it is unlikely that the European Stability Mechanism (ESM), the 500bn planned successor to the EFSF, will be brought forward early and allowed to run concurrently, as has been mooted in the media. This would require greatly increasing the German bailout contribution, which goes against Chancellor Merkels promises to the German parliament and would likely be challenged by the German constitutional court.26
Exhibit 25: Selling pressure on Italian and Spanish bonds could be as large as 650-700bn
Estimated secondary market selling pressure in Italian and Spanish bonds that mature in 2013 or after; bn Debt stock that Amount of debt
matures post-2012* Fraction Investor type Domestic Italy Non-domestic All Domestic Spain Non-domestic All Domestic Italy + Spain Non-domestic All 55% 45% -62% 38% -57% 43% -held (%) held (bn) 675 552 1227 268 165 433 943 717 1660 (%) 20% 66% -20% 66% -20% 66% --
sold (bn) 135 364 499 54 109 162 189 473 662
Par v alue Fraction Par v alue
* We include all conventionals, international bonds, linkers, and zero-coupon and floating rate notes that mature in 2013 or later. We assume that 1) 2/3rd of all nondomestic investors and 1/5th of all domestic investors wish to sell their holdings of Italian/Spanish paper and 2) T-bill markets remain open to Italy and Spain. Domestic and non-domestic ownership shares taken from Exhibit 8 above.
Under these assumptions, total available funding of 725bn is sufficient to fund Italy and Spain for a little over two years. This may be enough to buy some time, but is unlikely to keep yields in check. In the long run, we believe that the ECB will eventually need to step in to stabilize markets by monetizing EMU debt. The ECBs role How much would the ECB need to buy to stabilise markets? Most observers, including ourselves, believe that the crisis has progressed to such an extent that the ECB is now the only institution remaining with enough firepower to short-circuit the contagion and stabilise markets. However, it is reluctant to do so as it would mean socialising a large amount of debt, exacerbating moral hazard, and potentially damaging its credibility by blurring the line between fiscal and monetary policy. Additionally, debt monetisation is not a long-term solution: there is no guarantee that debt capital markets will reopen for peripheral sovereigns once the ECB stops buying. In this section, we assess just how active the ECB would need to be in order to deliver market stability, and whether they would be able to intervene to the extent required. We focus on the immediate pressures facing the market, specifically in the next one year. As mentioned in the previous section, Italy and Spain have combined
22 Also, note that Spain owns considerably less gold and FX reserves than Italy does. 23 Please see IMF 101, Global Fixed Income Markets Weekly, 21 October 2011. 24 See Overview, Global Fixed Income Markets Weekly, 14 October 2011 and 28 October 2011. 25 See Overview, Global Fixed Income Markets Weekly, 8 July 2011. 26 See Overview, Global Fixed Income Markets Weekly, 9 September 2011.
14
Exhibit 26: suggesting that the supply/demand imbalance in Italian/Spanish paper over the next 12 months would be around 75bn/month, or an imbalance of around 40bn/month assuming that the SMP continues its current rate of purchase of 35bn/month
Estimated supply/demand imbalance* in Italian and Spanish bonds over the next one year, with and without IMF/EU/SWF commitments**; bn Italy Spain Italy + Spain
Without Gross issuance / funding needs Selling pressure on bonds* Domestic inv estor selling Non-domestic inv estor selling One-off w ealth tax es, asset sales Funds available** SWFs IMF/EFSF Supply/demand imbalance* bn total bn/month 233 135 364 135 --598 50
With 233 135 364 135 40 233** 324 27
Without 113 54 109 ---276 23
With 113 54 109 -20 113** 142 12
Without 347 189 473 135 --873 73
With 347 189 473 135 60 347** 467 39
IMF/EU/SWF IMF/EU/SWF IMF/EU/SWF IMF/EU/SWF IMF/EU/SWF IMF/EU/SWF
* Estimates of funding needs and secondary market selling pressure taken from Exhibits 23 and 25 above. We assume T-bills continue to be rolled. ** Estimates of amounts raised from one-off wealth taxes and SWFs are taken from Exhibit 24 above. However, we assume that the IMF and EFSF contribute just enough to cover gross issuance and funding needs over the next 1Y, rather than providing three years funding in advance.
primary market funding needs of 350bn over the next year, and we believe that demand for new Italian and Spanish bonds will be limited. In addition, we estimate between 650-700bn of secondary market selling pressure over the next 12 months (Exhibit 25), based on the following assumptions: 1) 1.6tn of Italian and Spanish bonds outstanding that mature in 2013 or later, 2) total weighted-average non-domestic ownership of 43%, and 3) an estimated sell-rate of 2/3rds for non-domestic investors and 1/5th for domestic investors.27 Combined, this nets to over 1tn of funding demand/ selling pressure over the next year. Given the private sector resources available to Italy and Spain of roughly 135bn, we estimate that the supply/demand imbalance would be nearly 900bn in Italian and Spanish paper over the next year (Exhibit 26). This equates to around 75bn per month of supply/demand imbalance. Even if Italian and Spanish gross issuance needs were funded from other sources, the supply/demand imbalance would still be around 450-500bn of Italian and Spanish
Exhibit 27: Despite negative net supply in Italy, and significant ECB buying since August,
Gross bond issuance*, redemptions*, and SMP purchases in Italy and Spain** since 8 August 2011; bn
100 90 80 70 60 50 40 30 20 10 0 Italy 61 52
92
Gross issuance Redemptions SMP purchases**
31 21 14
Spain
* Excluding T-bills. Bond redemptions include conventional and non-conventional bond redemptions. ** The SMP has purchased a total of 123bn of peripheral paper in the roughly 3 1/2 months since they initiated purchases on 8 August 2011. We assume that these SMP purchases are split 3:1 between Italy and Spain. Source: Bloomberg, ECB, J.P.Morgan
paper over the next year, or roughly 40bn per month pre-SMP buying. To put these numbers in context, we note that the ECB has been purchasing about 35bn of peripheral paper per month since August. While the ECB does not identify individual countries, we assume the split has been roughly 3:1 between Italy and Spain. Over this same time period, net issuance has actually been negative for Italy (Exhibit 27). Despite the negative net issuance and
15
27 The estimate of investors looking to sell is purely a J.P.Morgan expectation, but we believe that our assumed sell rates are reasonable given anecdotal investor discussions and the empirical pattern of the past few months: Italian yields have risen over 200bp since August, despite sizable ECB buying.
sizable ECB buying, Italian yields have risen over 200bp since August (Exhibit 28). This indicates the strength of the selling pressure currently going through the secondary markets. Will the ECB be willing to double its current rate of purchases to eliminate the supply/demand imbalance? We make several points here. On the one hand, there are no hard capacity constraints that would prevent it from doing so.28 On the other hand, while the ECB has increased its balance sheet less than other central banks, it has arguably taken on a good deal more credit risk. The lions share of the growth in the ECBs balance sheet stems from increased support for peripheral country banks and peripheral sovereigns (please see Appendix4). Finally, based on recent official commentary, there appears to be growing reluctance to expand the size of SMP purchases. The German contingent remains firmly opposed to this and insists that monetary financing to sovereigns is against the ECBs mandate. Further, recent media reports suggest that this view is spreading. For instance, there is reported to be a limit on purchases of no more than 20bn per week. This is said to be a reduction from the original limit, and a limit that the ECB Council discussed lowering further, as late as last week.29 To be sure, such limits are self-imposed and can always be changed if market conditions change. However, it appears that at least over the near term, the ECB remains reluctant to step up the size of its buying. Going forward, our baseline view is that the ECB will not be willing to purchase debt in large quantities, unless the market forces its hand. This is due to 1) the need to maintain oversight and aid conditionality in order to mitigate moral hazard; 2) the desire to avoid blurring monetary and fiscal policy; and 3) a reluctance to take any steps that might damage credibility or independence.
Exhibit 28: peripheral yields have continued to rise after initiation of the SMP program
Portuguese and Italian 10Y yields around the start of ECB bond purchases for each country*; %
8.00 7.00 6.00 5.00 4.00 -3 -2
Italy
Portugal
-1 0 1 2 3 4 Months around first day of SMP purchases*
* The SMP started buying Portuguese and Italian bonds on 10 May 2010 and 8 August 2011, respectively. Portugal was eventually bailed out by the IMF/EFSF on 3 May 2011.
have been taken, these efforts will take far too long to bear fruit in order to deliver market stability in 2012.30 In the near term, as in all negative feedback loops or crises of confidence, outside intervention is required in the form of a lender of last resort. With the ECB reluctant to play that role, the path of least resistance is for the crisis to worsen. Thus, our baseline view is that the crisis stays on a slow boil over the next few months, possibly resulting in another country losing market access. In the event that such a risk plays out, we believe that the ECB will be compelled to start monetizing debt since the alternative would be significantly worse. There are a large number of risks that could go wrong in other countries as well (Exhibit 29). These include, for instance, the prospect of another round of PSI in Greece, a hard default in Greece, potential PSI in Ireland and Portugal, another country needing to tap the EFSF, and a French downgrade.
The 2012 outlook

The long-term solution to the EU sovereign crisis is consistent fiscal and structural reform, better governance in the Euro zone, more fiscal integration, and potentially jointly guaranteed debt issuance. While some small steps
The only hard restriction is that the ECB cannot purchase bonds directly from a sovereign (i.e. on the primary market) or set up credit facilities for sovereigns. 29 See Upper limit for bond purchases/Obergrenze fur Anleihekaufe, Frankfurter Allgemeine, 17 November 2011; and ECB hits back at calls for intervention, FT, 18 November 2011
16
28
30 See Overview, Global Fixed Income Markets Weekly, 19 August 2011, for discussion of the process required to institute Eurobonds.
Exhibit 29: The number of risks on the horizon is extremely large; we expect the peripheral crisis to escalate in 1H12
Risks in 2012, and possible policy responses / positive surprises
Effect Severe
Risk Greece fails to get further funding, leading to a hard default and potential ex it from the Euro zone Italy and/or Spain lose access to capital markets Italian coalition collapses, leading to early elections More sev ere PSI in Greece, triggering contagion fears in Italy and Spain
Significant Risks
French dow ngrade Ireland / Portugal require PSI Another country requires EFSF assistance (such as Cy prus) Italy drags its feet on fiscal consolidation / structural reforms Bank delev eraging depresses economic grow th Euro area in deep recession prev ents planned fiscal consolidation LCH continues to boost margin requirements, leading to further delev eraging in peripheral paper ECB w illing to monetise debt in large quantities
Long term
Potential policy responses / positive surprises
Germany agrees to fund ESM alongside EFSF Significant new money commitments from IMF/BRICs/SWFs Upside surprise on economic grow th Upside surprise on implementation of austerity measures
Given the large number of risks on the horizon, one or more of these threats are likely to play out in 2012. The combination of a steady stream of negative news and market disappointments will keep liquidity poor, spreads wide, and investors skittish.
Exhibit 30: Bond yields in bailed-out European countries have continued to soar even after the bailout request
Average of Greek, Irish, and Portuguese 10Y benchmark yields around the date when bailout was first requested*; %
11.00 10.00 9.00 8.00 7.00 6.00 -3 -2 -1 0 1 2 3 months around bailout request*
* Greece, Ireland, and Portugal formally requested a bailout on 2 May 2010, 21 November 2010 and 6 April 2011, respectively.
Trading strategies
In our view, the crisis is likely to worsen in 1H12, before the ECB feels compelled to step in and stabilize markets by monetizing debt. We therefore expect peripheral yields to continue to climb going into 1H12, especially if another country loses market access. Note that bond yields of a country that loses market access do not immediately stabilise; as demonstrated in Exhibit 30, official intervention has thus far provided only a temporary respite. Thus, short positions in peripheral debt have been profitable even when initiated postbailout. Additionally, in case of loss of market access, the debt of a country is likely to be downgraded which would exacerbate the selling pressure. Recent rating agency commentary around Italy and Spain, and the pattern of Greece, Ireland, and Portugal when they lost market access, suggests that Italy would be downgraded if it lost market access (Exhibit 31). Indeed, sovereign CDS spreads are trading significantly wider than their
similarly-rated corporate brethren, suggesting that the risk of further sovereign downgrades is high (Exhibit 32). Of course, the ECB will be quick to realize the severity of the situation if a large country loses market access, suggesting that this will be the tipping point where it will be compelled to start monetizing sovereign debt in large
17
Exhibit 31: Rating agencies have stated that Italy will be downgraded further if it loses access to debt capital markets. We expect another three-notch downgrade (from A to BBB) in that event
Number of notches that bailed-out sovereigns were downgraded by Moodys in the 3 months prior and 6 months after the bailout request* Dow ngrade (notches) Total dow ngrade
Exhibit 32: Indeed, sovereign CDS spreads are trading significantly wider than their corporate brethren, suggesting that the risk of further sovereign downgrades is high
CDS spreads of iTraxx components regressed against corporate Moodys rating vs. CDS spreads of sovereigns regressed against Moodys rating*; triangles show sovereigns and circles show corporates; bp
-3M to aid request Greece Ireland Portugal Average Italy 1 0 3 1 3
Aid request to +6M 4 7 4 5 -
since Jan-10 (notches) 14 9 9 11 3
2000 1500 1000 IT 500 0 ES PT IE
* Italy is currently rated A2 by Moodys and A with a negative outlook by S&P. Source: Bloomberg, Moodys, S&P
quantities. We expect that the eventual response from the ECB, IMF and EU will be on a scale sufficient to ultimately stabilise peripheral yields in 2H12, unlike the limited attempts made thus far with Greece, Ireland, and Portugal (Exhibit 33). We also see 10Y Bund yields falling to 1.25% by 2Q12 in a flight-to-quality, before rebounding to 1.75% by end-2012. Thus, we recommend underweighting peripheral duration and buying German duration going into 1H12. Below, we also present three different types of trading strategies that allow investors to take advantage of our bearish view on both the European economy and the peripheral situation. These strategies, in turn, offer 1) significant upside in a global recession, 2) positive convexity to peripheral risk, and 3) diversification away from the peripheral sovereign debt crisis. 1. Identifying the best long-duration trades in a lowyield world Valuations for fixed income are clearly stretched in the major markets. For instance, US, UK, and European 10Y real yields are currently below zero, and central bank policy rates are near lower bounds. Thus, we look for global markets where long duration positions offer more upside, without compromising too much on credit risk. In Exhibit 34, we highlight sovereigns that are rated A+ or above and, at the same time, offer positive real yields, attractive policy rate downside, and strong economic linkages to Europe (based on 10Y regression of annual nominal GDP growth). As a simple metric of attractiveness, we average each countrys cross-sectional z-scores across the three measures listed above. This captures countries that are most likely to be impacted by
18
Aaa
Aa2
A2 Baa2 Moody 's rating
Ba2
B2
Caa2
* Using a unit scale for Moodys rating (AAA=1, Aa1 = 2, Aa = 3, etc.), we model the following relationships based on current cross-sectional data: 5Y Sov. CDS = 76.6*(Moody's) + 38.7; R2 = 88% Sample based on 16 developed market sovereigns (US, UK, Japan, Australia, Sweden, Denmark, and 10 Euro area countries excluding Greece). 5Y Corp. CDS = 7.2*(Moody's)^2 - 73.8*(Moody's) + 358.9; R2 = 66% Sample based on nearly 200 corporates from the iTraxx High-grade, High-yield, and Cross-over indices. Source: Moodys, Bloomberg, Markit
Exhibit 33: Spreads and yield forecast: We expect the crisis to get worse initially, before concerted policy action pushes peripheral yields lower in 2H12
J.P. Morgan 2012 forecast for 10Y yields and benchmark spreads to Germany
Spread to Germany (bp) 18-Nov Germany France Greece Ireland Italy Portugal Spain Wtd av g., peripherals* -147 2425 625 494 910 469 699 2Q12 -225 2400 850 775 1200 700 955 4Q12 18-Nov -190 2400 900 625 1300 600 855 1.97 3.46 26.32 8.15 6.95 11.04 6.70 -Yield (%) 2Q12 1.25 3.50 25.25 9.75 9.00 13.25 8.25 -4Q12 1.75 3.65 25.75 10.75 8.00 14.75 7.75 --
* Weighted by proportional contribution to the J.P. Morgan EMU Bond Index.
the slowdown in growth in Europe and where policymakers have the most leeway to respond by slashing rates. Based on this metric, long Australian and Chilean 10Y duration appear to be the most attractive trades. We particularly like longs in Australia: the RBA is expected to cut rates and domestic technicals are strong (see Australia).
Exhibit 34: Look to global markets for higher real yields, potential policy rate cuts, and high sensitivity of GDP growth to Europe. Long 10Y duration in Australia and Chile appear to be the most attractive trades
Statistics on global benchmark nominal and real yields, central bank policy rates, and beta of GDP growth to Europe; % Policy Rate (%) Rating 10Y Yield Avg Inflation, Real yield
Continent Americas Americas Americas Europe Europe Europe Europe Europe Europe Asia/Pacific Asia/Pacific Africa/ME Asia/Pacific Asia/Pacific Asia/Pacific Asia/Pacific Av erage Country Chile Canada USA Czech Sw eden Germany Denmark UK Sw itzerland Australia China Israel New Zealand Japan Taiw an S Korea (S&P) AA AAA AA+ AA AAA AAA AAA AAA AAA AAA AAAAAA+ AAAAA+ -(%) 4.8 2.1 2.0 3.9 1.7 2.0 2.0 2.3 0.9 4.1 3.7 4.1 4.0 1.0 1.3 3.8 2.7 L10Y (% YoY) 3.1 2.1 2.5 2.4 1.8 2.0 2.1 2.5 0.8 2.9 2.7 2.4 2.9 -0.2 1.0 3.3 2.1 (%) 1.7 0.1 -0.5 1.4 -0.05 -0.04 -0.2 -0.2 0.1 1.2 1.0 1.7 1.0 1.2 0.3 0.5 0.6 5.25 1.00 0.13 0.75 2.00 1.25 1.20 0.50 0.13 4.50 6.56 3.00 2.50 0.05 1.88 3.25 2.12 -1.25 0 0 0 -0.75 -0.75 -0.75** 0 0 -0.75 0 -0.50 0 0 0 0 -0.30 Avg nominal GDP growth 10.0 4.9 4.2 5.1 4.1 3.2 3.0 4.1 2.9 7.0 14.8 5.7 5.3 -0.3 3.8 7.1 5.3 1.3 1.3 1.0 1.1 1.1 1.0 1.0 1.0 0.8 0.8 0.2 0.3 0.6 0.9 0.8 0.4 0.8 66% 90% 91% 84% 91% 100% 86% 92% 81% 86% 16% 26% 80% 89% 58% 43% 74% Average Zscore* 1.5 0.0 -0.7 0.5 -0.1 -0.3 -0.3 -0.5 -0.6 0.6 0.3 0.1 0.1 0.0 -0.2 -0.3 --
18-Nov 2H12 FC 18-Nov L10Y, % YoY Beta to Europe Corr to Europe
* Defined as the average of three cross-sectional z-scores: 1) 10Y real yield 2) level of central bank policy rate and 3) 10Y GDP beta to Europe. Z-score defined as (level crosssectional sample average)/cross-sectional sample standard deviation. ** J.P.Morgan does not forecast central bank rates for Denmark but here we assume that they cut rates in line with the ECB, as is their typical pattern. Note: Yields shown for benchmark government bonds. Real yield defined as nominal yield realised annual headline inflation over the past 10 years. Beta and correlation of GDP growth to Euro area GDP growth based on a 10Y regression of annual nominal GDP growth rate data. Source: Bloomberg, J.P.Morgan
2. Identifying trades that are positively convex in a risk-off (or risk-on) world Given the bimodal distribution of risks, trades that offer positive convexity to the crisis are attractive. We define a positively convex risk-off trade as one which performs well if the crisis worsens but does not lose much if the crisis stabilises. Similarly, we define a positively convex risk-on trade as one which performs well if the crisis abates but does not lose much if the crisis worsens. To identify such trades, we look at a wide selection of global assets. We regress asset levels against 10Y weighted peripheral spreads since 2010. Assets that have a sufficiently strong and convex relationship are presented in Exhibits 3536. Specifically, these are trades where the R-squared exceeds 65% and the tstatistic on the squared peripheral spread term exceeds +/- 4. Using the regression betas, we project gains and losses to the recommended position from a +/-100bp shock to peripheral spreads. The convexity in each position is defined as the additional gain that one makes
in the targeted scenario due to the convex performance profile, divided by the average absolute trade P/L for a +/-100bp move. For risk-off positioning, we find that Japanese 2Y swap spread wideners, long Aussie duration, and short France vs. Germany offer the greatest positive convexity (Exhibit 35). However, long Aussie duration is our preferred trade in that it is amongst the very few that offer positive carry and it is expected to benefit from policy rate cuts and strong technicals (see Australia). For risk-on positioning, short Kiwi 2Y duration, UK 10Y swap spread narrowers, and short gold offer the greatest positive convexity (Exhibit 36). As before, there are few assets that simultaneously offer positive carry and positive convexity. One that does is long US municipal credit.
19
Exhibit 35: Trades that offer the greatest positive convexity to peripheral spread widening include 1) Japanese 2Y swap spread wideners; 2) long Aussie duration; and 3) short France vs. Germany. Long Aussie duration offers positive carry as well
Global risk-off trades (trades with a positively convex performance profile when peripheral spreads are widening) Regression of asset against
wtd periph. spreads^ Trade Class Yields Yields Region Australia/NZ Australia/NZ Item Aussie 2Y Aussie 10Y UK 2Y SS JPY 2Y SS France - Germany Itrax x UK Fins Senior Itrax x Asia EU banks Italy MIB Positioning Long duration Long duration Wideners Wideners Wideners Wideners Wideners Wideners Short Short Curr 320 405 96 26 147 189 216 315 212 125 15233 R2 * 65% 71% 75% 66% 90% 76% 75% 71% 65% 82% 72% Beta* 26.1 -9.3 -3.5 -8.8 -7.6 6.8 1.5 -6.0 -11.0 -4.8 -589.2 Beta_sq* -6.6 -2.1 1.2 0.9 3.1 1.5 2.3 3.9 3.8 -1.7 -109.3 P/L for +/-100bp change in wtd periph. spreads^ Riskoff** 67 39 14 4 35 27 34 49 42 29 2120 Riskon** -54 -35 -11 -2 -29 -24 -29 -41 -35 -26 -1901 Difference 13 4 2 2 6 3 5 8 8 3 219 % 22% 12% 20% 65% 19% 11% 15% 17% 20% 13% 11% (Risk-off - Risk-on)*** Convexity^^
Sw ap spreads UK Sw ap spreads Japan EMU spreads Credit Credit Credit Credit Equities Equities Europe Europe Europe UK Asia Europe Europe
Maggie Banks Senior Wideners
* Trades based on polynomial regression vs. 10Y weighted peripheral spreads since January 2010. Trades shown have R-squared >65% and statistically significant convexity (T-stat vs. squared peripheral spread term > +/- 4). Beta = the regression beta vs. 10Y weighted peripheral spreads measured in % and Beta_sq = the regression beta vs. 10Y weighted peripheral spreads^2, again measured in %. A positive (negative) value of beta_sq indicates a convex (concave) asset. ** Modelled gain (loss) for a +100bp (-100bp) spread move, assuming peripheral spreads at 650bp. Gains and losses are calculated with respect to the recommended trade (e.g. reversing the betas where appropriate, or where beta_sq<0). *** Defined as the absolute value of gain in the risk-off scenario minus absolute value of gain in the risk-on scenario. ^ 10Y weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bond market). ^^ % Convexity defined as the difference in the absolute value of gains divided by the average of the absolute values of gains in both the risk-off and risk-on scenarios. Units: yields, curve, swap spreads, credit spreads shown in bp of yield. Equities and commodities shown in index points. FX are in ratios.
Exhibit 36: Trades that offer the greatest positive convexity to peripheral spread narrowing include 1) short Kiwi 2Y duration 2) UK 10Y swap spread narrowers; and 3) short gold. Long US muni credit is one of the few that offers positive carry as well
Global risk-on trades (trades with a positively convex performance profile when peripheral spreads are narrowing) Regression of asset against
wtd periph. spreads^ Trade Class Yields Curv e Region Australia/NZ UK Item Kiw i 2Y 2s/10s UK 10Y SS Denmark 10Y SS 10Y Muni Gold CHF/HUF Positioning Short duration Steepeners Narrow ers Narrow ers Narrow ers Short Short CHF Curr 250 178 31 65 216 1720 246 R2 * 78% 75% 83% 80% 82% 89% 89% Beta* -72.5 -30.0 12.4 12.6 7.9 225.0 20.6 Beta_sq* 4.8 1.3 -0.7 -0.7 -0.3 -13.1 -1.2 P/L for +/-100bp change in wtd periph. spreads^ Riskoff** -6 -12 -3 -3 -3 -41 -3 Riskon** 15 14 4 4 4 67 6 Difference 10 3 1 1 1 26 2 % 90% 20% 41% 36% 18% 49% 53% (Risk-on - Risk-off)*** Convexity^^
Sw ap spreads UK Sw ap spreads Scandis Credit Commodities FX USA Global FX
*, **, ^,^^: See footnotes in Exhibit 35 above. *** Defined as the absolute value of gain in the risk-on scenario minus absolute value of gain in the risk-off scenario.
20
Exhibit 37: The elusive quest for diversification from the peripheral crisis in Europe: US prime RMBS and Auto ABS carry positively and spreads are expected to be stable to firmer in 2012
Trades with low correlation* to EU peripheral spreads
Canada: 5s/10s; EUR/CAD
Denmark: UK: 2s/10s; Rate Vega
USA: Prime RMBS; AAA Auto ABS; CDS-cash basis;
Japan: 10s/30s; Korea: USD/KRW; EUR/KRW;
Brazil: Real/CLP
Notes: * Trades based on polynomial regression vs. 10Y weighted peripheral spreads since January 2010, as described in footnotes in Exhibit 35 above. Trades shown here have R2 <20% over entire sample period and R-squared < 30% since January 2011. Around 5% of the nearly 200 trades sampled meet these criteria.
3. Identifying trades that are orthogonal to the peripheral debt crisis and therefore offer diversification benefits One of the difficulties of investing in the current market environment is that return correlation is extreme. Indeed, one could be forgiven for assuming that we live in a onefactor world (with that factor being sovereign risk). Trades that are relatively uncorrelated with peripheral spreads can add value in a portfolio context, especially if they carry positively. We screen nearly 200 asset classes globally and find that just a few, in fact around 5% of those sampled, demonstrate a significant lack of correlation. Specifically, Exhibit 37 highlights asset classes for which the R-squared in a polynomial regression vs. peripheral spreads is both 1) less than 20% in 2010-2011, and 2) less than 30% in 2011. Of these, we note that prime RMBS and Auto ABS in the US carry positively. Moreover, J.P.Morgan strategists expect spreads in these sectors to be stable to firmer in 2012.
21
Trading themes
Be long German and UK duration; enter 2s/10s curve flatteners We expect the peripheral crisis to escalate in 1H12, driving peripheral yields higher and German and UK yields lower in a flight-to-quality dynamic. Further, a Euro area recession, slow growth in the UK, and disinflation will add downwards pressure to intermediate yields, flattening the 2s/10s curve. Underweight intra-EMU paper vs. Germany We expect intra-EMU spreads to widen as the peripheral crisis escalates in 1H12. Also, ratings downgrades and poor technicals will contribute to upwards pressure on peripheral yields. Go long 10Y duration in Australia and Chile Global markets that offer high real yields, potential policy rate cuts, and high sensitivity of GDP growth to Europe are attractive. Long Australian and Chilean 10Y duration appear the most attractive; we particularly like longs in Australia given advantageous technicals and expected central bank rate cuts.
Position in trades that are positively convex to peripheral spread widening Trades that perform well when peripheral spreads are widening but give up little when they are narrowing are attractive risk-off trades. We find that Japanese 2Y swap spread wideners, long Aussie duration, and short France vs. Germany offer the greatest positive convexity to peripheral spread widening. Long Aussie duration is also attractive as it is one of the few trades which offer positive carry as well. On the other hand, short Kiwi 2Y duration, UK 10Y swap spread narrowers, and short gold offer the greatest positive convexity to peripheral spread narrowing. US Prime RMBS and Auto ABS offer diversification away from the peripheral crisis We screen nearly 200 asset classes globally and find that just a few, around 5%, demonstrate a significant lack of correlation to the sovereign debt crisis. Of these, we note that US prime RMBS and Auto ABS carry positively; J.P.Morgan strategists expect spreads in these sectors to be stable to firmer in 2012.
Exhibit 38: Birds eye view of our major trade recommendations by currency
Duration Curve Swap spreads Euro area Long 10Y Long 6Mx6M EONIA 2s/10s flatteners 1s/5s bull flatteners 2Y wideners UK Long 10Y 2s/10s flattener 10s/30s flattening bias 10Y wideners 10Y wideners FRA/OIS wideners Short end wideners Long end narrowers TIBOR/LIBOR narrowers 3s/6s wideners Swap spread curve Flattening bias Gamma Vega Inflation Cross-market Long gamma in 10Y tails Short gamma in 2Y tails Neutral Neutral Long gamma in 10Y tails Short gamma in 2Y tails Neutral Long Short Short breakevens Flattening bias Neutral Neutral Neutral US Long 10Y Long end steepening bias Japan Long bias Swap curve flattening bias
1s/10s HICP swap curve flattener Long intermediate real yields Long French CPI-linked vs. Euro HICP linked Short 10Y breakevens Wider intra-EMU spreads Underweight Italy and France vs. Germany Credit curve flatteners Short Euro vs. US breakevens
22
Appendix-1: Timeline of key events in the EU sovereign debt crisis

2009 04-Oct-09 20-Oct-09 2010 13-Jan-10 14-Jan-10 02-Feb-10 11-Feb-10 08-Mar-10 16-Mar-10 18-Mar-10 24-Mar-10 25-Mar-10 30-Mar-10 12-Apr-10 21-Apr-10 23-Apr-10 27-Apr-10 02-May -10 03-May -10 05-May -10 06-May -10 07-May -10 07-May -10 10-May -10 12-May -10 18-May -10 21-May -10 26-May -10 28-May -10 14-Jun-10 15-Jun-10 16-Jun-10 23-Jul-10 18-Oct-10 01-Nov -10 01-Nov -10 10-Nov -10 16-Nov -10 21-Nov -10 28-Nov -10 23-Dec-10 European Commission report accuses Greece of statistics fraud; new Finance Minister say s: "There is no skeleton in the closet" Greece adopts three-y ear plan to bring the European Unions biggest budget deficit to w ithin the EU limit in 2012 Greek gov ernment announces austerity package to get deficit to 3 percent of GDP in 2012 EU leaders hold first emergency summit on Greece. EU agrees to take determined and coordinated action to protect financial stability of euro area, w ithout giv ing further details Portuguese gov ernment announces new budget cuts, more asset sales and a freeze on public w ages Euro region finance ministers lay groundw ork for making emergency loans av ailable to aid Greece Papandreou calls on EU partners to come up w ith specific aid measures w ithin a w eek to help Greece, hints he might seek support from IMF if EU partners dont act Fitch cuts Portugals credit rating to AATrichet say s that the ECB w ill continue to accept bonds rated as low as BBB- as collateral. An EU summit agrees to general principles behind a Greek bailout, including IMF inv olv ement, but again prov ides no details Ireland say s country s banks need to raise an additional 31.8bn of capital Euro area finance ministers agree to prov ide up to 30bn of loans to Greece ov er the nex t y ear w ith the IMF agreeing to put up another 15bn in funds The EU say s Greeces 2009 budget deficit w as w orse than it prev iously forecast and could top 14 percent of GDP as off- market sw aps cloud its estimates Papandreau asks the EU for aid S&P cuts Greece rating to junk (BB+), dow ngrades Portugal to AEuro area agree to upsize its rescue package to Greece to 110bn. Greece agrees to 30bn in austerity cuts ov er the nex t 3 y ears in ex change for the aid The ECB say s it w ill indefinitely accept Greek collateral regardless of the country s credit rating Protests in Athens against the gov ernments austerity plans turn v iolent and 3 people are killed w hen they become trapped in a bank set ablaze by demonstrators Flash crash in US stocks European leaders agree in principle to set up a 440bn emergency temporary fund to stem the sov ereign crisis, w hich w ill come to be know n as European Financial Stability Facility (EFSF) In Spain, the Gov ernment and the main opposition party reach agreement ov er mergers betw een cajas and the need of sav ings-bank regulation reform EU finance chiefs agree to establish a 750bn permanent aid facility to succeed the EFSF (the European Stabilisation Mechanism, or ESM). Also, the ECB 1) ex tends unlimited liquidity for banks, 2) re-opens $ sw ap lines w ith the Fed, and 3) announces a gov ernment bond purchase program to focus on peripheral sov ereign debt (the SMP). This is the biggest attempt y et to solv e the peripheral crisis Spain announces public-w age cuts and a pension freeze Germany surprises the markets w ith a ban on naked short selling of debt securities, CDS and shares in 10 financial institutions. German parliament approv es its share of the EFSF commitments Italian Prime Minister Silv io Berlusconi's ministers approv e 24bn in budget cuts for 2011-2012 Fitch cuts Spains AAA rating one lev el to AA+ Moody s cuts Greek rating to junk (Ba1) Chinese sign multibillion euro contracts w ith Greece hours after Moody 's debt dow ngrade The French gov ernment Wednesday unv eiled a controv ersial plan to gradually raise the age at w hich w orkers can retire w ith pension benefits Europe publishes the results of the first round of bank stress tests. 7 banks failed w ith a total capital shortfall of 3.5bn German Chancellor Angela Merkel and French President Nicolas Sarkozy agree that priv ate inv estors may be required to contribute to future EU bailouts v ia a mandatory solv ency rev iew of countries requesting aid; they also declare that official loans granted out of the ESM after 2013 should hav e preferred creditor status ov er priv ate debt inv estors Portugals gov ernment and biggest opposition party agree to let nex t y ears budget pass, aiming to stem the euro regions fourth-biggest fiscal shortfall Greek minister Pangalos suggests to Greek media that: "Demonizing debt restructuring is w rong. Debt ex ists to be restructured. We may pursue it ourselv es or it may be proposed to us and it may be too adv antageous to turn it dow n." LCH.Clearnet increases haircuts for Irish bonds December aid tranche for Greece is delay ed until January , after the Austrian and Finnish ministers say Greece hasn't met EU criteria Ireland say s it w ill apply for a bailout Ireland gets 85bn bailout. European leaders scale back proposals to inflict losses on bondholders in future EU bailouts Fitch cuts Portugal to A+ rating George Papandreou leads Socialist Pasok Party to landslide v ictory in Greek elections New Greek Finance Minister Papaconstantinou say s deficit w ill rise to 12.5 percent of GDP this y ear, triple the prior forecast
Source: Bloomberg, Reuters, WSJ, FT, NY Times, Telegraph
23
Timeline of key events in the EU sovereign debt crisis, continued

Date 2011 24-Jan-11 11-Feb-11 25-Feb-11 10-Mar-11 11-Mar-11 15-Mar-11 21-Mar-11 23-Mar-11 25-Mar-11 27-Mar-11 31-Mar-11 01-Apr-11 06-Apr-11 15-Apr-11 17-Apr-11 03-May -11 16-May -11 16-May -11 17-May -11 24-May -11 01-Jun-11 02-Jun-11 05-Jun-11 15-Jun-11 16-Jun-11 17-Jun-11 21-Jun-11 24-Jun-11 29-Jun-11 29-Jun-11 30-Jun-11 02-Jul-11 04-Jul-11 06-Jul-11 07-Jul-11 08-Jul-11 08-Jul-11 11-Jul-11 15-Jul-11 21-Jul-11 02-Aug-11 05-Aug-11 05-Aug-11 Spain announces new capital requirements for banks Ax el Weber resigns from Bundesbank after opposing the ECBs crisis policy Fine Gael w ins elections and knocks out the current party from the Irish gov ernment, establishing a coalition w ith Labour. Backtracks from demands for bank debt burden-sharing but continues its demands for low er interest rates on official loans Bank of Spain announces 15bn capital shortfalls in national bank stress tests EU summit agrees to ex pand pow ers of EFSF by 1 ) allow ing it to buy debt in primary markets and 2) upsizing guarantees to tap its full 440bn in firepow er EcoFin meeting agrees on new Stability & Grow th competitiv eness pact EU finance ministers decide on mechanisms for allow ing the regions permanent bailout mechanism, the ESM, to lend 500bn in 2013, Greece's loan rates are low ered by 100bp and loan maturity is ex tended to 7.5y rs; Ireland's loan terms are unchanged Portuguese prime minister resigns ov er loss of support stemming from austerity measures Agreement on capital structure and lending terms of the ESM at the EC heads of state meeting Merkel's coalition loses in state elections in Baden-Wurttemberg, narrow ing her support in the upper house of Parliament. Irish bank stress tests announce 24bn shortfall across the remaining 4 un-nationalized banks; 5bn w ill come from sub debt w ritedow ns ECB agrees to drop its ratings requirement on Irish-backed paper at its tender operations Portugal requests aid from the EU Papandreou announces 76bn of austerity measures, later increased to 78bn, running through the end of 2015 In the Finnish national election, True Finns, a Euro-skeptic party , polls third in a tight 4-w ay race, taking roughly 20% of seats in the legislature Portugal reaches agreement on a 78bn aid package Eurogroup meeting endorses Portuguese bailout package Eurogroup meeting endorses Draghi as the nex t ECB chief European finance ministers propose talks w ith bondholders around ex tending Greeces debt-repay ment schedule Greece releases a new medium-term plan w ith more aggressiv e fiscal consolidation and priv atization targets IMF holds off releasing the nex t tranche of Greece's aid package until a new bailout package, w hich meets neccessary funding needs for at least 1 y ear, is approv ed The ECB and EU to come to truce around Greek PSI v ia a "Vienna initiativ e" approach Centre-right party w ins conv incingly in Portuguese parliamentary election; most Portuguese parties hav e already signaled their support for required austerity measures A parliamentary crisis in Greece sends markets into a tailspin. Faced w ith party defections, the Greek prime minister calls for a no-confidence v ote to be held on 21 June. EU emphasises that the disbursement of the nex t tranche of aid package requires approv al of all new austerity measures The IMF agrees to release the nex t tranche of Greek aid w ithout formal agreement on a new package in place Merkel and Sarkozy agree in principle to a "Vienna-initiativ e" debt rollov er for Greece Greek prime minister w ins no-confidence v ote EU finance ministers agree on details around EFSF upsizing and remov e preferred creditor status from ESM loans to Greece, Ireland and Portugal Greek parliament approv es austerity measures (principles) French banks propose a Greek debt rollov er w ith modest haircuts Greek parliament approv es austerity measures (implementation measures) EU releases nex t tranche of Greek aid package S&P announces that the French proposal for the Greek debt rollov er w ould garner an SD rating Follow ing rating agency comments that Vienna initiativ e approach w ould most likely garner an SD rating, German and Dutch officials return to calls for a formal debt restructuring in Greece. The ECB remains adamantly opposed ECB agrees to drop the ratings requirement on Portuguese-backed paper at its tender operations IMF releases nex t tranche of Greek aid package Contagion spreads to Italy . Negativ e headlines around finance minister and banks under stress (Unicredit trading halted) Munitions blast at Cy prus nav al base leads to sev eral multi-notch rating dow ngrades and fears of a potential Cy priot bailout. Sev eral ministers resign and protests call for the dissolution of the ruling gov ernment coalition 2nd round of EBA bank stress test results released; net 2.5bn capital shortfall estimated across 8 banks Details of 2nd Greek aid package announced; IIF banking federation launches a v oluntary Greek debt rollov er Italian 10Y y ield closes abov e 6%, at highest lev els since joining the EMU Italian Prime Minister Berlusconi announces new reforms and promises to front-load budget cuts in an apparent quid pro quo; the ECB begins buy ing Italian and Spanish debt the w eek after S&P dow ngrades the US to AA+ Event
24
Timeline of key events in the EU sovereign debt crisis, continued

Date 08-Aug-11 16-Aug-11 02-Sep-11 ECB starts buy ing Italian and Spanish debt in its SMP purchase program Finland's deal to get collateral in ex change for loans granted to Greece in Greece's 2nd aid package becomes public and sparks controv ersy amongst other Euro area countries IMF/EU/ECB officials end Greece's 5th quarterly rev iew early , unhappy w ith the pace of structural reforms. They delay decision on granting the September aid tranche to Greece Sw iss central bank announces it w ill target a 1.2 floor for Sw iss/Euro ex change rate German constitutional court rules that the EFSF structure is constitutional but requires more strict ov ersight form Parliament (mandatory Parliamentary Budget committee approv al before aid is distributed). The ruling has bearish implications for Eurobonds Greece announces it w ill speed up reforms and collect a new property tax to help bridge the budget gap Italian parliament giv es final approv al in a confidence v ote to a 54bn austerity package to balance the budget by 2013 S&P dow ngrades Italy 1 notch to A Slov enian gov ernment loses a confidence v ote w hich threatens to delay EFSF ratification Bundestag v otes to approv e EFSF reforms Greeces gov ernment approv es the draft budget for 2012 EU finance ministers w ork out a rev amped deal on collateral for Greek loans that satisfies Finnish demands and those of other euro-region gov ernments opposed to a bilateral deal for Finland Fitch cuts Spain rating to AA- and Italy rating to A+ Troika releases statement on the fifth rev iew of Greek economy and suggests that the six th tranche of the bailout pay ments, w orth 8bn, w ill be paid Slov akia becomes the final country to approv e the upsized EFSF facility . The ruling gov ernment lost a confidence v ote tied to the initial EFSF v ote w hich w ill lead to new parliamentary elections Papandreou w ins parliamentary approv al of the latest austerity bill in Greece, w hich includes w age and pensions cuts and plans to lay off 30,000 state w orkers European leaders say a summit on the euro crisis w ont produce decisions and set another meeting for 26 October EU leaders agree to 1) lev erage the EFSF to boost its firepow er to 1tn, 2) force priv ate inv estors to accept a 50 percent haircut on Greek bonds, 3) push Euro zone banks to raise 104bn in new capital, and 4) ex tend a new aid package w orth 130bn for Greece Surprising his ow n party , EU officials, and markets, Papandreou calls for a public referendum on the second bailout agreement in Greece European leaders suspend aid pay ments to Greece and say Greece must decide w hether it w ants to stay in the euro Papandreou backs dow n on referendum Papandreou agrees to step aside to make w ay for a gov ernment of national unity in Greece Berlusconi fails to secure a majority in a v ote on public finance in the low er house of the Italian parliament. He later announces that he w ill resign once he has secured passage of the upcoming budget bill. Italian 10Y y ield crosses the 7% mark LCH.clearnet increases margin requirements on Italian paper by 3.5-5.0% points across the curv e Greek political leaders agree on a new gov ernment, Papandreou steps dow n Former ECB v ice-president Lucas Papademos is tapped to head Greece's new coalition gov ernment The Italian parliament passes the budget bill and Berlusconi resigns President Napolitano nominates former European Commissioner Mario Monti to become prime minister of Italy , heading a new technocratic gov ernment The Papademos-led Greek gov ernment w ins a v ote of confidence Mario Monti is sw orn in as Italian Prime Minister Event
6-8 Sep 2011 A w ide sw athe of German politicians declare that Greece w ill not get its nex t aid tranche w ithout a positiv e IMF rev iew and sufficient progress on reforms 06-Sep-11 07-Sep-11 11-Sep-11 14-Sep-11 19-Sep-11 20-Sep-11 29-Sep-11 02-Oct-11 3-4 Oct 2011 07-Oct-11 11-Oct-11 13-Oct-11 21-Oct-11 23-Oct-11 26-27 Oct 2011 31-Oct-11 02-Nov -11 03-Nov -11 06-Nov -11 08-Nov -11 08-Nov -11 09-Nov -11 10-Nov -11 12-Nov -11 13-Nov -11 16-Nov -11 16-Nov -11
17-18 Nov 2011 Mario Monti w ins confidence v ote in both houses of the Italian Parliament
25
Appendix-2: What is the maximum amount of funds that Italy can raise via one-off wealth taxes and asset sales?
There has been speculation in the media around the extent of funding that Italy can raise via one-off wealth taxes and asset sales. In this appendix, we attempt to quantify the maximum amount that Italy can potentially raise via three sources of funding: 1. 2. 3. A one-off tax on bank deposits, securities, and property, Privatisation of financial and non-financial assets, and Partial sale of gold + FX reserves held at the central bank
Exhibit A1: Even assuming onerous one-off wealth taxes can be imposed on Italian households, the maximum that the government can raise is 6070bn, or 3 months of Italian funding needs
J.P.Morgan expectation of maximum one-off taxes that may be raised by Italy
Wealth bn Housing Land Non-residential buildings Valuable items Other Deposits, cash, post office Italian gov ernment bonds Other securities and mutual funds Insurance technical reserv es Commercial credits Total 4830 241 335 124 347 1062 189 1576 631 107 9442 % of total 51% 3% 4% 1% 4% 11% 2% 17% 7% 1% 100% 66 60 100 60 6 2 9 bp 100 bn 48 Assumed one-off tax
Exhibit A1 shows the wealth in Italy as of 2009, as estimated by the Bank of Italy. Around half of the total wealth of around 9.4tn is in housing, and another almost one-third is held in cash deposits, Italian government bonds, and other securities. We assume that, in extremis, Italy will be able to impose a one-off tax of 100bp on housing and Italian government bonds,31 and 60bp on cash deposits/other securities and mutual funds.32 Such a move would raise 6070bn, or around 3 months of Italian funding needs. On the privatisation front, it is difficult to get accurate data on holdings of financial and non-financial assets held by the Italian government.33 Additionally, it may be hard to make assumptions on the possibility of largescale privatisations given the difficult economic environment. We therefore look at historical precedent to estimate how much money can be raised via privatisations, since Italy has been privatising assets over the past several years. Exhibit A2 shows the amount of financial and non-financial assets that Italy has privatised on average over the past several years, as a percentage of GDP.34 Also shown is the amount of financial assets privatised in the top 3 years. Although data on peak non31 In the US, an annual tax on property is levied by the local (town) authority to fund the local school system. This tax ranges from 0.2% to 4.0% (average 1%) of the value of the property. We therefore assume that Italy can impose such a one-off tax. For more information, see: http://www.taxfoundation.org/publications/show/1913.html. 32 There is historical precedent for a 60bp tax on cash deposits. In 1992, Italy imposed such a tax under Prime Minister Giuliano Amato. 33 For example, a recent presentation on the Italian Treasurys website has data on Italian government holdings going no further than 2005. 34 Note that since we use two different data sources, the data periods for privatisation of financial and non-financial assets do not match.
Source: Household wealth in Italy 2009, No. 67, Supplements to the Statistical Bulletin, Banca D'Italia, Eurosystem
Exhibit A2: Historically, Italy has been able to achieve peak annual privatisation of 1.9% of GDP for financial assets, and an estimated 0.9% of GDP for non-financial assets. Achieving a similar high rate of privatisations in 2012 will likely raise 4045bn of funding*, or 2 months of Italian funding needs
Average and maximum level of privatisation receipts achieved by Italy in past years Priv atisation receipts;
% of GDP Financial assets (1987-09) Av erage Top 3 y ears** Non-financial assets (2000-06) Av erage Top 3 y ears*** 0.3% 0.9% 0.7% 1.9%
* Italian GDP is around 1.5tn, so 2.8% of GDP works out to around 4045bn. ** Top 3 years for financial asset privatisation were 1997, 1999, and 2005 when 19bn, 27bn, and 26bn were privatised, respectively. Source: Privatisation barometer. *** Italy privatised non-financial assets worth 0.3% of GDP between 2000-2006. Although data on peak non-financial asset privatisation is not available, we estimate it from the ratio of Top 3 years to average financial assets privatised. Source: Should Italy Sell Its Nonfinancial Assets to Reduce the Debt?, Stefania Fabrizio, IMF Policy Discussion Paper, April 2008.
financial asset privatisation is not available, we estimate it from the ratio of Top 3 years to average financial assets privatised. We give Italy the benefit of the doubt and assume that it can privatise assets at the maximum rate achieved over the past few years. Since Italy has historically achieved peak annual privatisation of 1.9% of GDP for financial assets, and around 0.9% of GDP for non-financial assets, we conclude that a similar high rate
26
of privatisations in 2012 will likely raise 4045bn of funding,35 or around 2 months of Italian funding needs. Additionally, the Bank of Italy holds gold + FX reserves in excess of 8% of GDP, compared to a Euro area average of just under 6% (Exhibit A3). Bringing them in line with the average would generate funding of around 2% of GDP (30bn), or around 1-1/2 months of funding needs.36,37,38 Note that the Bank of Italy also owns around 44bn of European sovereign debt (excluding BTPs), a part of which may also be saleable. Overall, we find that, in extremis, Italy may be able to raise around 125150bn over the next year via a wealth tax and aggressive asset sales, allowing it to avoid tapping capital markets for as long as 68 months. In the long run, however, it is going to be difficult for Italy to raise taxes significantly in order to achieve a primary surplus as it is already one of the most heavily taxed nations in the OECD world (Exhibit A4). Instead, it will need to rely on increasing the tax base and boosting growth.
Exhibit A3: Italys holdings of gold + FX reserves are significantly higher than the Euro area average; bringing these assets in line with the average should yield around 2% of GDP (30bn), or around 1-1/2 months of funding needs*
Gold and FX reserves of various Euro zone countries as a % of GDP
Gold** Tonnes Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Total 280 228 49 2,435 3,401 112 6 2,452 613 383 282 10,239 bn 12 10 2 103 143 5 0 103 26 16 12 431 FX reserv es (bn) 5 6 4 22 29 0 0 26 7 1 12 112 GDP (bn) 301 370 190 1,988 2,567 218 156 1,586 607 172 1,075 9,231 bn 17 16 6 125 172 5 1 129 33 17 23 543 Gold + FX res. %GDP 5.7% 4.2% 3.3% 6.3% 6.7% 2.2% 0.4% 8.1% 5.5% 9.7% 2.2% 5.9%
* Note that the Bank of Italy also owns around 44bn of European sovereign debt (excluding BTPs), a part of which may also be saleable. ** Gold spot level used: 1310/oz. Source: World Gold Council, ECB
Exhibit A4: It may be difficult for Italy to raise taxes significantly on an ongoing basis, given that it is already one of the highest taxed countries in the OECD world
Total tax revenue as % of GDP*; %
50 45 40 35 30 25 20 24 28 29 31 34 35 37 42 43 43 39 43 44
45
Portugal
Netherlands
Austria
Ireland
Spain
France
Finland
Greece
Italy
US
UK
35 Italian GDP is around 1.5tn, so 2.8% of GDP works out to around 4045bn. 36 Technically, governments cannot use gold for funding needs since central banks are legally independent and free from government influence. Indeed, according to an FT article entitled Italy to use gold reserves to cut national debt, 1 August 2007, Italys government cannot unilaterally infringe the central banks independence and autonomy, although the bank could freely agree to contribute some of its reserves. Given the severity of this crisis, we believe that the central bank will not hesitate in making such a contribution, especially since it would contribute to financial stability. 37 The ECB will need to approve the sale of gold and FX reserves by the Bank of Italy. Since Italy will be selling such assets merely to bring them in line with the Euro area average, we believe that such approval will be forthcoming. 38 In 2009, central banks in the Eurosystem (together with Sweden and Switzerland) agreed to not sell more than 400 tonnes of gold per year in total, or around 17bn at current market prices. Given the severity of the crisis, however, we believe that such restrictions may be relaxed in order to help Italy.
* Data as of 2008 for Netherlands and Portugal, and as of 2009 for the rest of the countries. ** Scandi average includes Norway, Denmark, and Sweden. Source: OECD
Germany
**Scandi
27
Belgium
Appendix-3: What is the maximum amount of funds that sovereign wealth funds (SWFs) might be willing to invest in EU peripherals?
EU officials are targeting emerging market countries such as Brazil, China, and Russia, countries with large amounts of FX reserves such as Japan, and countries with significant natural resources such as Norway and Qatar to help peripheral EU sovereigns. In particular, the plan to leverage EFSF capacity increases the importance of finding equity or debt investors from amongst this pool. Despite stepped-up entreaties from EU officials, we do not think these nations will be willing to contribute a large amount of funds. We base our view on several factors. 1) While sovereign wealth funds do control a large amount of resources, nearly 2tn in total, the pool of resources available to EU sovereigns is much smaller than widely believed. This is because of several reasons. First, SWFs are typically highly concentrated in equity. Based on 1H12 data from Monitor, a consulting group, we estimate that less than a quarter of SWF assets are allocated to fixed-income. Second, SWF assets are diversified across currencies; we estimate that a little over one-third is allocated to European currencies. This implies that in total, just around 170bn is currently allocated to European fixed income (170bn = 2tn *
Exhibit A5: Despite having nearly 2tn of resources, very little of sovereign wealth fund resources are allocated to European fixed income (170bn). We assume that this allocation to European fixed income could be increased at most by one-third, or 5560bn
SWF holdings and amount allocated by region and sector as of 2Q11; bn Region w eight (%) and allocation (bn)
Total Sector Total --Fix ed Income --Equity --Other --Cash Sector w eight (%) 100% 24% 67% 8% 1% 100% 2,012 476 1,340 165 30 Europe 36% 720 170 479 59 11 Americas 38% 771 183 514 63 12 Asia/EM 26% 520 123 347 43 8
Note: Converted from USD using 1.35 EUR/USD FX rate. Source: Monitor Company Group, J.P. Morgan estimates
24% * 36%, as shown in Exhibit A5). Third, this amount includes non-Euro area currencies such as sterling and Swiss franc, and non-sovereign exposures such as corporate credit, structured finance, covered bonds, etc. Although few SWFs break out portfolio composition in much greater detail than this, it is likely that the pool of resources available to euro area sovereign debt is smaller still. 2) Recent commentary suggests that these countries are reluctant to provide large amounts of aid. This is especially true given that EU officials have indicated they do not plan to make political concessions in order to obtain commitments. Exhibit A6 highlights recent statements by high-ranking officials in various countries.
Exhibit A6: Recent commentary suggests that most non-Euro area countries are reluctant to provide aid, particularly countries that are less wealthy than the EU peripherals
Selection of recent comments regarding SWF or FX reserve investments in peripheral sovereign debt Country Speaker Date Comment
Russia Sergey Ignatiev , Central bank chairman Zhu Guangy ao, Vice Finance Minister Oey stein Olsen, Central Bank Gov ernor Norway Yngv e Sly ngstad, Chief Ex ecutiv e of the Norw ay oil fund Jens Stoltenberg, Prime Minister Yoshihiko Noda, Prime Minister Senior Japanese finance minister 18-Nov regarding the rescue fund. "We, of course, must w ait until its structure is ex tremely clear. And moreov er, this inv estment must be decided on after serious, technical discussions." "It's not on the agenda to contribute any ex tra inv estments in the special purpose inv estment v ehicle (SPIV) that hav e elements of help or aid. That's outside of the mandate." "We do not hav e sufficient detail really to comment on w hat it w ould do for us in terms of inv estment 28-Oct opportunity ... Of course w e are committed to Europe although w e hav e signalled that ov er time the fund w ill hav e less of its inv estment in Europe." 26-Oct 02-Nov Norw ay w ill not participate in any aid package...That is not our task. I believ e its w rong that Norw ay should take part in such support measures. We w elcome the [EU summit] package, and... "from the standpoint of supporting such efforts, w e w ill continue to study the purchase of EFSF bonds." Negativ e Positiv e Positiv e Negativ e Pos/Neg Negativ e Negativ e Negativ e Negativ e
Russias central bank isnt planning to purchase debt issued by the EFSF until theres more clarity
02-Nov Its now too soon for his country to contribute. 28-Oct 28-Oct
China
Japan
01-Nov "I told [Mr. Regling] that w e w ill continue to purchase EFSF bonds."
Source: WSJ, Bloomberg, Reuters

28
For instance, Norway looks opposed to making any investments on concessionary terms, and China and Russia have also emphasised that any investment will depend on the structure of the investment, how attractive the investment is, and how extensive EU co-investment is. It is especially unlikely that countries which are less wealthy will see a compelling reason to support relatively wealthy EU countries. Thus, we believe that the 170bn allocated to European fixed income will at most be increased by one-third (or 5560bn) as a contribution towards peripheral sovereign debt purchases. Even this would represent a large portfolio concentration of risk and would likely require considerable arm-twisting and/or political quid pro quos.
Exhibit A7: The ECB has grown the size of its balance sheet by a smaller percentage than the Fed and BoE over the past few years...
Total assets held on central bank balance sheets, current vs. 4Q06 3Q11 4Q06 Change
bn of ntl ccy ECB Fed BoE 2,289 2,871 243 1,134 863 81
bn 1,154 2,008 162
% 102% 233% 199%
Source: ECB, Federal Reserve, Bank of England
Exhibit A8: but asset growth is heavily skewed towards riskier exposures such as peripheral bank repo and ELA* and SMP purchases of peripheral sovereign debt
ECB balance sheet composition since 2007**; bn
2400 2200 2000 1800 1600
Cov ered bond purchases SMP
repo to periphery banks/ELA* repo to core banks
Appendix-4: The extent of ECB aid to peripheral countries

Since the onset of the global financial crisis in mid-2007, the ECB has grown its balance sheet by a substantial amount, over 1.1tn. This is about half the rate of increase seen in other central bank balance sheets (Exhibit A7). However, the lions share of the increase (nearly 700bn) comes from increased support to peripherals, via repo operations to banks in peripheral countries, emergency liquidity assistance (ELA39) to banks in peripheral countries, purchases of covered bonds (largely from Spanish banks), and purchases of peripheral sovereign debt (Exhibit A8).
1400 1200 1000 800 600 2007 2008
All other balancesheet items** 2009 2010 2011
* ELA, or emergency liquidity assistance, is typically included in Other assets or Other claims on central banks balance sheets. We estimate the current amount of ELA outstanding based on Irish and Greek central bank data. Since the onset of the financial crisis in 2009, we estimate ELA reliance has risen to nearly 90bn in these two countries. See Overview, Global Fixed Income Markets Weekly, 1 April 2011 for more description of ELA. ** Includes items such as gold and FX reserves, claims on non-Euro area residents, and other assets and securities. We subtract the estimated ELA outstanding from this figure. Source: ECB, Central Bank of Ireland, Bank of Greece. Data as of Sep 2011.
The rest of the increase comes from items such as gold reserves (up around 250bn), other securities and assets excluding ELA (up 250bn), and claims on non-euro area residents (up 100bn). Thus, while the ECB has expanded its balance sheet less than its counterparts have, it has arguably taken on a good deal more credit risk in doing so. Given this substantial increase in balance sheet risk, the ECB is reluctant to further expand the size of its
39 See Overview, Global Fixed Income Markets Weekly, 1 April 2011 for more description of ELA.
peripheral sovereign debt purchases. This is confirmed by recent comments made by EU and ECB officials. For example, while some German advisors have begun to speak out in favour of a lender-of-last resort commitment, key German leaders such as Merkel and Schuble, as well as the German Bundesbank, appear to remain firmly opposed (Exhibit A9).
29
Exhibit A9: Indeed, comments from senior officials in Germany and at the ECB suggest that they are not ready to monetise peripheral debt in large quantities just yet
Selection of recent policymaker comments regarding ECB purchases of peripheral sovereign debt
Country Speaker Date 18-Nov Mario Draghi, ECB President Comment Losing credibility can happen quickly -- and history show s that regaining it has huge economic and social costs." Keeping prices stable is the major contribution w e can make in support of sustainable grow th, employ ment creation and financial stability . And w e are making this contribution in full independence. 03-Nov 03-Nov "Our securities market programme has three characteristics: it is temporary ; it is limited; it is justified in restoring the functioning of monetary transmission channels." The [ECB] w ill "not be forced by any body " to buy bonds... it w as pointless to think sov ereign bond y ields could be brought dow n for a protracted period by outside interv ention. The economic costs of any form of monetary financing of public debts and deficits outw eigh its benefits so clearly that it w ill not help to ECB/ Eurosystem Jens Weidman, head of the Bundesbank 12-Nov 12-Nov Jrgen Stark, ECB Ex ecutiv e Board Member Jose Manuel Gonzalez-Paramo, ECB Ex ecutiv e Board Member Angela Merkel, Prime Minister Germany Wolfgang Schuble, German Finance minister 11-Nov 11-Nov 16-Nov 18-Nov 18-Nov stabilize the current situation in any sustainable w ay . The lack of success in containing the crisis does not justify ov erstretching the mandate of the central bank and making it responsible for solv ing the crisis." "We hav e a mandate and w e hav e to stick to our mandate. Fix ing an interest rate for a country is certainly not compatible w ith our mandate... y ou could not argue that this w as not monetary financing." "This w ould v iolate Article 123 of the EU treaty . I cannot see how y ou can ensure the stability of a monetary union by v iolating its legal prov isions." "The ECB w ill nev er be lender of last resort... The ECB can't compensate for deficits in crisis nations." The ECB is last resort lender for banks because that is its function, but not for gov ernments because the unions treaty ex plicitly forbids it." "The w ay w e see the (EU) treaties is that the ECB does not hav e the possibility of solv ing these problems here." If w e did that, the consequence w ould be that for some months w ed hav e a certain quiet. But in the longer term, the financial markets w ould assume that the euro isnt a stable currency , w hether financial markets w ould then think that its not that bad, I doubt it. Thats w hy i dont think its the right solution.
Source: Bloomberg, Reuters, Telegraph, Businessweek
30
Economic Research Global Fixed Income Markets 2012 Outlook November 24, 2011 David MackieAC (44-20) 7325-5040 david.mackie@jpmorgan.com JPMorgan Chase Bank N.A., London Branch
Economics
Sovereign stress is expected to drive Euro area into a mild recession, but there is significant risk that the downturn could be deep The ECB main policy rate is expected to be cut to 0.50% We do not forecast ECB QE, but SMP purchases will push the agenda for a fiscal union The rest of the world escapes recession, but growth will be lackluster Fiscal tightening will be a significant headwind in the US The UK MPC will continue to expand its gilt purchase programme, possibly extending that to other assets EM growth forecasts have been cut in line with the historical relationships between DM and EM
Exhibit 1: Global GDP growth

10 Emerging markets
Realised real GDP growth and J.P. Morgan forecast; %oya
forecast 5 Global
-5 Dev eloped markets -10 2007 2008 2009 2010 2011 2012 2013
moderate pace of growth than we expected earlier: an average quarterly annualised rate of 2.1%ar. Essentially, after the above-potential performance of the first 18 months of the recovery, the global economy is now expected to experience below-potential growth through 2011 and 2012. There are a number of reasons why the global economy is now performing at a sub-par pace. There were clearly two huge unexpected shocks in 1H11: the surge in commodity prices, which dramatically reduced purchasing power for commodity consumers, and the Japanese earthquake and tsunami, which not only dealt a huge blow to the Japanese economy but also disrupted the global manufacturing supply chain. Although the impact of these shocks started to fade in the summer, other drags entered the picture. Essentially, the structural drags have become bigger and are now dominating over the cyclical lift. The debt ceiling debate in the US has led to more fiscal tightening next year and beyond. Meanwhile, the Euro area sovereign crisis has pushed the region back into recession, reflecting the direct effects of fiscal consolidation and the indirect effects of a tightening of financial conditions. The tipping point for the Euro area was the spread of sovereign stress to Italy, which led to a sharp fall in confidence and asset prices, and a significant tightening in bank funding conditions. Even though monetary policy has eased further in both the US and the Euro area, and emerging market growth is holding up relatively well, the structural drags in the US and the Euro area appear to be sufficient to restrain global growth to a sub-par pace through the end of 2012.
The outlook for the global economy in 2012

Since the global recovery began in mid-2009, there has been a fair amount of tension between the forces of cyclical lift (easy monetary policy, low levels of spending on durables, low inventories and healthy nonfinancial corporates) and the structural drags (deleveraging by households, banks and governments, elevated uncertainty about fiscal and regulatory issues, and a decline in underlying growth potential). As the recovery began to unfold, we expected the forces of cyclical lift to dominate and ensure above-trend global growth. Even so, the structural drags were always expected to ensure that the return to full employment would be a long journey. Indeed, this was the experience of the global economy in the first 18 months of the recovery, from mid-2009 to the end of 2010, when GDP growth averaged 3.8%ar. However, 2011 has not followed this script at all (Exhibit 1). It now looks like the pace of global growth in 2011 will be around twothirds of what we expected earlier in the year (a quarterly average growth rate of 2.3%ar compared with our earlier expectations of 3.9%ar). Further, our expectations for the global economy in 2012 show a much more
31
Deleveraging shifts from households to sovereigns and banks
Exhibit 2: US household financial position* and stock of debt outstanding

% of GDP % of GDP
Over the past few years, there has been much focus on household deleveraging. However, as far as the broader economy is concerned, the drag from household deleveraging is now very limited. The sharp move in household financial positions ended some time ago, and we are currently in a situation where households are generating free cash flow and using that to repay debt (Exhibit 2). The big deleveraging stories of this year and the next concern sovereigns and banks. After easing fiscal policy significantly during the crisis, fiscal authorities in the US and Europe are turning towards tightening. Having already begun this year, it is due to get much more intense in 2012 (Exhibit 3). Meanwhile, the sovereign crisis in the Euro area and the policy response encouraging recapitalisation looks to be accelerating the deleveraging process in Euro area banks. It is hard to know exactly where equilibrium configurations of primary positions and debt levels stand, but the sovereign deleveraging is likely to be a long and hard journey, most probably longer and harder than the journey for households. The key difference is that sovereigns hope to take more time to complete this journey, although across the periphery of the Euro area, financial market impatience is accelerating the adjustment with inevitable consequences for growth. Fiscal tightening is a key reason why global growth is expected to be sub-par in 2012, and this basic picture is unlikely to change in 2013 and beyond. Sovereign deleveraging is likely to dominate the macro landscape for years to come.
Euro area slides into recession
8 6 4 2 0 -2 -4 1980
Financial position
Debt stock
100 90 80 70 60 50 40
1985
1990
1995
2000
2005
2010
* US households financial position: ((disposable income - personal outlays)-( gross private investment -consumption of fixed capital))/GDP
Exhibit 3: US fiscal policy changes for 2012

$bn
Current law Payroll tax Other taxes Infrastructure, S&L support Unemployment relief Other spending Total % of GDP -110 -25 -125 -50 -40 -350 -2.3
Net change +65 +45 +15 +12 -40 +97 +0.6
Imapct on GDP (%pt) -1.8 0.5 Note: Obama plan incorporates changes that will not be fully implemented in 2012
Exhibit 4: Change in Euro area composite PMI

6M change in Euro area composite PMI; points
15 10 5 0 -5 -10 -15 1998 2000 2002 2004 2006 2008 2010 2012
Recent data confirm that the Euro area moved into a recession in the autumn. When gauging whether an economy is transitioning into a recession, we look at both changes and levels of key variables. Over the past 6 months, the area-wide composite PMI has fallen by more than during any other 6-month period except in the immediate aftermath of the Lehman bankruptcy (Exhibit 4). Further, the level of the PMI is now well below where it stood at the start of the 2008-09 recession (April 2008) (Exhibit 5). The labour market data also are sending a compelling message of recession. Over the past three months, unemployment has risen by an average of 125,000, well ahead of the average in the three months up to the start of the 2008/09 recession (Exhibit 6).
32
At the moment, our forecast is of a mild recession in the Euro area. The contraction is expected to last through next autumn and to involve a peak-to-trough decline in the level of GDP of just over 1%. In the recessions of the mid-1970s, early 1980s, and early 1990s, the Euro area saw peak-to-trough declines in the level of GDP of 2.5%, 0.5%, and 1.9%, respectively. And in the 2008/09 recession, GDP fell by a staggering 5.5%. Given the relatively unusual nature of the current environment, and the difficulty in gauging the behaviour of policymakers, we would emphasise that there is a lot of uncertainty around our central projection for the Euro area real economy. Given the stress in the financial system, the recession could easily be deeper than our central projection, especially if policymakers do not manage the situation well.
What usually causes a recession?
Exhibit 5: Euro area composite PMI

Index
60 55 50 45 40 35 2007
Lev el at the start of 2008/09 recession: 51.9
2008
2009
2010
2011
2012
Exhibit 6: Euro area unemployment

Monthly change; 3M average; 000s
500 400 300 200 100 0 -100 -200 2007
Lev el at the start of 2008/09 recession: 29
Recessions occur when one or more negative shocks hit a vulnerable economy. These shocks can take a variety of forms, but they involve either a squeeze in current real disposable income (interest rates, commodity prices, fiscal policy) or a shift in the forces that influence current spending relative to current real disposable income (interest rates, asset prices, wealth, credit availability, uncertainty). Meanwhile, an economy becomes vulnerable to such shocks for a number of reasons: households or non-financial corporates have overextended their spending on durable goods and inventory relative to their income (through excessive borrowing, partly driven by elevated asset prices); or banks have over-levered their balance sheets relative to their capital positions (again partly driven by elevated asset prices); or corporate profits have been squeezed to such an extent that drastic action is required to restore profitability. The Euro area recessions of the 1970s, 1980s, and 1990s were driven by monetary policy shocks that hit a private sector with overextended levels of spending and a corporate sector where profitability had been depressed by an overheated economy. The 2008/09 recession too involved a monetary policy shock, but there was also a huge asset-price shock; the overextension was in the financial sector as well as in the non-financial sector. In 2008/09, the classic recessionary adjustment in the nonfinancial private sector was amplified by a severe adjustment in the financial sector.
2008
2009
2010
2011
2012
Exhibit 7: Euro area bank lending standards

80
Net tightening of loans to non-financial corporations and households; %
Net tightening of loans to non-fin.corporates 60 40 20 0 -20 2003

Source: ECB
Net tightening of loans to households 2005 2007 2009 2011
33
What does the current Euro area recession look like?
Exhibit 8: Government budget balance projections

Official* and J.P. Morgan budget balance projections**; % of GDP
Greece Official JPM, no extra aust. JPM, more austerity Official JPM, no extra aust. JPM, more austerity Official JPM, no extra aust. JPM, more austerity Official JPM, no extra aust. JPM, more austerity Official JPM, no extra aust. JPM, more austerity Official JPM, no extra aust. JPM, more austerity 2011 -7.3 -9.1 -9.1 -10.0 -9.6 -9.6 -5.9 -6.8 -6.8 -6.0 -7.1 -7.1 -3.9 -4.3 -4.3 -3.6 -3.8 -3.8 2012 -5.6 -10.3 -8.3 -8.6 -9.3 -8.9 -4.5 -5.9 -5.2 -4.4 -6.6 -5.9 -1.6 -2.8 -2.2 -2.8 -4.1 -3.5 2013 -4.4 -9.0 -6.9 -7.2 -9.3 -8.3 -3.0 -5.4 -4.2 -3.0 -6.0 -4.5 -0.1 -2.1 -1.0 -1.8 -3.6 -2.6 2014 -2.2 -8.2 -6.7 -4.7 -7.9 -6.4 -1.8 -5.3 -3.6 -2.1 -5.9 -4.0 0.2 -2.3 -0.9 -0.8 -3.1 -1.9
At first blush, the Euro area recession that we are sliding into now looks more like the 2008/09 experience. Instead of a monetary policy shock, the region is experiencing a confidence shock confidence in sovereign solvency which is being amplified by a still-vulnerable financial system. As asset prices have fallen, bank funding stress has risen dramatically. Although banks have access to unlimited funding from the central bank, funding stress does lead to higher retail borrowing rates, tighter lending standards (Exhibit 7), and a desire on the part of banks to de-lever their balance sheets. This suggests that the Euro area will experience something of a credit crunch in the coming quarters. As we saw in 2008/09, tight credit availability can take a significant toll on the real economy. However, a key difference between now and early 2008 is that levels of spending on durable goods and inventory are currently low, and financial positions in both the household and corporate sectors are currently good. There is still a deleveraging process under way in some of the peripheral economies as households generate free cash flow to reduce the size of their balance sheets but conditions are not nearly as stretched as they were at the start of the 2008/09 recession. Moreover, the significant fiscal consolidations are concentrated in the peripheral economies, and it is hard for the Euro area to have a deep recession without a significant fall in Germany and France. Moreover, the global backdrop looks better now than it did in 2008/09: what we are experiencing at the moment is an asymmetric Euro area shock, rather than a common global shock. The differences between now and early 2008 and the uneven spread of fiscal consolidation are why we think the present recession will be relatively mild. The distinguishing feature of recessions tends to be depth rather than duration. The 2008/09 recession in the Euro area was almost three times deeper than the early 1990s recession, but it was only one-quarter longer in duration. This suggests that policy plays a key role in arresting the negative dynamic in recessions. However, in contrast to the experience of previous recessions, when policymakers were adding stimulus to the economy, the policy action that we anticipate in the coming months asset purchases by the EFSF and the ECB, loans to sovereigns and bank recapitalisation is more about crisis containment. This leads to the question of how effective policy will be in containing the negative
34
Ireland
Portugal
Spain
Italy
Belgium
* Official projections are from the Stability and Growth Programmes for Greece, Ireland, Spain and Belgium (the Greek one was published in August rather than in the spring). Portuguese projections are from their fiscal strategy and Budget 2012 documents, published in October. Italian projections are from their updated strategy document published in September. ** JPM, no extra austerity assumes that governments do not correct the cyclical fiscal slippage due to growth undershooting the official forecasts JPM, more austerity assumes additional fiscal tightening worth 50% of the cyclical slippage
dynamic in the real economy and creating a base for the ultimate recovery. All of this indicates that it is difficult to have a lot of conviction about what the present Euro area recession will look like. The conditions are unusual and it is unclear how confidence, asset prices, and bank funding stress will interact, and how that will impact the real economy. There is also uncertainty about policy. This is not simply about whether policymakers will respond in a timely and effective manner. It is also about the limited scope for the addition of outright stimulus. Our inclination is to think that downside risks prevail around our central projection.
Recession makes fiscal and structural adjustments harder to achieve
As the Euro area slides into a recession that looks likely to be particularly deep in parts of the periphery, budgetary positions will slip due to the automatic stabilisers. Given the need for peripheral sovereigns to keep moving towards more sustainable fiscal positions, there will be enormous pressure to offset some of the budgetary slippage with additional tightening measures, even though these will weigh additionally on demand. Our forecast does incorporate some additional fiscal
tightening next year, beyond the plans already announced. Essentially, we assume that governments seek to offset half of the cyclical slippage (Exhibit 8). In addition to putting significant pressure on individual sovereigns, the recession dynamic is also affecting the rescue mechanism put in place by the region. As Frances AAA rating looks less secure, due to both the discussion about leveraging the EFSF and the evidence that France itself may be falling into a meaningful recession, the EFSFs own situation looks less robust.
The response of the ECB to the macro landscape
Exhibit 9: Italian debt-to-GDP under different scenarios

% of GDP*
140 Mild shock case 130 120 110 100 90 2000 2005 2010 2015 2020 J.P. Morgan baseline Official plan
In early November, the ECB cut the main policy rate from 1.50% to 1.25%. Although this move was not widely expected, it is not hard to understand given the deterioration in the macro landscape over the past few months. Indeed, the new ECB President, Mario Draghi, spoke of the region heading toward a mild recession, which would exert downward pressure on wage and price inflation. We expect another rate cut to follow in December, and for the main policy rate to ultimately reach 0.50%. If the recession turns out to be a deep one, there will be talk about whether the ECB should engage in quantitative easing, which we would interpret as a targeted amount of bond purchases spread across the region in order to ease the monetary stance further by lowering longer-term interest rates. We would contrast this with the Securities Markets Programme (SMP), which is intended to improve the transmission of the prevailing monetary stance rather than add more stimulus. At the moment, we do not anticipate the ECB engaging in QE, although the central banks balance sheet will still expand further due to SMP purchases.
The response of the ECB to sovereign stress
* See Exhibit 10 for assumptions under different scenarios
Exhibit 10: Key variables for Italian debt dynamics
Evolution of macro variables under alternative scenarios*; % of GDP, unless otherwise specified
Official plan Key assumptions Marginal borr. rate (from now), % 5.8 Real long run growth pot., % 1.2 Maximum sustained prim. balance 5.7 Key variable averages over the period 2012-20 6.5 0.8 5.0 6.5 0.6 3.5 J.P. Morgan baseline Mild shock case
Average borrowing rate, % 5.0 5.4 5.4 Nominal GDP growth, % 3.0 2.0 1.7 Government debt 104.8 120.8 130.2 Primary balance 5.4 4.8 3.3 Yearly change in debt -3.4 -0.7 1.4 * Official projections are partly the result of our assumptions, as they are not available beyond 2014.
As the stress in the sovereign bond markets has spread to core countries such as France and Austria, it has become clear that something more dramatic from policymakers is needed to stabilise markets and ensure financial stability across the region. Not surprisingly, all eyes are on the ECB as the only institution able to step in quickly and decisively. The obvious next step would be for the ECB to announce its intention to use the SMP aggressively to ensure financial stability in the region. In our view, there is no legal or technical limit to the amount of explicit support that the ECB can give secondary markets and the amount of implicit support that the central
bank can give to primary markets (by purchasing around the primary auctions to help dealers get short and shed unwanted positions, but not at the auction itself). This more-aggressive stance could be adopted with explicit yield targets, although we doubt that the ECB will go that far. Whether this stabilises the situation remains unclear. What lies at the heart of the crisis is a fear that the process of sovereigns returning to solvency via austerity alone, i.e. without local currency depreciation and a locally orientated monetary stance, is dynamically unstable. Adding to the prospect of a dynamically unstable process is the widely held perception that peripheral economies are profoundly uncompetitive and that some of the structural reforms needed to improve competitiveness will depress growth and damage social cohesion before they deliver any macroeconomic
35
benefits. This fear of dynamic instability has increased with the growing evidence that the region is sliding into recession, possibly a deep one. Greece provides a clear example of how badly things can turn amid political and social instability and sovereign debt restructuring. At some point, sizable ECB interventions would stabilize the system. But, the situation for Italy remains a key uncertainty (Exhibit 9 and 10). In our view, the SMP is not an appropriate mechanism for funding Italys gross financing need over the medium term. Thus, the region does need to create some kind of liquidity hospital to accommodate Italy. The EFSF as currently constructed cannot do this. Thus, it is not surprising that France has put the idea of turning the EFSF into a bank back onto the agenda and that there are ongoing discussions about how to involve the IMF more substantially. Alongside increased ECB intervention and the creation of a liquidity hospital for Italy, the Germans will continue their campaign for further fiscal integration. As this crisis has continued to unfold, it has become increasingly evident that the current blueprint for the future more peer surveillance and pressure but no de jure loss of fiscal sovereignty, and with a permanent liquidity hospital based on the idea of shared fiscal capacity is unlikely to work, unless sovereign debt levels are much lower. The risk of sovereign liquidity crises is just too great. But, in terms of dealing with this fundamental issue in the functioning of the Euro area, policymakers are moving incrementally. In the near term, the Germans want limited treaty change to introduce automatic sanctions in the workings of the stability and growth pact, the possibility of referring sovereigns to the European Court of Justice, and the possibility of greater interventions in national budgets. It seems that they continue to want an orderly sovereign default mechanism as part of the permanent crisis resolution framework. It is not clear, however, that this will be sufficient.
Impact of the Euro area recession on the rest of the world
Exhibit 11: US labour market

3M moving average of monthly change in non-farm payrolls and unemployment rate for 25 years and older % %
0.5
Unemployment rate
9 8
0.0
Montlhly pay rolls
7 6 5 4
-0.5
-1.0 2006 2007 2008 2009 2010 2011 2012
surely impact demand from that market. If that were the only transmission channel, the risks would appear to be contained. However, the potential financial spill-overs could be larger. Thus far, most measures show only a modest impact of the European crisis on domestic credit availability. Nonetheless, if conditions in Europe worsen, the US economy would not be immune from a seizing up of global financial markets. The second risk is a tightening of domestic fiscal policy. The recession produced an unprecedented fiscal policy response as the Federal deficit increased to 10% of GDP. The political tolerance for such large deficits has lessened recently, and several temporary fiscal stimulus measures are set to fade in 2012, which could subtract 1%-2%-points from GDP growth next year. There has been some talk about extending the temporary measures, but nothing has been agreed upon so far. The anticipated drag from fiscal policy is the main reason behind our view that growth will be below potential in 2012, when we anticipate US growth will average 1.75%. This expected tightening of fiscal policy is likely to hit the economy where it is weakest: the household sector. Slack conditions in labour markets have eroded labours bargaining position, resulting in very tepid wage gains (Exhibit 11). This outcome has been positive for corporate profits which have received a significant lift from restrained unit labour costs but has held back consumers purchasing power.
In the middle of 2011, there was a serious growth scare in the US, but the economy has settled at a close-topotential growth rate in the second half of the year. Even so, two risks loom over the 2012 outlook for the US. The first risk relates to spill-overs from the economic and financial crisis in Europe. Around 20% of US exports are destined for Europe, and the economic contraction will
36
The low rate of inflation has given the Federal Reserve leeway to be creative in supporting the recovery. The two experiments this year were the explicit mid-2013 funds rate guidance, and Operation Twist. We think the most likely next step for the Fed is to replace the mid-2013 rate guidance with language that is more explicitly dependent on a limited number of observable economic variables, particularly the unemployment rate. While we are not forecasting more asset purchases in 2012, the risk of this is high and rising (subjectively, we would put it at around 40%). Moreover, Fed officials have shown an increasing openness to considering returning to MBS purchases. A significant setback in the Euro area, or an unwelcome fall in inflation expectations, could be all that is required for QE3 in the US. The UK has conspicuously chosen to front-load its fiscal adjustment. While that has shielded it from direct attack on the gilt market, public sector job losses, tax increases, and spending restraint have weighed on growth. Through noisy data, growth has run at a near-1% pace since mid2010. As we look forward, fiscal consolidation is ongoing, while confidence and bank funding conditions are caught in the back wash of Euro area turmoil. Leaning against this is the fact that the direct drags on household purchasing power from rising VAT, energy, and food bills are dissipating. And as inflation begins to ease, the MPC will continue to expand its gilt purchase programme, plausibly extending that to other assets (such as bank term debt) as 2012 progresses. All told, the forecast shows the UK seeing near stagnation in output through what will remain volatile GDP data, with unemployment continuing to rise. Given expected sluggish developed market growth, the emerging market outlook has also markedly weakened. The drag from DM to EM operates through two channels. The first is through trade linkages. A slowing in EM growth from weaker exports to DM is unavoidable. We estimate that EM export growth will slow to just 4% in 1H12, which is weaker than at any time during the 2000s expansion. The second channel of transmission is through financial markets. As risk appetite wanes, capital flows to the EM slow or even reverse, equity markets sell off, credit conditions tighten, currencies fall, and the prices of commodities a source of revenue for many EM countries decline. In turn, domestic demand in the EM is hit. This time around, the financial transmission channel could be magnified by the rapid rise in credit
since the start of the recovery in 2009. While domestic funding sources have increased, deleveraging of European banks could still pose a risk as their foreign claims in EM are sizable particularly in EMEA and Latin America. The interaction of these complex forces has generated a unit-beta response in the EM to shifts in US/Euro area growth over the past decade. In other words, a 1%-pt decline in US/Euro area growth has translated, on average, into a 1%-pt decline in EM growth. EM Asia has a smaller beta of 0.6 (owing to very low betas in China and India), Latin America has a beta of 1.2, and EMEA has the highest beta of 1.4. Currently, the channels of transmission are operating as usual. Sluggish DM import demand has produced a sharp slowdown in EM export growth. Risky assets have sold off across the world and commodity prices have fallen. In turn, economic activity has decelerated in most EM countries. J.P. Morgans recently lowered GDP growth forecasts for the next five quarters align well with the implied beta relationships for the three EM regions, with the average projected at a sub-trend pace of 5% in 2012. Should the US/Euro area slide into a deeper downturn than currently forecast, the greater flexibility of policy in the EM could lead to a relative outperformance in the region. This outcome, however, would require a change in EM policymakers reaction function. On the downside, increased leverage in the region could exacerbate financial stresses, leading to a worse-than-unit beta outcome.
37
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Gianluca SalfordAC (44-20) 7325-4334 Aditya Chordia (44-20) 7777-9841 J.P. Morgan Securities Ltd
Euro Cash
2011 saw a further escalation of the Euro area sovereign crisis as stress reached Italy and core countries In our central scenario the crisis will worsen in the first part of 2012 on the back of Euro-wide recession and further market pressure on sovereigns forcing policymakers to react aggressively to stabilise the situation in the second part of the year The ECB will be forced to cut the refi rate to 0.50% and expand its balance sheet even further Although our fair value model suggests that 10Y Bund yields could drop below 1.00%, as the PMIs/core inflation drop and peripheral spreads widen, we forecast a floor in 10Y Bunds at 1.25% We recommend long 10Y duration exposure; but this view is predicated on policy makers being able to ultimately contain the crisis; if they fail Germanys role as a safe haven will be challenged The short end of the curve has little scope to rally under our assumption that the depo rate will not fall below 0.25%: we therefore recommend 2s/10s flatteners targeting around 100bp by 1H12 With regard to intra-EMU spreads, we forecast generalized widening vs. Germany We also recommend credit curve flatteners: in line with the experience of Greece, Ireland and Portugal, investors will switch from focussing on yields to prices in case of severe sovereign stress The empirical price relationship between short and long dated bonds is convex: we recommend flatteners weighted by the empirical beta to get protection from tighter spreads and steeper curves We present a framework to analyse the yield spread of high- and low-coupon bonds of similar maturity under sovereign stress: Bonos 2032 are rich vs. Bonos 2037 We analyse the investor base of Euro area, German, French, Italian and Spanish bonds. The reliance on foreign banks and foreign central banks is high for all of them, except Spain
Exhibit 1: 2011 recap: markets and data peaked in the spring and have been on a downward trend ever since
10Y Bund yield and Euro area composite PMI % index
3.75 3.50 3.25 3.00 2.75 2.50 2.25 2.00 1.75 1.50 Jan-11 Mar-11 May -11 Jul-11 10Y Bund
PMI
60 58 56 54 52 50 48 46
Sep-11
Nov -11
On the supply side, higher redemptions will be generally offset by lower deficits. We expect 710bn of conventional bond issuance in 2012, slightly below 2011 levels Italy faces heavy bond redemptions in 2012, especially in the February-April period, whereas Spains redemption calendar is more manageable
2011 recap: Shattered hopes

It is going to get worse before it gets better was investors mantra into 2011, a view that we shared. The beginning of the year caught many investors on the wrong foot as policymakers seemed to calm fears with little effort and macro data continually surprised to the upside, leading to stronger equities and higher yields, with 10Y bund yields peaking at 3.50% in midApril (Exhibit 1). In addition, higher inflation prints prompted the ECB to surprise market participants and increase interest rates twice after being on hold for almost two years. Declining excess cash in the system added to the perception of normalisation. It was not meant to last. The sovereign crisis raised its ugly head again in midApril. Portugals request for aid was met with a yawn, but once again Greece was the catalyst: amid clear signs of fiscal slippage, the request for private sector involvement in exchange for further official support broke a taboo and hit investors confidence hard. In less than 6 months, Bund yields halved on the back of
38
Exhibit 2: The curve directionality to the level of the market shifted from the short and medium end to the ultra-long end in the second part of the year
2s/5s, 5s/10s, and 10s/30s regressed against 2Y German government par rates; 1 Jan 2011 to 11 Apr 2011 and 12 Apr 2011 to 18 Nov 2011; 2s/5s 5s/10s 10s/30s Jan-April -0.27 -0.30 -0.01 Beta Apr-Nov 0.09 -0.14 -0.19 Jan-April 75% 93% 1% R-squared Apr-Nov 42% 65% 68% * Peak in 10Y Bund yields.
Exhibit 3: Sovereign risk escalated to yet new highs in 2011

350% 300% 250% 200% 150% 100% 50% 0% -50% Jan-11 Mar-11 May -11 Jul-11 Sep-11 France Greece Italy
YTD % change in 10Y spreads to Germany for France, Greece, and Italy; %
pressure from peripherals, a collapse in the PMIs and the August downgrade of the US by S&P. Since October, yields have been range-bound and hostage to headline news related to the sovereign crisis. We are ending the year with 2s/5s significantly flatter YTD, 5s/10s almost unchanged and 10s/30s steeper. The 2s/5s and 5s/10s curves exhibited typical directionality in the first part of the year, flattening in the bearish move, but they lost directionality in the second part. On the contrary, the 10s/30s Bund curve showed poor directionality in the first part of the year refusing to flatten further from already stretched levels, but bullsteepened with decent directionality in the second half of the year (Exhibit 2). Sovereign risk escalated to yet new highs in 2011. In the first part of the year, the problem remained isolated to the smaller peripheral countries (Greece, Ireland, and Portugal). However, policy makers decision to include private sector involvement (PSI) in the new package for Greece spooked investors, spreading contagion to the larger peripheral countries (Italy and Spain) and beyond. It is interesting to highlight that in relative terms, France or supranational issuers such as the EFSF underperformed more than Italy and Spain did in the latter part of the year (Exhibit 3), making the distinction between core and peripheral debt markets somewhat obsolete. Although all intra-EMU spreads to Germany reached their all-time highs since the Euro inception, France and Italy were the worst performers in relative terms, whereas Ireland was the best performer (Exhibit 4). Rating agencies were very busy with downgrades (Exhibit 5), adding to the pressure. Once again, policy makers were forced to take unprecedented actions: 1) ECB purchases: In August, the ECB was forced to intervene by buying Italian and Spanish paper through the SMP after a 4 month hiatus (Exhibit 6). We estimate that the ECB has purchased
Nov -11
Exhibit 4: Intra-EMU spreads widened aggressively after mid-April 2011

10Y benchmark spreads to Germany; bp
% change from 18-Nov-11 Austria Belgium Finland France Netherlands Greece Ireland Italy Portugal Spain 147 282 67 147 56 2425 625 494 910 469 11-Apr-11* 30-Dec-10 31 72 25 30 22 933 561 125 521 170 46 104 28 42 24 948 611 186 374 248 30-Dec to 11-Apr to 30-Dec to Current Current 11-Apr 216% 170% 138% 249% 131% 156% 2% 166% 144% 89% 372% 291% 168% 394% 152% 160% 11% 297% 75% 177% -33% -31% -11% -29% -8% -2% -8% -33% 39% -32% 2011 high 189 309 75 188 65 2471 1142 575 1161 486 2011 low 31 72 24 30 19 747 529 125 349 170
* Peak in 10Y Bund yields.
Exhibit 5: Rating agencies had a busy year cutting peripheral sovereign debt ratings further
Average* long-term local currency debt rating for peripheral countries at the end of 2009, 2010 and 18 Nov 2011;
2009 CCC B BB BBB A AA AAA Greece Ireland Italy
2010
2011
Portugal
Spain
* Average of Moodys, S&P and Fitch. AAA/Aaa=1, AA+/Aa1=2, etc.
39
more than 120bn worth of Italian and Spanish paper since it restarted the SMP (Exhibit 7); 2) Expansion of EFSF and ESM lending capacity to 440bn and 500bn, respectively, with media speculation of combined fire power1 and still pending projects to leverage the EFSF firepower through different mechanisms;2 3) A second bailout package for Greece with PSI in July, which is being renegotiated; and 4) After a mild stress test in the summer, authorities pushed for a 100bn plan to recapitalise banks in October. What has been the impact of this flurry of announcements? Since the beginning of the crisis, policymakers promises of future action and surprise ECB purchases have had a large impact both on German yields and on peripheral spreads, but typically the impact has been short-lived as evidenced by sovereign spreads trading close to their widest levels ever (Exhibit 8).
Exhibit 6: The ECB started to buy Italian and Spanish bonds on 8 August 2011 and currently holds almost 200bn of peripheral country debt in the SMP on a cash basis
Weekly and cumulative ECB bond purchases through SMP*; official purchases data on a cash basis; bn
200 175 150 125 100 75 50 25 0 May -10 Aug-10 Nov -10 Mar-11 Jun-11 Sep-11 10 5 0 Weekly bond purchases (rhs) Cumulativ e bond purchases (lhs) 20 15 25
* Excludes roughly 6.7bn of matured bonds. Source: ECB
2012: Flirting with danger

We expect that the outlook for the Euro area fixed income market will be dominated by a combination of 1) macro developments, and 2) evolution of the sovereign crisis and policy making response. On the macro front, our economists are more negative on growth than either the latest set of European Commission forecasts or consensus, forecasting an area wide recession that only Germany will avoid. Even though we are far from forecasting deflation, we believe that in 2012 the ECB will be forced to do all it takes to support the region, cutting the refi to 0.50% (Exhibit 9) in quarterly 25bp steps. As discussed at length in the Overview, we fear that, as was the case in 2011, in the first part of 2012 domestic political constraints will ensure that policymakers will continue to be reactive rather than proactive. Italy and Spain may lose access to the primary market, in which case our view is that policy makers will realise the mistake, step up the effort and manage to contain the
Exhibit 7: We estimate that the ECBs SMP purchases account for roughly 20% of the bond market in Greece, Ireland, and Portugal, and 7% in Italy and Spain
J.P.Morgan estimate of notional amounts of peripheral bonds purchased through the ECB SMP* and their current bond market size; bn Greece Ireland Portugal Italy Spain Notional bought** 50 18 20 95 32 Bond market size 247 85 104 1359 474 % 20% 21% 19% 7% 7% * We estimate notional amounts using official total weekly purchases, cash prices and our estimate of country and maturity split. ** Includes roughly 6.7bn of matured bonds, mostly Greek ones. Source: J.P.Morgan, ECB
Exhibit 8: Although policymakers announcements had a positive impact on markets, the effect was short-lived, as evidenced by sovereign spreads currently trading close to their widest levels ever
Behaviour of the 10Y Bund and 10Y weighted peripheral spread* after significant policy announcements or actions; detailed announcements in grey
10Y Bund Date Event Days of Peak sell Days of sell off off (bp) tightening 0 7 4 20+ 20+ 15 1 0 2 0 8 2 20+ n/a 16 6 35 42 24 2 n/a 11 n/a 46 20 46 0 20+ 12 0 20+ 14 1 0 6 19 4 2 20+ 10Y wtd. peri. spreads* Peak % tightening n/a -46% -8% n/a -27% -8% -1% n/a -12% -17% -6% -5% -46%
03-May-10 First bailout plan for Greece 10-May-10 EFSF announcement, SMP 23-Jul-10 EBA stress test 28-Nov-10 Irish bailout 10-Jan-11 Discussion of comprehensive plan 25-Mar-11 Comprehensive plan announcement 04-May-11 Portuguese bailout 15-Jul-11 21-Jul-11 EBA second stress test EFSF 2, second Greek bailout
We think it is unlikely to have the EFSF and the ESM in place together (with combined capacity) as the German Supreme Court ruled that Germanys exposure should be limited and that the German government must now obtain the approval of the Parliamentary Budget committee before giving any guarantees. See Overview, Global Fixed Income Markets Weekly, 9 September 2011. 2 See Overview,Global Fixed Income Markets Weekly, 14 October 2011 and 28 October 2011.
40
08-Aug-11 SMP 2 04-Oct-11 Discussion of comprehensive plan 2 26-Oct-11 Announcement of comprehensive plan 2 Max
* 10Y weighted peripheral spread computed against Germany for Greece, Ireland, Italy, Portugal, and Spain (weighted by the size of their outstanding bond market).
Exhibit 9: Our economists forecast a mild Euro area recession, declining inflation and official rates troughing at 0.50%
Forecasts for 2012 real GDP growth, HICP inflation, and end-of-2012 ECB refi rate; %
Exhibit 10: We forecast a trough in 10Y Bund yields at 1.25%
J.P.Morgan interest rate forecast and spread vs. forwards; German benchmarks unless otherwise stated; %
18-Nov -11 ECB refi 1.25 0.65 0.46 1.10 1.97 2.61 151 64 1Q12 0.75 0.38 0.35 0.75 1.55 2.15 120 60 2Q12 0.50 0.32 0.30 0.70 1.25 1.95 95 70 3Q12 0.50 0.32 0.40 0.85 1.50 2.25 110 75 4Q12 0.50 0.32 0.50 1.05 1.75 2.50 125 75 2Q12 v s. 4Q12 v s. fw d (bp) n/a -13 -37 -69 -86 -70 -49 16 fw d (bp) n/a -30 -44 -57 -48 -18 -4 30
2.00 1.50 1.00 0.50 0.00 -0.50 -1.00
J.P. Morgan
Consensus
EC
1M EONIA 2Y 5Y 10Y 30Y 2s/10s (bp) 10s/30s (bp)
n/a
Exhibit 11: Intra-EMU spreads will reach new highs by mid-2012

Forecast of 10Y curve-adjusted spread to Germany; bp
Grow th
Inflation
ECB refi
Germany Austria Belgium Finland France Greece Ireland Italy Netherlands Portugal Spain Weighted spread* Weighted peripheral spread**
Spread to Germany (bp) 18-Nov -11 147 282 67 147 2425 625 494 56 910 469 461 699 Q2 200 400 100 225 2400 850 775 85 1200 700 625 955 Q4 175 350 80 190 2400 900 625 70 1300 600 555 855 1.97 3.50 4.80 2.61 3.46 26.32 8.15 6.95 2.52 11.04 6.70 -
Yield (%) 18-Nov -11 Q2 1.25 3.25 5.25 2.25 3.50 25.25 9.75 9.00 2.10 13.25 8.25 Q4 1.75 3.50 5.25 2.55 3.65 25.75 10.75 8.00 2.45 14.75 7.75 -
Source: J.P.Morgan, European Commission, Consensus Economics, Reuters
crisis. If everything plays out as expected, 2H12 should see a marginal improvement in financial conditions. Based on these views, we recommend investors: 1) Go long duration. All the drivers of our 10Y Bund model point to lower yields in 2012 (see below). We target 10Y Bunds yields at 1.25% (Exhibit 10). 2) Enter 2s/10s flatteners. We see limited scope for much lower short-term yields and recommend flattening exposure in 2s/10s, targeting around 100bp by mid-2012. 3) Position for further sovereign stress, underweighting all intra-EMU countries vs. Germany. We expect the credit curve to flatten with wider spreads, but we favour weighting trades for the empirical convexity of the price curve to minimise risk. Our forecast for intra-EMU spreads is shown in Exhibit 11. Short-end yields: Not far from the bottom Even with our downbeat view, we struggle to find value at the short end of the German curve. Shortterm German yields are close to historical lows, already pricing in 1) an ECB deposit rate at 0.25% for a prolonged period of time, 2) extra liquidity in the system to be abundant (Exhibit 12), and 3) a significant premium for the perceived safety of German government bonds. It is unlikely that German 2Y yields will move much in either direction in the first part of the year if our macro forecast of declining GDP for four consecutive quarters
* Weighted spread computed against Germany for the 10 largest Euro area countries (weighted by the size of their outstanding bond market). ** Weighted peripheral spread computed against Germany for Greece, Ireland, Italy, Portugal and Spain (weighted by the size of their outstanding bond market).
Exhibit 12: High liquidity and significant premium for German bonds leave limited scope for a further decline in short-dated German yields
Average monthly seasonally adjusted excess liquidity* and 3M German GC-3M EONIA OIS; monthly data bn bp
200
3M GC-OIS
-10
150
-15
100 Ex cess liquidity ; sa 50
-20
-25
0 Jan-11 Apr-11 Jul-11 Oct-11
-30
* Seasonally adjusted excess liquidity = OMOs Autonomous factors Reserve requirements.
41
plays out. However, an improvement in 2H12 should push yields higher as market participants might start to see some light at the end of the tunnel. We forecast 2Y yields at 30-35bp until 2Q12 and at 50bp by the end of 2012. The 50bp year-end yield target for 2Y is consistent with our expectations of 3M GC rates around 10bp and a 2Y benchmark/3M GC rate spread in line with average since the beginning of 2010 of around 40bp (Exhibit 13). 10Y Bunds to drop to record low: Go long In 2011, quite a few investors complained that low German yields did not reflect fundamentals in their view. Indeed, a decelerating but still-growing economy and core inflation not far from the ECB target are difficult to reconcile with 10Y Bund yields almost 200bp lower than the 10-year average. We find that incorporating intraEMU spreads in a macro model explains the dramatic decline in Bund yields over the past 2 years quite well (Exhibits 14 and 15). Looking ahead, all the explanatory variables point to falling 10Y Bund yields over the next 6-12 months. On the macro side, to be consistent with our growth forecast, the composite PMI should fall further by 1H12, before rebounding above 50 by the end of the year, whereas core inflation should decline gently in 1H12, and more aggressively in 2H12. We expect peripheral spreads to widen further in 1H12 before retracing a bit. Plugging the inputs into the model would give a fair value for the 10Y Bund around 1.00% by the middle of the year (Exhibit 16). We expect non-linearities to eventually emerge, preventing the Bund from reaching such an extreme level, but 10Y Bund yields at 1.25% certainly seems possible. As discussed in detail in the demand section below, we believe that Germanys position as safe haven is at risk in case of an escalation of the crisis if policy makers are not able to contain the fallout from the crisis. Given our views on the short and long end of the curve, we recommend 2s/10s flatteners, targeting around 100bp mid-year. However, we have repeatedly found that bouts of risk appetite can hurt flatteners, as was the case in 4Q10 and in September-October 2011. When the market is pricing in low-for-long, any short-term improvement in sentiment is felt first at the long end of the curve, creating a convex behaviour as the 2Y rate catches up with a lag. Investors who seek protection from bearish steepening should consider 1s/5s conditional bull flatteners in swaps (see Euro Derivatives).
Exhibit 13: Under the assumption of 3M GC repo at 10bp, we forecast 2Y rates at around 50bp by the end of 2012, 40bp above 3M GC repo, in line with the average of past two years
Evolution of 2Y German par govt rate minus 3M general collateral repo rate; %
1.00 0.80 0.60 0.40 0.20 0.00 -0.20 -0.40 Jan10 Apr10 Jul10 Oct10 Jan11 Apr11 Jul11 Oct11 Av erage: 41bp
Exhibit 14: In our fair value model, 10Y Bunds yields are explained by a combination of macro variables and sovereign stress
10Y Bund yields regressed against composite PMI, HICP ex food and energy and 10Y weighted peripheral spread*; daily interpolated data over past 10Y; % Beta T-stat Intercept 0.716 9.3 Composite PMI (units) 0.031 25.0 Core inflation (%oya) 0.901 70.1 10Y weighted peripheral spread (bp) -0.003 -47.7 R-squared 83% Standard error 0.28 * Weighted peripheral spread computed against Germany for Greece, Ireland, Italy, Portugal and Spain (weighted by the size of their outstanding bond market). Note: In the presence of the above factors, addition of front-end yields as another explanatory variable does not materially improve the goodness of the fit.
Exhibit 15: with intra-EMU spread widening accounting for around 175bp of decline in 10Y Bund yields over the past two years
Marginal contribution to 10Y Bund yields*: deviation from average; %
1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 PMI Inflation 10Y w td. peripheral spread
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
* See Exhibit 14 for 10Y Bund yield model details.
42
Exhibit 16: Under our baseline scenario, a further drop in the PMI, lower core inflation and wider peripheral spreads should pressure Bund yields below 1.00%; however, given non-linearities, we expect Bund yields to bottom around 1.25% in 1H12
10Y Bund fair value yield based on the model described in Exhibit 14 and J.P.Morgan forecasts for the explanatory variables; 1H12 Level Beta Composite PMI (units) 43.75 0.031 Core inflation (%oya) 1.50 0.901 10Y weighted peripheral spread (bp) 955 -0.003 Fair value 10Y Bund yield (%) 0.85
Exhibit 17: We do not recommend buying 5Y in 2s/5s/10s fly as a bullish duration trade because the directionality can change when yields drop substantially
6M beta of 50:50 2s/5s/10s vs. 5Y JGB yield; 1 Jan 2001 18 Nov 2011;
0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 0.00
Given our forecast for record-low yields, we look at the example of Japan to analyse the behaviour of the 2s/5s/10s fly under extreme conditions. Typically, in a scenario of unchanged monetary policy stance, the fly is fairly directional, allowing investors to choose between outright longs or barbell trades based on RV and relative carry and slide considerations. However, even with unchanged monetary policy, when interest rates drop significantly the directionality of the 2s/5s/10s can flip, with the 5Y underperforming a combination of 2Y and 10Y amid falling yields; this was the case in Japan (Exhibit 17), and for a few days in the Euro area during mid-September. We therefore do not recommend buying 5Y in a 50:50 2s/5s/10s fly, despite our bullish outlook. At the ultra-long end of the curve, we anecdotally witnessed a reduction in client interest in 10Y+ core bonds, with an impact on relative liquidity compared with shorter segments of the curve. This forced primary dealers to increasingly use the Buxl futures as a hedging instrument (Exhibit 18). We recommend neutral exposure on 10s/30s curve. Finally, on RV, we highlight that the 5s/10s curve is trading close to the cheapest level of the past 10 years vs. 2s/30s (Exhibit 19). We recommend 5s/10s flatteners against weighted 2s/30s steepeners.
0.50
1.00 5Y JGB yield; %
1.50
2.00
Exhibit 18: Buxl futures volumes have increased amid declining liquidity in the cash market
3M rolling average of Bund and Buxl futures volumes; 000s of contracts
1,600 1,400 1,200 1,000 800 600 400 2006 2007 2008 2009 2010 2011 Bund Bux l
9 8 7 6 5 4 3 2
Exhibit 19: 5s/10s remains close to the cheapest level of the past 10 years vs. 2s/30s
Residual of 5s/10s regressed against 2s/30s*; German government par rates; past 10 years; bp
20 15
Current: 15bp
Intra-EMU spreads: Wider in 1H12

Given the view highlighted in the Overview and above, we believe investors should underweight all intraEMU countries vs. Germany in the first part of the year. However, stronger policy action should promote a mild reversal in most countries in 2H12 (Exhibit 11). To refine our views, we analyse the bond market along two axes: 1) Country selection, and 2) Maturity/coupon selection.
10 5 0 -5 -10 -15 -20 2001 2003 2005 2007 2009 2011
* 5s/10s = 0.35*2s/30s 2.94; R-squared: 96%
43
In our view, country selection hinges less than in previous years on fundamental, medium-term analysis and more on technical factors, due to the progressive deterioration of market liquidity. We define technical factors as events that change market dynamics in a non-linear fashion, such as rating downgrades and loss of market access among the negatives, and explicit or implicit forms of market support on the positive side. We expect France to continue to underperform in relative terms on the back of rating-sensitive selling by foreign accounts. The importance of the rating is not confined to France. If Italy loses market access and gets downgraded to BBB space, as stated by rating agencies under such an occurrence, there is likely to be further pressure on the inflation-linked segment due to potential index-inclusion rules (see Inflation-linked Markets section), in addition to further erosion of its investor base. As discussed in the Overview, we expect any request for support to be market-negative: in Greece, Ireland, and Portugal, selling pressure accelerated after bailouts (see Exhibit 30 in the
Exhibit 20: Sovereign risk index: nobody is perfect
J.P.Morgan sovereign risk index* and its components; lower-than-average risk in grey;
2012 debt/GDP (%) 2012 deficit/ GDP (%) Total private debt/GDP 2012 current account balance (% of GDP) Data Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain 73 99 52 89 81 198 118 121 65 111 74 -0.6 0.0 -1.2 -0.2 -0.4 2.5 0.5 0.6 -0.8 0.3 -0.6 -3.1 -4.6 -0.7 -5.3 -1.0 -7.0 -8.6 -2.3 -3.1 -4.5 -5.9 -0.4 0.2 -1.4 0.5 -1.3 1.1 1.8 -0.8 -0.4 0.1 0.7 150 131 153 138 111 122 341 126 219 223 211 -0.4 -0.6 -0.3 -0.5 -0.9 -0.8 2.4 -0.7 0.6 0.7 0.5 2.8 2.1 0.0 -3.3 4.4 -7.9 1.5 -3.0 7.0 -5.0 -3.0 -0.7 -0.6 -0.1 0.7 -1.1 1.7 -0.4 0.6 -1.7 1.0 0.6 5.14 5.20 5.47 5.14 5.41 3.92 4.77 4.43 5.41 4.40 4.54 -0.5 -0.6 -1.1 -0.5 -1.0 1.9 0.2 0.9 -1.0 1.0 0.7 2 7 2 2 3 4 2 4 4 1 0 -0.4 2.2 -0.4 -0.4 0.1 0.6 -0.4 0.6 0.6 -1.0 -1.5 0 79 73 132 6 132 38 35 6 11 60 -1.1 0.6 0.4 1.7 -1.0 1.7 -0.3 -0.4 -1.0 -0.9 0.2 Global competitiveness index Political
Overview). On the positive side, positive technical pressure can come from aggressive SMP-buying, which could offset the market signal, and from idiosyncratic issues such as the Irish re-investment of the promissory note proceeds.3 In terms of valuations, our updated sovereign risk index (Exhibits 20 and 21) shows that: 1) Austria, Portugal and Spain are cheap vs. fair value. We agree with Austria, but not on Portugal and Spain. We believe the case for an Austrian downgrade is dubious if the authorities take appropriate action (see below). On Portugal, we believe that there is a high probability that investors will be at least asked to extend maturities of existing Portuguese bonds later in the year, making current valuations (especially at the short end) unattractive. Despite a low debt/GDP ratio, we believe that Spanish bonds can come under significant pressure if the country loses market access. 2) France, Greece and Netherlands trade on the
Industrial
2012 GDPpc (EU=100)
risk index action index
124 116 123 104 108 63 119 90 126 54 80 Sov. risk index -0.9 -0.6 -0.9 -0.1 -0.3 1.5 -0.7 0.4 -1.0 1.8 0.8 -0.6 0.1 -0.6 0.1 -0.7 1.3 0.4 0.2 -0.6 0.4 0.2
Cross sectional z-score
* Calculated as weighted average of cross-sectional z-spreads; debt/GDP ratio accounts for 25% of the weight, other variables have equal weight. Source: Eurostat, EC, ECB, National Central Banks, World Economic Forum, European Industrial Relation Online
3 See Euro cash, Global Fixed Income Markets Weekly, 14 October 2011 for details.
44
expensive side. We agree on all accounts. We have been highlighting for a while that Frances fundamentals are weaker than its AAA peers. 10Y Greek bonds are fairly priced for the second PSI (with 50% haircut), but we expect another debt restructuring in the not too distant future. With regards to the Netherlands, we believe that domestic support is likely to help only at the ultralong end of the curve. We cover the most interesting issues for the major Euro area countries in the section below. Austria We believe most macro variables for Austria look very good, and certainly consistent with a AAA rating. However, we see two sources of weakness for Austria that need to be addressed: 1) The bond market is small, with limited domestic participation (less than 25% according to financial accounts data) and therefore potentially vulnerable to the vagaries of the market; and 2) Austrian banks net exposure to countries outside the Euro area, in particular to Eastern Europe, remains the countrys Achilles heel (Exhibit 22). Our working assumption is that the government has the resources and the will to provide a stronger support framework in place: S&P recently expressed the desire to see non-voting participation capital that some banks raised during the crisis transformed into more solid ordinary capital. Belgium Despite investors focus on higher-than-average debt/GDP ratio, most of Belgiums economic variables are pretty good, especially the net international investment position. Political issues have been in the spotlight for more than a year now, and despite the country working well on autopilot for quite some time, there is a need for a consensus on roughly 3% of GDP worth of fiscal consolidation in 2012. Discussions are ongoing but have been so far unsuccessful. The fragility of financial institutions is the second issue that concerns us, especially given the amount of support that was needed in the 2008-09 crisis, and more recently, with Dexia (Exhibit 23). Finland Finland is a very solid AAA country from a macro point of view, but is likely to be penalised vs. Germany by the relative lack of liquidity. However, low borrowing requirements and the potential support coming from its extensive general government assets should limit excessive underperformance. Finland is in a unique
Exhibit 21: We find that Austria, Portugal, and Spain trade on the cheap side vs. fundamentals, whereas Greece, Netherlands and France are on the expensive side
10Y spread to Germany vs. fair value implied by sovereign risk index*; bp; Index Current Model Difference % difference
Austria Belgium Finland France Greece Ireland Italy Netherlands Portugal Spain
-0.6 0.1 -0.7 0.1 1.5 0.4 0.2 -0.6 0.4 0.1
147 282 67 147 2,425 625 494 56 910 469
81 299 71 304 4,128 565 399 84 550 298
66 -18 -4 -156 -1703 60 95 -28 360 172
81% -6% -6% -51% -41% 11% 24% -33% 65% 58%
* 10Y spread = exp (1.88*sov risk index + 5.6); R-squared: 86%; see Exhibit 20 for the computation of the sovereign risk index.
Exhibit 22: Austrias main weakness lies in the net exposure of its banking sector to extra Euro area countries, second only to Ireland
Net external assets of the banking sector; as of September 2011; %GDP
80% 60% 40% 20% 0% -20% Netherlands Portugal Austria Ireland Italy France Belgium Germany Greece Finland Spain
45
Source: ECB
Exhibit 23: Belgian financial institutions have received considerable state support since the beginning of the crisis
Details of state support to major financial institutions with significant activities in Belgium; bn
2011 Asset Asset protection protection Capital Guarantees* scheme Capital Guarantees scheme Dexia Belgium 3.0 91 4.0 54 Others 3.4 59 36 KBC Belgium 7.0 20 Fortis** Belgium 4.7 Others 6.5 Ethias Belgium 1.5 2008-09
* Total guarantees agreed for Dexia were 150bn in 2008-09, peak was <100bn, currently <30bn. ** The original plan for Fortis (capital injection) was eventually changed, with nationalisation of the Dutch arm and sale of the Belgian arm to BNB Paribas.
Exhibit 24: The Finnish government balance sheet is very strong as government assets are significantly higher than liabilities
Net general government assets (financial assets minus financial liabilities) of Euro area countries; % of GDP; latest data available
Exhibit 25: Downgrades of a AAA country have always been preceded by either a negative outlook or a negative watch
Negative outlook/watch and rating actions on AAA countries over the past 5 years;
Country Agency Negativ e outlook Negativ e w atch Outlook (months) 20-May -08 02-Jul-09 30-Mar-09 30-Jun-10 30-Sep-10 12-Jan-09 19-Jan-09 21-May -09 18-Apr-11 14-Jul-11 05-Aug-11 3.6 3.5 0.7 1.6 2.5 5.0 2.7 3.0 0.2 2.5 Watch (months) Dow ngrade dow ngrade dow ngrade Current rating Baa3 Ba1 BBB+ A1 AAAAA Aaa AA+
75% 50% 25% 0% -25% -50% -75% -100% -125%

Netherlands Portugal Austria Ireland Italy Finland France Belgium Germany Greece Spain
Iceland Moody 's 05-Mar-08 Ireland Ireland Spain Spain UK* US US Av erage Moody 's 30-Jan-09 17-Apr-09 S&P Moody 's S&P S&P S&P 09-Jan-09
Moody 's 02-Aug-11 13-Jul-11
* The UK outlook was changed back to stable on 26 October 2010. Source: Bloomberg, Moodys, and S&P
Source: National financial accounts
Exhibit 26: Fiscal problems have been the most quoted reason for a review/watch/rating action on a AAA rated country
Problems Country Iceland Agency Moody's Moody's S&P Moody's S&P S&P Moody's S&P 4 X X 6 X 4 X 2 X X X X X X X X X X X X X X 3 X 1 Economic Fiscal Financial Funding sector costs X X X X X Political External position X
position among the major Euro area countries, with general government assets exceeding liabilities by a wide margin (Exhibit 24). France We believe that the main issue that investors have with France relates to the countrys ability to hold to its AAA rating. Moodys and S&P have laid the ground for an action on France with their recent comments. Moodys focused on the deterioration in debt metrics and the potential for further contingent liabilities to emerge are exerting pressure on the stable outlook of the governments Aaa debt rating. S&P instead focused on macro performance, suggesting that in case of a double-dip in the Euro area, France would likely be downgraded. We review rating actions on AAA countries over the past five years (Exhibits 25 and 26). We find that: 1) Negative outlooks and negative watches are unlikely to be reversed. 2) In all instances, a downgrade has been preceded by either a negative outlook or a negative watch. 3) Downgrades have occurred with an average lag of roughly 3 months from a negative outlook. 4) Fiscal problems are (unsurprisingly) the most quoted reason for action, followed by deteriorating economic outlook, problems in the financial sector which lead to
46
Reasons quoted by rating agencies for review/watch/rating action on AAA countries over the past 5 years;
Technical issues
Ireland Ireland Spain Spain UK US US Grand total
X X 2
Potential issues for France
Source: Moodys and S&P
Exhibit 27: In the event of a French downgrade, we expect the 10-15Y part of the French curve to find better support than other sectors, thanks to the large domestic insurance sector
Total assets (financial and non-financial) of insurance corporations of the Euro area countries; 2010; Insurance corp. assets
bn Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Total
Source: ECB
% GDP 40% 67% 27% 96% 60% 7% 33% 68% 25% 100%
119 250 52 1910 1549 16 525 414 274 5108
contingent liabilities, in addition to question marks about political resolve. 5) Only the UK has been able to return to a neutral outlook thanks to a combination of lower-than-expected cost of contingent liabilities, strong political resolve, and improving economic outlook. In terms of macro and fiscal variables, France starts from a weaker position than other AAA countries (Exhibit 20). A further deterioration in the macro outlook and in financial stress is likely to add to the pressure on France. Worries about contingent liabilities, increasing cost of funding, and political uncertainty associated with presidential elections are likely to prompt a rating action in 2012. We expect pressure to come from foreign central banks diminished appetite for short-end paper. We are more sanguine about the prospects for the 10-15Y part of the curve, given the massive size of the domestic insurance industry (Exhibit 27) and the likelihood of asset rotation into long-dated French bonds. Official data showed that domestic insurance companies hold around 220bn in French government securities, out of 1.9tn of assets. Germany At the moment of writing, Germany is the only Euro area country that has always traded like a safe-haven asset, thanks to solid macro fundamentals but also helped by the liquidity that its size and bond futures provide. In our central scenario this will continue, with yields expected to drop even further on the back of intra-Euro area reallocation flows. However, policy makers inability to provide an effective backstop and an escalation of the crisis beyond our central scenario would likely challenge our view: at that point, a run on the currency by extraEuro area investors would dent Germanys safehaven status. Greece The appointment of a technocratic government in Greece should pave the way for successful implementation of the second attempted PSI even though the NPV of the new structure might be below 30c. We expect the voluntary bond exchange to go through, but we believe that bond holders will eventually be asked for a further contribution, although it is difficult to predict the timing. The bond price curve has flattened aggressively since the announcement of the first PSI (Exhibit 28). We expect further selling pressure into the end of the year as
Exhibit 28: The Greek bond price curve is very flat as investors attach a high probability to a debt restructuring
Clean-price GGB curve after announcement of first and second PSI and current*; points
90 80 70 60 50 40 30 20 2012 2016 2020 2024 2028 2032 2036 Bond maturity date
* Dates used: First PSI: 22 July 2011, Second PSI: 27 October 2011 and current: 18 November 2011.
PSI 1 PSI 2 Current 2040
Exhibit 29: We remain sceptical about Irelands ability to deliver the needed fiscal consolidation from a deficit/GDP above 10% in 2011
General government budget balance for Ireland and average of peripheral countries ex Ireland; realised numbers for 2007-2010 and forecasts for 2011-2012; % of GDP
2 0 -2 -4 -6 -8 -10 -12 -14 -16
0.1
Ireland
Av g. peripheral ex Ireland
-2.3 -5.2 -7.3 -10.6 -14.2 2007 2008 2009 2010* 2011 2012 -8.6 -11.2 -10.3 -6.3 -8.6 -4.9
*Irish deficit number for 2010 exclude the 20.1% one-off banking sector support measures adjustment. Source: 2012 EC autumn forecasts
European financial institutions continue to clean their balance sheets. Ireland Ireland is the success story of 2011, with 10Y spreads to Germany close to unchanged on the year despite a selloff at the short end. We believe the reasons for this relative outperformance are both fundamental and technical: 1) For the first time in 4 years, there appears to be some stabilisation in the debt/GDP ratio trajectory; 2) The peak in Irish outperformance coincided with heavy domestic purchases on the back of the capital injections in the
47
summer. Irish banks increased their bond portfolio by more than 20% in August and September alone. Going forward, we remain sceptical about Irelands ability to deliver the required fiscal consolidation from a deficit/GDP above 10% (Exhibit 29), still-enormous net external liabilities, and still-falling house prices. We are medium-term negative on short-dated bonds as we believe they do not discount a high enough probability of restructuring, even in the milder form of maturity extensions. Italy We cover Italy at length in the Overview and in the demand/supply section below. The bottom line is that we see a risk that access to the market may become impaired, prompting significantly higher yields and downgrades, with technicals especially negative for inflation-linked bonds. The Netherlands Dutch bonds have remained, until recently, remarkably stable compared with Germany, thanks to their strong perceived creditworthiness. Amid overall sound macro fundamentals, the ongoing decline in house prices raises some concerns in the context of a heavily leveraged household sector, as highlighted by the Dutch National Bank in the Financial Stability Report. Going forward, we see the emergence of further market segmentation in the DSL market (Exhibit 30), with domestic insurance companies and pension funds likely to keep the long and ultra-long ends of the curve supported, and more pressure on the short end of the curve due to lower structural domestic demand. Portugal Our assessment of the latest troikas Portugal review is mixed: the 5.9% 2011 deficit/GDP target will only be met through an account expedient due to expenses overruns, but the troika expressed confidence about hitting the 4.5% 2012 target. The need for structural reforms was also stressed. We believe that a long period of fiscal tightening and internal devaluation to restore competitiveness in a highly leveraged economy is unlikely to prove successful. An ex-official Treasury official recently stated that Portugal might need additional 20-25bn in addition to the 78bn packaged agreed in the spring as public companies are unlikely to be able to roll their debt, an event not contemplated when the package was approved. Given the risks of PSI in the latter part of the year when a new package will be negotiated, we are
48
Exhibit 30: We expect domestic insurance companies to keep the long and ultra-long ends of the Dutch curve well supported, whereas the short/intermediate part of the curve will remain under pressure due to lack of structural domestic demand
Dutch bonds spread to interpolated German curve; bp
80 60 40 20 0 2012 2018 2024 2030 Bond maturity date 2036 2042
Exhibit 31: Spains strong domestic investor base provides support to its bonds markets
% of domestic investors in central government securities in selected Euro area countries; %
70% 60% 50% 40% 30% 20% 10% 0%

Portugal Ireland France Italy Germany Greece
62% 55%
34% 19%
33% 17% 20%
Source: J.P.Morgan estimates, national central banks, debt management agencies
generally negative on Portuguese bonds, especially in the 2-3Y part of the curve.4 Spain Spains attempt at rebalancing its economy is impressive, especially in terms of current account balance, which moved from -9.6% of GDP in 2008 to a more manageable -3.4% in 2011. The fiscal journey has also been laudable, with more than 4% points of deficit reduction between 2009 and 2011, but Spain continues to
4
See Overview, Global Fixed Income Markets Weekly, 4 July 2011 for details.
Spain
suffer from 1) poor fiscal discipline at the local level; and 2) uncertainty about the final cost of measures to clean banks balance sheets of non-performing assets. Among the positives for Spain, it is worth highlighting the high percentage of domestic investors (Exhibit 31). However, we see a risk that Spain may lose market access at some point during the course of 2012, pushing yields higher. Peripheral duration neutral curve flatteners and weighted flatteners are attractive After a few years of a sovereign debt crisis, it is empirically well established that credit spread curves are directional to the level of spreads: the wider the spread to Germany, the flatter or inverted the credit curve (Exhibit 32). The process is very common in credit markets when bonds transition from high-grade to highyield. Investors can analyse the relative merits of curve positions (boxed vs. Germany) vs. outright spread positions looking at relative value and/or carry and slide. Based on our view that spreads are going to widen across the board and our flattening view on the German curve, we recommend duration-neutral curve flatteners in France, Italy, and the Netherlands. Selling high-price, short-dated bonds and buying low-price, long-dated ones can make the strategy more attractive (see below). Price convexity between short and long dated bonds For investors who want to be protected in case of a tightening of spreads to Germany that would result in a steepening of the credit curves, a source of relative value can be found in the empirically convex relationship between prices at the short and long ends of the curve. As an example, we look at 3Y and 30Y Portuguese bonds (Exhibit 33) to highlight that 1) initially, 30Y bond prices fall more forcefully than prices at the short end of the curve, but 2) eventually, the short end of the curve catches up as the probability of a shortdated credit event increases. Based on the empirical regression, when 3Y Portuguese prices were at 100, 90, and 80, the sensitivity of 30Y prices were 1.7, 1.2, and 0.6, respectively. At an extreme, under most debt restructurings, bond prices across maturities tend to converge to similar levels. The Italian and Spanish price curves have, until now, exhibited limited convexity, but we expect this to change in case of further stress, in line with the experience of Greece, Ireland, and Portugal. Empirically, 30Y Italian bond prices are exhibiting roughly twice as much price volatility as 3Y ones (Exhibit 34). Selling
Exhibit 32: Credit spread curves are directional to the level of spreads: the wider the spread to Germany, the flatter or inverted the credit curve. Given our view that spreads are going to widen across the board, we recommend duration-neutral curve flatteners in all countries
3s/10s spread curve to Germany for the Euro area countries ex Greece, regressed against 3Y spread to Germany; bp
100 0 -100 -200 -300 -400 -500 0 500 1000 3Y spread to Germany ; bp 1500 y = -0.0002x 2 - 0.014x + 7.4 R 2 = 98% Ireland Belgium Spain Italy Portugal
Exhibit 33: Under stress, the price curve exhibits convex behaviour as the short end finally catches up with the long end in terms of losses
30Y Portuguese bond price vs. 3Y Portuguese bond price; since May 2010*; clean prices; points
100 90 80 70 60 50 40 60 70 80 90 3Y Portugal bond price; points 100 110 y = 0.021x 2 - 2.7x + 134.5 R 2 = 94% Slope: 0.6 Slope: 1.2 Slope: 1.7
* The ECB started the SMP on 10 May 2010.
Exhibit 34: We recommend selling 100mn of BTPs Jun14 vs. 50mn of BTPs Feb37 to exploit the convexity of the price curve relationship
Recommended trade weight on short BTP Jun14-long BTP Feb37 based on empirical price relationship since the beginning of the first ECB SMP*; BTP Jun14 BTP Feb37 Ratio Dirty price 95.6 69.9 0.73 Empirical price beta 1 2.01 Nominal weight 100.0 49.7 Cash weight 100.0 36.3 Yield (%) 6.23 6.67 Modified duration 2.25 12.94 PVBP 2.15 9.04 * BTP Feb37 = 235.3 - 5.64*(BTP Jun14) + 0.041*(BTP Jun14)^2; R-squared: 95%; calculated using clean prices.
49
100mn of BTPs Jun14 into 50mn of BTPs Feb37 would result in a trade that is locally hedged for small price changes and would start to be profitable in case of further sovereign stress. How to price high- low-coupon bond yield spreads under sovereign stress As the sovereign crisis evolved from liquidity to potential solvency issues, investors have been switching out of high-coupon, high-price bonds into low-coupon, low-price bonds with similar maturity. Exhibit 35 shows how the yield spread of the pair with the highest coupon mismatch (BTP Aug23-Nov23) has moved from virtually zero to around 60bp. Although it seems intuitive to recommend switches between bonds with similar maturity and yield, and different coupon and price under sovereign stress, it is not straightforward to estimate the fair value once yield spreads have already moved as in the example above. In the current environment of relatively flat yield curves like the Italian or Spanish ones, traditional cheap/dear analysis based on bootstrapping would suggest that wide yield spreads between high- lowcoupon bonds with similar maturities are not justified and have a source of relative value.5 One needs to incorporate in the analysis risk free rates, the probability of a credit event throughout the bond life, and the recovery rate in order to justify higher yields for higher coupon bonds. We analyse how bond maturity and coupon rates impact yield spreads for same-maturity bonds in a simplified world (with constant conditional probability of default per year, constant recovery rate and constant zero risk-free rate). We calculate NPVs for theoretical bonds with different characteristics and compare their fair value IRRs. Typically, for bonds with the same maturity and fixed coupon differential: 1) the longer the maturity of the bond the wider the yield spread (Exhibit 36), 2) the lower the coupons the wider the yield spread, 3) the higher the conditional probability of debt restructuring, the higher the impact on the yield spread (Exhibit 37).
Exhibit 35: The yield spread between high- and low-coupon bonds has risen with higher Italian yields
Yield spread between BTP 9% Nov23 (current clean price 111.6) and BTP 4.75% Aug23 (current clean price 82.6) regressed against yield of BTP 4.75% Aug23; last six months; %
70 60 50 40 30 20 10 0 4.00 5.00 6.00 7.00 BTP 4.75% Aug23 y ield; % 8.00 9.00 y =20.1x + -87.5 R 2 = 81%
Exhibit 36: The high- low-coupon impact on yield spread is positively correlated with bond maturity but negatively with the coupon rate
Theoretical yield spread between two bonds with the same maturity and coupon of the first bond 2% lower than the coupon of the second bond under different maturities and coupon assumptions; NPV and IRR calculations are based on 0% risk free rate, conditional default rate of 5% per year, 40% recovery rate; %
0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0.00 2.00 4.00 6.00 8.00 10.00 10Y 30Y
Coupon rate of low er-coupon bond; %
Exhibit 37: The high- low-coupon impact on yield spread is positively correlated with the expected probability of restructuring
Theoretical yield spread between two bonds with the same maturity and coupon of the first bond 2% lower than the coupon of the second bond under different coupon assumptions and conditional default probabilities per year; NPV and IRR calculations are based on 30Y bonds, 0% risk free rate, 40% recovery rate; %
1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 0.00 5% conditional prob. of default 2.00 4.00 6.00 8.00 10.00 10% conditional prob. of default
In the typical textbook world of risk free government bonds and upward sloping yield curves, high-coupon, high-price bonds should trade with a higher yield than low-coupon, low-price bonds with similar maturities.
50
Coupon rate of low er-coupon bond; %
We can now apply this stylised framework to two real cases: 12Y BTPs and 20Y-25Y Bonos (Exhibit 38). What is the fair value spread between 12Y BTPs with coupons of 4.75% and 9.00% when the low coupon bond is yielding around 7.00%? We assume a fixed recovery rate of 40% and fixed risk free rate of 2%, and we calculate the annual conditional probability of default that is consistent with the market yield of the low coupon bond. We then use the same parameters to calculate the theoretical yield on the high coupon bonds6. Based on our model, the spread should be around 75bp vs. current market level of 60bp, making the market spread broadly fair. What is the fair value spread between the 20Y 5.75% Bono and the 25Y 4.20% Bono when the latter is yielding around 7.00%? Based on our model, the spread should be around -80bp vs. current market level of -7bp, suggesting a significant mispricing. We therefore recommend investors sell the Bono Jul32 and buy the Bono Jan37. In conclusion, under sovereign stress, the standard cheap/dear analysis is not useful to deal with bonds issued by the same issuer, with similar maturities but with significantly different coupons. The fair value yield spread between bonds with different coupons is positively influenced by maturity and default probability, and negatively influenced by the coupon rate for a constant coupon difference. We believe the BTP Aug23Nov23 yield spread is not far from fair price, but the Bono Jul32-Jan37 spread is too tight.
Exhibit 38: We believe that the BTP 4.75% Aug23-9% Nov23 spread is close to fair value, but the Bono 5.75% Jul32 is too expensive vs. the Bono 4.2% Jan37
Market and theoretical* yields on BTPs 4.75% Aug23 and 9% Nov23 and Bonos 4.20% Jan37 vs. Bono 5.75% Jul32; % Maturity Coupon Dirty price Mkt yield Theoretical yield Aug-23 4.75 84.1 7.06 7.06 BTP BTP Nov-23 9.00 112.2 7.65 7.81 Spread (bp) 59 75
Bono Jan-37 4.2 71.3 7.01 7.01 Bono Jul-32 5.75 88.7 7.08 7.80 Spread (bp) -7 -79 * We adjust a flat conditional probability of default under 40% recovery rate and 2% risk free rate to match the market yield of the first bond and calculate the theoretical yield of the second bond under the same parameters.
Exhibit 39: Euro area banks and extra-Euro area investors are the largest holders of Euro area government bonds
Split of Euro area general government bond holdings; as of 2Q11 2Q11 bn % MFIs 1,579 23% Eurosystem (ECB, NCBs)* 469 7% Investment funds 671 10% Insurance companies + pension funds 1,228 18% Others 633 9% Rest of the World 2,183 32% Total 6,763 100% * The number includes holdings other than SMP. Source: ECB
Exhibit 40: but banks turned net sellers of bonds in 3Q11

Change in Euro area general government bond holdings*; bn 2009 2010 1H11 3Q11 MFIs 228 41 53 -43 Eurosystem (ECB, NCBs)** 42 104 15 81 Insurance co. + pension funds 100 101 13 n/a Others -13 199 47 n/a Rest of the World 248 154 150 n/a Total 604 598 278 n/a * SMP purchases from 1 October 2011 till 18 November 2011. ** Including revaluation/devaluation changes up to 3Q11. Source: ECB
Bond demand: In whose hands is the hot potato?

Given the increasingly technical nature of the Euro area government bond market, we look at its investor base, first in aggregate (Euro area) and then for the four major countries. We conclude that despite different investor bases, no country is in the position of Japan, where 95% of the countrys domestic investor base acts as a potent dampener of sovereign stress (see Japan). Euro area According to ECB data, in mid-2011, Euro area banks (monetary and financial institutions [MFIs], to be precise) held roughly 23% of Euro area (general) government securities, whereas the Euro system held 7% of the total. It is important to highlight that SMP holdings account for less than 50% of that amount. Euro area
6
4Q11* n/a 33 n/a n/a n/a n/a
insurance companies and pension funds had sizeable holdings (18%), whereas investors outside the Euro area held almost 1/3 of the total (Exhibit 39). We estimate that central banks constitute the lions share of external holdings. For more specific cross-border exposure in the banking sector, please refer to Exhibits 11 and 12 in the Overview. Flow-wise (Exhibit 40), Euro area banks were decent buyers of bonds in 1H11, but consistent with anecdotal evidence, they became heavy sellers in 3Q11 (and likely in 4Q11, based on individual bank updates). In the recent words of the president of the
51
Using inputs that are extracted from market prices is computationally more complex, but does not change the results significantly.
European Banking Federation: The banks are doing exactly what they should be doing: they are reducing their risk We can see that clearly as now Italian bonds are being sold off. The country wise split shows that in 3Q11, French banks were the most aggressive in reducing their government bond portfolios (-16bn), followed by Spanish and German ones, while Italian banks were still adding to their government bond portfolios.7 After a relatively quiet 1H11, the ECB had to significantly increase purchases: the central bank releases on a weekly basis their SMP holdings, providing the timeliest information. We estimate that between the end of 3Q11 and mid-November, another 33bn was bought by the ECB in addition to 81bn in 3Q11. For the other investors, data is available only until 2Q11. It is worth highlighting the strong buying from extra-Euro area investors and subdued purchases by insurance companies and pension funds in 1H11. Germany Given the relatively large amount of local debt in Germany (almost 40% of the total, between securities and loans), we focus on central government securities. In trying to estimate central government holdings, we faced the problem that, based on different sources, there do not seem to be enough German government bonds and Tbills around to satisfy all the reported demand. We estimate that domestic holdings of central government securities are a small percentage of the total, and foreign central banks constitute more than one third of the investor base (Exhibit 41). Germanys role of an international safe haven is confirmed by the 6% jump in foreign holdings of general government debt in 2010 to 60%, with a further increase in 1Q11. However, as discussed previously, we believe that strong international participation, in particular from overseas investors, makes German bonds vulnerable to an escalation of the crisis. If the fabric of the Euro zone is perceived to be under threat, even German bonds might lose their safe-haven status and be sold by foreign central banks. France Exhibit 42 shows our best estimate of the investor holding of French government securities.8 Nearly two
7
Exhibit 41: We estimate that foreign central banks alone hold more than one third of German bonds and T-bills, making even the German bond market vulnerable to a severe escalation of the sovereign crisis
J.P.Morgan estimate of German central government securities holdings; as of 1Q11; bn
bn Domestic Foreign MFIs Insurance companies Mutual funds Central banks Others Total 107 40 39 4 20 210 210 150 116 395 10 881 Total 317 190 155 399 30 1091 10% 4% 4% 0% 2% 19% % Domestic Foreign 19% 14% 11% 36% 1% 81% Total 29% 17% 14% 37% 3% 100%
Source: Bundesbank, ECB, BIS, EFAMA, IMF
Exhibit 42: We estimate that almost 30% of French bonds and T-bills are held by foreign central banks that are likely to be sensitive to a rating downgrade
bn Domestic Foreign MFIs* Insurance companies Mutual funds Central banks Others Total 147 221 21 n/a 53 442 141 100 133 363 129 866 Total 288 321 154 363 181 1308 Domestic 11% 17% 2% n/a 4% 34% % Foreign 11% 8% 10% 28% 10% 66% Total 22% 25% 12% 28% 14% 100%
J.P.Morgan estimate of French central government securities holdings; as of 1H11; bn
* Domestic central bank might be under MFIs Source: AFT, ECB, BIS, EFAMA, IMF
Exhibit 43: We estimate that foreign banks are by far the largest foreign players in Italian bonds, putting pressure on prices as they de-lever their portfolios
bn Domestic MFIs Insurance companies Mutual funds Central banks Others* Total 188 165 64 67 335 820 Foreign 189 85 77 66 258 675 Total 377 250 141 133 593 1494 Domestic 13% 11% 4% 4% 22% 55% % Foreign 13% 6% 5% 4% 17% 45% Total 25% 17% 9% 9% 40% 100%
J.P.Morgan estimate of Italian central government securities holdings; as of May 2011; bn
* 5% of general government securities are held abroad, but is attributable to Italian savers (4.3% of total debt). Source: Bank of Italy, ECB, BIS, EFAMA, IMF
ECB data shows changes due to outright buying/selling and to revaluations/devaluations. 8 See Euro cash, Global Fixed Income Markets Weekly, 21 October 2011 for more details.
52
thirds of French government securities are held by foreign investors, of which the major chunk is owned by foreign central banks, based on our estimates (28% of the total). We believe that, from a technical point of view, France is fragile due to the large percentage of foreign investors who are likely to be sensitive to its AAA rating. Official data shows that foreign holdings have been declining since mid-2010 after steady increases since the introduction of the Euro.
Italy Banca dItalia publishes lagged data on the split between various domestic investors (MFIs, central banks, mutual funds, insurance companies, pension funds, and others) and foreign investors in government securities. They do not provide the holding split among foreign investors, who hold nearly 45% (675bn, Exhibit 43) of government securities as per the most recent data. However, they do estimate that 5% of general government securities held abroad are attributable to Italian savers9. We collect information from various sources to get an idea of the foreign investor split in Italy. BIS data shows that non-Italian banks have exposure of around 200bn to the Italian general government, and we assume that around 95% of that is in Italian government securities. Euro area investment funds hold nearly 670bn of Euro area government securities, and we estimate that holdings of Italian bonds are around 140bn (20% of total), based on Italys contribution to the JPM EMU government bond index. IMF (CPIS) data shows that foreign central banks held nearly 50bn of Italian securities on their books at the end of 2009. Given that the IMF survey captures only a small sub-sample of reporting entities, and reserves have increased since 2009, we estimate that total numbers might be 30% higher (around 65bn). Lastly, based on our insurance companies equity analysts database, we estimate that nearly 85bn of Italian government securities is held by non-Italian insurance companies. Flow-wise, available data on government bond securities shows stable foreign holdings for the first 5 months of the year, but timelier Balance of Payments data shows a sharp decline in Italys foreign liabilities between July and August (Exhibit 44). The 11bn increase in domestic banks holding of government bonds in 3Q11 and aggressive purchases of Italian bonds through the ECB SMP (we estimate 95bn of purchases of BTPs) are also consistent with domestics and the ECB offsetting large international selling. As discussed in the Overview, we expect selling pressure from international investors to continue in coming months.
Exhibit 44: Foreign investors have reduced exposure to Italian financial assets by 40bn over July-August
Cumulative changes in foreign holdings of Italian assets*; bn
120 100 80 60 40 20 0 -20 Aug-09 Dec-09 Apr-10 Aug-10 Dec-10 Apr-11 Aug-11
* Italian government securities account for roughly 2/3rds of foreign liabilities. Source: Bank of Italy Balance of Payments data
Exhibit 45: Spanish banks are the key players in the Bono market, making Spain less vulnerable to selling pressure from international investors
bn Domestic Foreign MFIs Insurance companies Mutual funds Central banks Others Total 175 36 23 63 32 328 51 25 45 73 11 205 Total 227 61 67 136 43 534 33% 7% 4% 12% 6% 62% % Domestic Foreign 10% 5% 8% 14% 2% 38%
J.P.Morgan estimate of Spanish central government securities holdings; as of August 2011; bn

Total 42% 11% 13% 25% 8% 100%
Source: Bank of Spain, ECB, BIS, EFAMA, IMF
Spain Tesoro Publico provides detailed domestic and foreign investor holding data for government securities (Exhibit 45), but we use our own estimates for the split of foreign investor types based on other sources.10 Spain is the country with the lowest share of foreign investors (latest data, 38%). Domestic investors, especially the domestic banking sector, are the most important players in the Spanish government bond market and based on available information, they have continued to support the segment in 2011. Domestic banks increased their exposure by nearly 6%-pts, from around 28% at the end of 2010 to 33% at the end of August 2011. From a technical point of view, Spain looks less vulnerable than Italy.
In the latest Financial and Stability report, Banca dItalia analysts stated, An estimated 4.3% (of total debt) is held by individually managed portfolios and investment funds administered by foreign intermediaries but attributable to Italian savers.
10
The foreign split by investor type published by the Spanish Treasury is based on non-resident withholding tax refunds.
53
Bond supply: Expected to decline

Under the assumption that no other country loses market access, we expect around 710bn of conventional supply in 2012, around 20bn lower than in 2011 (Exhibit 46), whereas net conventional supply would decline by almost 75bn due to higher redemptions. Our economists forecast Euro area central government deficits to decline by 50bn to around 275bn in 2012 (Exhibit 47), despite their view that non-core countries will fail to hit the targets. The most important conclusions on conventional supply are: 1) In core countries, we expect gross supply to remain stable or decline marginally 2) We project Italian gross conventional issuance to decline by 5bn to around 145bn. On the other hand, we expect Spanish gross supply to remain stable as the increase in redemption is offset by a decline in central government deficit. 3) Even though Irish officials have expressed their intention to return to the issuance market in 2012, we believe this is unlikely and expect them to remain out of the bond market, along with Greece and Portugal, in 2012. Other supply We estimate that floaters, inflation-linked, and zerocoupon bonds add another 60-65bn to Italys supply; the numbers for the other countries is significantly smaller. Italy may also increase its T-bill issuance to meet its funding needs. For the rest of the Euro area countries, we forecast zero net T-bill issuance due to governments reluctance to further shorten the maturity of the debt after cutting the T-bill outstanding by nearly 50bn in 2011. -denominated, non-conventional net issuance would likely remain close to zero, but we expect a decline in non--denominated issuance, given the declining appetite of international investors. Most countries will announce issuance guidelines in December, and we will publish the Euro area 2012 supply update in early January. Italy and Spain funding in 2012 Italian redemptions in 2012 are very heavy, with around 200bn of marketable bonds maturing (Exhibit 48).
54
Exhibit 46: Government conventional bond supply will decline marginally in 2012
Gross conventional bond issuance history and J.P.Morgan 2012 forecast*; bn
900 850 800 750 700 650 600 550 500 450 400 512 584 535 533 506 487 562
820
839 733
711
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012f
* Under the assumption that no other country loses market access.
Exhibit 47: 2012 net conventional supply will decline by around 75bn due to an increase in conventional bond redemptions
Gross and net conventional bond issuance* and deficit for 2011 and J.P.Morgan forecasts for 2012; bn
2012 Gross bond issuance Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Total 20 40 13 180 175 0 0 145 48 0 90 711 Net bond issuance 10 16 7 95 18 -30 -6 35 18 -10 49 202 Gross Deficit 10 13 7 82 26 18 14 37 12 9 45 274 bond issuance 2 2 1 -4 -6 0 0 -5 -5 -7 1 -22 v s. 2011 Net bond issuance 0 -3 0 -10 -16 -8 -1 -27 -6 -8 4 -75 Deficit -1 -2 -1 -10 4 2 -2 -26 -7 -1 -3 -47
* Gross conventional issuance is not equal to the sum of net conventional issuance and deficit as a part of the deficit is met by non-conventional supply and other sources of funding.
Redemptions concentration is particularly high between February and April (around 90bn). Based on the last official release, Italy had around 14.5bn in Treasurys payment account (conto disponibilit) at the end of September 2011. We estimate the number to be in the range of 40-50bn by the end of 2011, based on Italys net issuance since the end of September and our estimate of Italian deficit seasonal pattern. This amount alone will not be sufficient to cover their funding requirements in 1H12, but see Overview Appendix-2 for other sources of funding for Italy.
Exhibit 48: Italy has a highly concentrated redemption schedule for 2012, while Spain has a light redemption schedule, which gives it an opportunity to front-load
Exhibit 49: The EFSF/EFSM has issued around 45bn of bonds so far in 2011, with the bulk of the issuance coming from EFSM
Details of the EFSF and the EFSM issuances in 2011; EFSF supply highlighted in grey; bn
Date 05-Jan-11 25-Jan-11 17-Mar-11 24-May-11 25-May-11 15-Jun-11 22-Jun-11 14-Sep-11 22-Sep-11 29-Sep-11 07-Nov-11 Issuer EFSM EFSF EFSM EFSM EFSM EFSF EFSF EFSM EFSM EFSM EFSF Maturity 04-Dec-15 18-Jul-16 04-Apr-18 04-Jun-21 03-Jun-16 05-Jul-21 05-Dec-16 21-Sep-21 04-Sep-26 04-Oct-18 04-Feb-22 Total EFSM Total EFSF Grand total Size 5.0 5.0 4.6* 4.8 4.8 5.0 3.0 5.0 4.0 1.1 3.0 29.2 16.0 45.2 Issuance level (spread to midswap; bp) 12 6 8 14 0 17 6 20 40 15 104 EMU beneficiary Ireland Ireland Ireland Portugal/Ireland Portugal Portugal Portugal Portugal Ireland/Portugal Ireland/Portugal Ireland
Monthly redemption data for marketable bonds* issued by the Euro area countries; bn Core Peripheral ATS BEF FIM FRF DEM NLG GRD IEP ITL PTE ESP Jan12 2 0 0 15 25 14 0 0 0 0 0 Feb12 0 0 0 0 0 0 0 0 36 0 1 Mar12 1 4 0 0 19 0 14 6 27 0 1 Apr12 0 0 0 18 16 0 0 0 28 0 12 May 12 0 0 0 0 0 0 9 0 1 0 0 Jun12 0 0 0 0 19 0 0 0 2 10 0 Jul12 10 0 0 28 27 15 0 0 17 1 13 Aug12 0 1 1 0 0 0 8 0 12 0 0 Sep12 0 12 6 12 21 0 0 0 11 0 2 Oct12 0 0 0 19 16 0 0 0 20 0 20 Nov 12 1 0 0 0 0 0 0 0 13 0 0 Dec12 0 8 0 5 17 0 2 0 30 1 0 Total 15 25 7 98 160 30 33 6 198 13 49 * Marketable bonds include conventionals, floaters, zero-coupons, linkers, and international bonds.
* Out of 4.6bn issued, 1.2bn was used for BoP and the remaining 3.4bn was given to Ireland.
In contrast, Spain has a very light redemption schedule in 1H12, giving it some time to front-load. Further, Spains most liquid cash balances have remained in the 50-60bn range since 2009, with the latest published balance of 59bn at the end of 2Q11. EFSF/EFSM supply The EFSF and EFSM have issued around 45bn of bonds so far in 2011 to fund the EU share of the Portuguese and Irish aid packages (Exhibit 49). The bulk of the issuance came from the EFSM as the EFSF remained out of the markets for most of 2H11 as the new version of the EFSF awaited parliamentary approvals. We expect the EFSM to tap the market one last time before the year ends. EFSF/EFSM funding requirements for 2012 to support Ireland and Portugal are quite manageable, at around 32bn, out of which 10bn would be covered by the EFSM and the remaining 22bn would be provided by the EFSF. However, the EFSF issuance numbers may get revised significantly higher once the second bailout package for Greece gets finalised. Given the EFSFs several guarantee structure (no joint guarantees) and significant uncertainty around Greece PSI and the leveraged structures, EFSF bonds have recently come under market pressure. EFSF funding costs, which used to be around 5-10bp above LIBOR, increased dramatically to 104bp when they last issued in early November. In contrast, EFSM bonds, which have a
Exhibit 50: EFSM bonds have outperformed their peers whereas EFSF bonds have been under market pressure, given the high level of uncertainty and the several but not joint guarantee structure
ASW spread for EFSF Jul16, EIB Jul16 and EFSM Dec15 bonds; bp
100 80 60 40 EIB Jul16 20 0 EFSM Dec15 -20 May -11 Jun-11 Jul-11 Sep-11 Oct-11 Nov -11 EFSF Jul16
joint and several guarantee structure, have outperformed their peers (Exhibit 50).
Trading themes
Enter long 10Y duration trades in Germany Outright recession in the region, falling core inflation and a worsening of the sovereign crisis will push German yields down in the first part of the year. We target a trough in 10Y Bunds at 1.25% but the failure to contain the crisis might eventually hurt even German bonds, the last safe haven asset in the region.
55
Enter 2s/10s flatteners The short end of the curve has little scope to rally even if the ECB cuts the refi to 0.50%: we therefore recommend 2s/10s flatteners, with a target of around 100bp by mid-year. Enter intra-EMU spread wideners We forecast generalised spread widening vs. Germany in 1H12. Enter flatteners in non-German curves We also recommend credit curve flatteners: in line with the experience of Greece, Ireland and Portugal, in case of sovereign stress, we expect investors to move from yield to price considerations. Technicals point to flatter curves also in core countries such as France and Netherlands, as we expect domestic support to the ultralong end of the curve. RV on peripherals: weighted flatteners and highcoupon low coupon We also recommend flatteners weighted by the empirical price beta to exploit the convexity between short-dated and long-dated bond prices. We present a framework to analyse the yield spread of high- and low-coupon bonds with similar maturity under sovereign stress: We recommend selling Bonos 2032 vs. Bonos 2037 as the current yield spread is not consistent with the coupon differential.
56
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Pavan WadhwaAC (44-20) 7777-3370 Fabio Bassi (44-20) 7325-8615 Khagendra Gupta (44-20) 7777-1980 J.P. Morgan Securities Ltd
European Derivatives
The EONIA curve was driven by ECB policy rates and liquidity conditions in 2011 We expect excess liquidity to remain elevated in 2012 as peripheral banks continue to rely on the ECB for funding, given limited market access and large bank debt redemptions in 1H12 The EONIA/refi bias is likely to remain around 70-75% of the corridor and, with the ECB expected to cut refi and deposit rates to 0.50% and 0.25% respectively, we expect EONIA fixings to fall to 30bp We recommend longs in 6Mx6M EONIA with a target of around 30bp and bullish option structures on Dec12 Euribor With peripheral sovereign stress expected to escalate in 1H12, we expect the swap curve to bull flatten and recommend carry efficient flatteners which are a proxy for bullish positions but a with better risk profile, and receiver structures on 6Mx5Y swaptions We present a framework to analyse the attractiveness of long-dated forward steepeners but refrain from recommending these trades due to the prevalence of technical factors in this part of the curve The 10s/30s swap curve will continue to be driven by technical factors and is expected to flatten as peripheral spreads widen Implied curve directionality has frequently underestimated delivered directionality during 2011 making conditional structures attractive at the very front end of the curve; we recommend 1s/5s bull flatteners Swap spreads will continue to be driven by peripheral risk and are likely to eventually trade wider than their Lehman peak
We target 2Y and 10Y swap spreads at 145bp and 90bp, respectively, by mid-2012 Delivered directionality of swap spreads has historically exceeded implied directionality as options markets have failed to price in greater delivered counter directionality in risk-off environments, suggesting that conditional swap spread wideners offer better risk/reward than outright swap spread wideners Although spread markets are being driven largely by a single-factor, i.e. sovereign risk, some spread markets such as OIS swap spreads and 3s/6s basis are still trading far below their local/Lehman peaks, suggesting that investors should express riskoff views in these markets EUR volatility in the belly of the curve will increase as the peripheral crisis escalates; we target 3Mx10Y at 9.4bp/day, up from its current level of 8bp/day Front-end volatility, however, will decline as the ECB eventually goes on hold and excess liquidity stays high and we recommend fading flare-ups in 2Y tail volatility Favour long positions in Bund volatility vs. swaption volatility On a cross-market basis, we prefer buying EUR gamma to GBP gamma
EONIA curve
The EONIA curve in 2011 was driven primarily by ECB activity and liquidity conditions. Early in the year, EONIA yields drifted higher on declining excess liquidity and a hawkish ECB. The ECB hiked the refi rate in April and July, and phased out some of the extraordinary liquidity measures to reduce European banks reliance on ECB funding. In the summer, however, further escalation of the peripheral crisis, declining inflationary pressures, and a deteriorating Euro area macro outlook led the ECB to initially extend liquidity measures and later cut refi by 25bp at its November meeting (Exhibit 1).
57
Exhibit 1: EONIA yields rose in the first half of 2011 on the back of ECB rate hikes and declining excess liquidity. Yields have fallen since early August as the ECB cut rates and re-introduced extraordinary liquidity measures
3M EONIA and ECB refi rate; 1 January 2011 18 November 2011; %
Exhibit 2: As excess liquidity increased to over 200bn, EONIA fixings declined but remained above the level implied by our model of EONIA/refi bias
EONIA/refi bias* regressed against non-seasonally adjusted excess liquidity; past 2Y; bp
1.60 1.40 1.20 1.00 0.80 0.60 0.40 04-Jan 08-Mar 11-May 13-Jul
Refi rate
60 40 20 0 -20 -40
y = -9E-06x 3 + 0.006x 2 - 1.1x R 2 = 60% Oct & Nov 11 Maint period
3M EONIA
-60 -80
18-Nov
15-Sep
100 200 300 Ex cess liquidity (NSA); bn
400
* EONIA/refi bias defined as EONIA fixings refi rate.
The dynamic of excess liquidity in 1H11 was broadly in line with our 2011 outlook1 of declining excess liquidity. Indeed, non-seasonally adjusted excess liquidity averaged 40bn and 20bn in 1Q and 2Q, respectively, close to our forecast of 50bn in 1Q and 20bn in 2Q. Our outlook for a further decline in excess liquidity in 2H11 was severely challenged, however. At its August meeting, the ECB, faced with increasing concerns surrounding peripheral sovereign debt and slowing economic growth, provided unlimited liquidity until 1Q12. At its October meeting, it extended liquidity measures even further and re-introduced two new 1Y LTROs, covering liquidity funding until the beginning of 2013. As a result of the extension of these extraordinary liquidity measures, excess liquidity increased to around 200bn (this is the peak 1M average of non-seasonally adjusted excess liquidity). The increase in excess liquidity has widened the bias between EONIA fixings and refi once again, broadly following the dynamic captured in our non-linear model of the EONIA/refi bias.2 However, segmentation in the EONIA market has reduced the impact of excess liquidity on the EONIA/refi bias in recent months. Over the past couple of years, whenever excess liquidity crossed 120bn, the EONIA/refi bias tended to widen to about 85% of the corridor between refi and deposit facility (which would currently be equivalent to -65bp).
1 2
Exhibit 3: likely because peripheral stress resulted in weaker banks finding it more difficult to access funding markets
Residual* from Exhibit 2 regressed against 10Y weighted peripheral spread**; past 3M; bp
30 20 10 0 -10 -20 -30 300 400 500 600 700 10Y w eighted peripheral spread; bp 800 18-Nov -11 y = -0.0004x 2 + 0.47x - 142.1 R 2 = 37%
* Residual obtained by regressing EONIA/refi bias against non-seasonally adjusted excess liquidity over the past 2Y. ** Weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).
See Global Fixed Income Markets 2011 Outlook, 26 November 2010. See Global Fixed Income Markets 2011 Outlook, 26 November 2010.
More recently, however, despite high excess liquidity, the EONIA/refi bias has averaged around -55bp. For instance, the EONIA/refi bias averaged around -45bp in the September maintenance period and around -55bp in the October and November maintenance periods (Exhibit 2). In our view, this decline in the EONIA/refi bias is driven by segmentation in the EONIA market, resulting from peripheral sovereign stress. Indeed, the residual from our long-term model of EONIA/refi bias increases (EONIA/refi bias becomes less negative) as peripheral spreads widen (Exhibit 3). Going forward,
58
we expect this bias to stay at 70-75% of the deposit/refi corridor (currently equivalent to -55bp) in 2012. We believe excess liquidity will remain elevated in 2012. Peripheral banks have funded an increasingly higher proportion of their assets at the ECB over the past few months, and we do not expect this to change next year. On the contrary, we expect an escalation of the peripheral crisis, which will make wholesale funding even more difficult for European banks, compelling them to further increase their reliance on the ECB for funding. The borrowing needs of peripheral European banks for 2012 are quite heavy as there are significant senior unsecured redemptions, especially in 1H12 (Exhibit 4). Absent a full re-opening of the unsecured bond market, we expect that most of these redemptions will be covered by the ECBs MRO/LTRO, and only partially by alternative sources of funding such as retail bonds, private placement and covered bonds. The last 1Y tender will take place in December. With about 140bn expiring at the 3M tender, we expect 75bn to be rolled and about 100-125bn to be tendered at the 1Y (13M tender), which will extend over both 2011 and 2012 year-ends. Overall, we expect an increase in excess liquidity from these tenders of about 30-60bn and a significant increase in the average duration of the existing tenders. Where does that leave the ECB in terms of policy rate? We expect the ECB to cut the refi rate to 1.00% at the December meeting and then to cut it a further 25bp each quarter, down to 0.50% by June 2012, while leaving the deposit facility rate at 0.25% after the December cut. In that scenario, we expect EONIA to trade as low as 32bp, with the EONIA/refi bias at 7075% of the deposit/refi corridor. Therefore, we see value in long positions in mid-to-late 2012 such as 6Mx6M EONIA (Exhibit 5). Although there is a risk that the ECB will not cut rates below 1%, we believe long positions at the front end are likely to offer an option-like profile, with significant downside only if rates are left unchanged at 1.25%. Recent widening of the FRA/OIS basis at the front-end of the curve has worsened the carry in long Euribor positions, which is now negative for the first year of the Euribor curve. Additionally, we expect funding pressures to increase, driving FRA/OIS close to its Lehman highs by mid-2012, before it starts to decline in 2H12. From both a carry and FRA/OIS perspective, we
Exhibit 4: With bank funding markets unlikely to open fully in early 2012, we expect a reasonable portion of European banks borrowing needs to be funded by the ECB
Cumulative total redemptions of senior unsecured bonds for peripheral* country banks in 2012; bn
160 140 120 100 80 59 60 40 1Q12 2Q12 3Q12 114 135
153
4Q12
* Greece, Ireland, Italy, Portugal, and Spain. Source: Dealogic
Exhibit 5: We are bullish on front-end EONIA as ECB refi and deposit rate cuts are expected to drive EONIA yields into the low 30s; receive 6Mx6M forward EONIA
Projected upside from outright longs across the EONIA curve; %
Market Date Dec11 Jan12 Feb12 Mar12 Apr12 May 12 Jun12 Jul12 Aug12 Sep12 Oct12 Nov 12 Dec12 ECB OIS 0.55 0.47 0.45 0.48 0.47 0.47 0.51 0.50 0.51 0.53 0.55 0.57 0.61 Refi 1.00 1.00 1.00 0.75 0.75 0.75 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 J.P.Morgan Depo EOINA/refi bias ECB OIS -0.55 -0.55 -0.55 -0.37 -0.37 -0.37 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 0.45 0.45 0.45 0.38 0.38 0.38 0.32 0.32 0.32 0.32 0.32 0.32 0.32 Upside from longs 0.10 0.02 0.00 0.10 0.08 0.08 0.19 0.18 0.19 0.22 0.24 0.26 0.29
see value in bullish structures only further out the Euribor curve. In Exhibit 6, we plot our projected level of EONIA, under the assumption that the refi rate will be cut to 0.50% by June, and the projected Euribor levels obtained by adding to the projected EONIA curve the maximum and minimum levels of the FRA/OIS basis since
59
1 January 2009.3 We then show the current and projected Euribor curves consistent with our FRA/OIS assumptions. We draw the following conclusions from this analysis. First, although the FRA/OIS is trading close to its widest levels since 1 January 2009, we do not see value in outright longs at the very front end of the curve, as we expect FRA/OIS to widen further. Second, Euribor fixings are expected to decline towards the end of 2012, with our forecast for Dec12 Euribor at 1.00%. Third, we see more value in bullish option structures than outright Euribor longs. We analyse the projected P&L of various bullish option structures on Dec12 Euribor under various yield projections in Exhibit 7. We draw the following conclusions. First, call spreads are relatively expensive and offer only modest upside in our central scenario of Dec12 Euribor at 1%. Second, symmetric structures such as call flies are relatively cheap and offer upside for a wide range of Euribor fixings (0.785% to 1.215%). Third, 1x2s are the most attractive structures as they can be implemented at a credit and are expected to be profitable as long as Dec12 Euribor remains above 0.68%, or a mere 3bp above the minimum Euribor fixing level reached between June 2009 and March 2011. We therefore recommend 1x2s on Dec12 Euribor.
1) Projected EONIA levels under the assumption that the refi rate is cut to 50bp, 2) min/max Euribor threshold based on the min/max* FRA/OIS since 1 January 2009, and 3) 18 November 2011 vs. forecast yield on front Euribor strip; %
Exhibit 6: We expect bank funding pressures to increase, driving the FRA/OIS basis close to its Lehman highs in 1H12 before declining in 2H12. On balance, we expect December 2012 Euribor to trade below its current level
1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 Projected EONIA Mar-12 Jun-12 Sep-12 Dec-12 EONIA + Min FRA/OIS EONIA + Max FRA/OIS Euribor forecast Euribor 18 Nov
Dec-11
* Min/max since 1 January 2009. See footnote 3 further discussion.
Swap curve
The swap curve was driven by ECB activity and the peripheral crisis in 2011. The ECB hiked the refi rate in April and July, driving the curve flatter in a selloff. Monetary policy turned at the August meeting when it
became clear that the ECB had gone on-hold. The 2s/10s swap curve steepened aggressively as the market started pricing in ECB easing. Since then, it has mostly bull flattened (Exhibit 8) going into the ECB cut in November. The 10s/30s swap curve remained uncorrelated with front-end yields, driven mostly by technical factors. Since the beginning of the year, the 2s/10s swap curve flattened 55bp, about 35bp more than was implied in the forward curve early in the year, whereas 10s/30s flattened only 8bp, 12bp less than the forward.
Exhibit 7: and therefore favour bullish option structures in December 2012 Euribor
Upfront cost and projected P&L under different Euribor projections for various bullish option structures on Dec12 Euribor futures; bp Dec 2012 Euribor futures; Forward = 126bp;
Ty pe Call spread Call spread 1x 2 1x 2 Ladder Fly Condor Strike 98.750/99.000 98.750/99.250 98.750/99.000 98.750/99.250 98.750/99.000/99.250 98.750/99.000/99.250 98.750/99.000/99.250/99.500 Contracts +1/-1 +1/-1 +1/-2 +1/-2 +1/-1/-1 +1/-2/+1 +1/-1/-1/+1 Cost (bp) 14.0 24.5 -7.0 14.0 3.5 3.5 7.5 Projected P&L (bp) at maturity when Euribor is 70 11.0 25.5 2.0 31.0 16.5 -3.5 12.5 80 11.0 20.5 12.0 31.0 21.5 1.5 17.5 90 11.0 10.5 22.0 21.0 21.5 11.5 17.5 100 11.0 0.5 32.0 11.0 21.5 21.5 17.5 110 1.0 -9.5 22.0 1.0 11.5 11.5 7.5 120 -9.0 -19.5 12.0 -9.0 1.5 1.5 -2.5 130 -14.0 -24.5 7.0 -14.0 -3.5 -3.5 -7.5 65* 11.0 25.5 -3.0 26.0 11.5 -3.5 7.5 130** -14.0 -24.5 7.0 -14.0 -3.5 -3.5 -7.5
* Minimum Euribor fixing between June 2009 to March 2011, a period when the refi rate was 1%. ** Maximum Euribor fixing between June 2009 to March 2011, a period when the refi rate was 1%.
3 The period since 1 January 2009 does not include the FRA/OIS peak witnessed during the Lehman crisis, when front FRA/OIS widened to 150bp.
60
Exhibit 8: A year of flattening: the swap curve bear flattened early in the year on the back of ECB rate hikes, and bull flattened later due to escalation of the peripheral crisis
2Y swap yields and 2s/10s EUR swap curve; 1 January 2011 18 November 2011; % bp
Exhibit 9: Changes in swap curve directionality may be seen in the evolution of PCA factor loadings. Going forward, with the front end expected to remain tethered, we expect the swap curve to continue to flatten in a rally and steepen in a selloff
3M rolling 2Y and 10Y first factor loading obtained from PCA*; %
50% 40% 30% 20% Bear 10% 0% Jan 11 Mar 11 May 11 flattening
Bull flattening
2.60 2.40 2.20 2.00 1.80 1.60 1.40 1.20 Jan 11 Mar 11 May 11
2Y sw ap rate
180 160 140 120
10Y
2Y
2s/10s sw ap curv e
100 80
Jul 11
Sep 11
Nov 11
Jul 11
Sep 11
Nov 11
* Principal component analysis (PCA) is a tool that recombines observed variables into factors such that a smaller number of these factors explain a large percentage of variation observed in the original data set. We run PCA on demeaned levels of 1Y, 2Y, 3Y, 5Y, 7Y, 10Y, 12Y, 15Y, 20Y, 25Y, and 30Y swap yields.
A notable characteristic of the swap curve during 2011 has been frequent changes in curve directionality observed on the back of central bank activity and the peripheral crisis. The change in swap curve directionality may be seen in the evolution of PCA factor loadings (Exhibit 9). Early in the year, when the market was expecting the ECB to remain on hold, the curve was driven by the intermediate sector, flattening in rallies and steepening in selloffs. This resulted in higher first-factor loading on 10Y yields relative to 2Y yields. This dynamic changed in 2Q and 3Q as the ECB started to hike rates and the curve became driven by the front end. Consequently, the first factor loading on 2Y yields increased relative to 10Y. This directionality has flipped once again over the past three months, when intermediates started driving the curve. Going forward, we expect the ECB to eventually cut the refi rate to 0.50% in 1H12, while keeping the deposit facility rate at 0.25%. With the market already pricing much of the expected easing (see EONIA curve), we expect the front end to remain tethered and the curve to exhibit positive directionality in 2012, flattening in a rally and steepening in a selloff. 2s/10s flatteners have been profitable on average during 2011 and their P&L performance has been correlated with the carry at inception. Exhibit 10 shows the success ratio and the average 3M P&L on 2s/10s flatteners initiated at different levels of ex ante carry during 2011. The analysis shows that flatteners have performed better when initiated at positive carry, and
Exhibit 10: Over the past year, flatteners have been more profitable when initiated at high levels of carry; by this metric, flatteners currently appear attractive
Success ratio and statistics on P&L of 2s/10s flatteners held for three months relative to the 3M carry at inception*; 1 January 2011 18 November 2011; bp 3M Carry <-8bp >-8bp and <0bp >8bp
# trades # profitable trades Success Ratio Av erage Min Max SD
88 34 39% -3 -38 24 17
48 32 67% 7 -28 30 17
25 25 100% 29 8 53 15
* We split the sample of trades into three categories based on the top, the central, and the bottom third of carry level at trade inception.
Exhibit 11: Further escalation of the peripheral crisis is expected to put flattening pressure on the swap curve
2s/10s EUR swap curve regressed against 10Y weighted peripheral spreads*; 1 January 2011 18 November 2011; bp
180 y = -0.08x + 156.5 160 140 120 100 80 200 300 400 500 600 700 10Y w eighted peripheral spread; bp 800 18-Nov -11 R 2 = 50%
* Weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).
61
have averaged negative returns if implemented when the carry was in the bottom third of the sample. Another significant driver of the curve has been the peripheral sovereign crisis. We believe that this crisis will likely worsen in 1H12 (see Overview and Euro Cash), resulting in wider peripheral spreads. This will likely put flattening pressure on the Euro swap curve (Exhibit 11). Empirical analysis over the past year indicates that the 2s/10s swap curve tends to flatten about 8bp for every 100bp of widening in 10Y weighted peripheral spreads. We expect about 250bp of widening in 10Y weighted peripheral spreads (see Euro Cash), which will put around 25bp of flattening pressure on the 2s/10s swap curve. Additionally, 3M carry on 2s/10s flatteners is +5bp; we therefore recommend that investors implement swap curve flatteners. Next, we further analyse carry in the swap curve to find the most attractive sectors to implement flatteners. Since early 2011, carry in long positions at the front end of the curve has flipped from positive to negative as the fronts/reds curve has inverted. This has turned carry in flattening trades positive, especially when anchored at the very front end of the curve (Exhibit 12). On a carry-to-risk basis, we believe 1s/3s and 1s/5s are the most attractive flatteners on the swap curve. These trades have a strong bullish bias and sport a better risk profile than outright longs. Investors wishing to express a bullish view on the intermediate sector may also consider call structures.
Exhibit 13: or bullish option structures on intermediate tails
Cost and projected payoff of various bullish option structures on 5Y EUR swap rates; bp
Exhibit 12: and investors wishing to position for this should consider flatteners that offer attractive risk-adjusted carry
Risk-adjusted carry in outright long positions in swaps and swap curve flatteners; bp
Trade 1Y 2Y 3Y 5Y 10Y 15Y 30Y 1Y/2Y 1Y/3Y 1Y/5Y 2Y/5Y 2Y/10Y 2Y/30Y 5Y/10Y 1.62 1.60 1.71 2.09 2.68 2.96 2.82 -2 10 47 49 108 122 59 -8 -1 4 6 4 3 1 7 12 14 7 5 2 -2 3M Carry (bp) 18-Nov -11 18-Nov -11 30-Jun-11 04-Jan-11 14 12 13 11 7 6 3 -2 -2 -4 -1 -5 -9 -3 9 12 14 12 8 6 2 4 5 3 0 -4 -10 -4 Ann Risk* Current ann. risk (bp) 43 71 82 93 102 107 116 32 43 59 39 60 78 29 adj. carry ** -0.7 -0.1 0.2 0.3 0.2 0.1 0.0 0.9 1.1 0.9 0.7 0.4 0.1 -0.2
* Annualised risk is defined as sqrt(251)*3M standard deviation of daily changes. ** Annualised risk-adjusted carry is defined as 4*3M Carry/Annualised risk.
With our 5Y German benchmark yield target at 0.65% (see Euro Cash), we see value in receiver structures on 5Y swaps despite our view that 5Y swap spreads will widen to 130bp (see Swap spreads below). In Exhibit 13, we analyse the projected upside for various bullish structures implemented with 6M receivers. Among them, we prefer symmetric structures such as receiver butterflies or condors, given their limited downside. However, investors less concerned about 5Y swap rates declining below 1.50%-1.60% may consider asymmetric
6Mx5Y Swaptions; Carry = 13.2bp; ATMF = 222.1bp; Spot = 208.9bp; Spot 1Y Min = 184.1bp Ty pe Call spread Call spread 1x 2 1x 2 Ladder Ladder Ladder Ladder Butterfly Butterfly Condor Condor Strike ATM/ATM-1C ATM/ATM-2C ATM/ATM-1C ATM/ATM-2C ATM/ATM-1C/ATM-2C ATM/ATM-1C/ATM-3C ATM/ATM-1C/ATM-4C ATM/ATM-2C/ATM-4C ATM/ATM-1C/ATM-2C ATM/ATM-2C/ATM-4C ATM/ATM-1C/ATM-2C/ATM-3C ATM/ATM-1C/ATM-3C/ATM-4C Contracts +1/-1 +1/-1 +1/-2 +1/-2 +1/-1/-1 +1/-1/-1 +1/-1/-1 +1/-1/-1 +1/-2/+1 +1/-2/+1 +1/-1/-1/+1 +1/-1/-1/+1 Cost (bp) 6.1 11.2 -16.4 -6.2 -11.3 -7.1 -3.7 1.4 1.0 3.6 1.9 2.7 Projected pay off Yield bounds (bp)** at maturity (bp)* 13.2 13.2 13.2 13.2 13.2 13.2 13.2 13.2 13.2 13.2 13.2 13.2 Low er 179.3 163.2 171.3 162.3 152.4 144.3 196.7 172.9 184.4 172.0 Upper 216.1 211.0 238.5 228.3 233.4 229.2 225.8 220.7 221.2 218.5 220.2 219.4 Max P&L (bp) Upside 7.1 15.2 29.6 32.6 24.5 20.3 16.9 25.0 12.2 22.8 11.3 10.5 Dow nside 6.1 11.2 1.0 3.6 1.9 2.7
Note: ATM-1C refers to the swaption strike that is 1-carry (13bp) away from ATM swap rate. * As the forward rate is expected to slide to spot rate, the projected payoff at maturity is equal to the current slide. ** Yield levels above and below which the trade becomes unprofitable.
62
structures such as receiver ladders, which may be implemented at a credit. Another set of trades that have captured investor interest over the past year has been long-dated forward steepeners on the swap curve. We analyse the attractiveness of such trades by building a carry efficient frontier of steepener trades that: 1) offer high risk-adjusted carry, and 2) are trading flat from a historical perspective (Exhibit 14). In addition, given our view that the spot 2s/10s curve will flatten, we eliminate trades that have large positive correlation to the 2s/10s curve. The best trades are in the top right quadrant of Exhibit 14 which shows forward curves that are relatively flat and offer high risk-adjusted carry. Unfortunately, however, we found that most trades have their long leg anchored in the 20Y+ area of the curve. Since long dated forward curves are positively correlated with 10s/30s, which is itself driven by technical factors (see below), we remain cautious on long-dated forward steepeners and refrain from recommending trades in this sector. Long-dated swap curve The 10s/30s swap curve has been in a 25bp range since the beginning of the year and is now trading close to the bottom of the range. Various technical factors have, over the past five years, overwhelmed the typical relationship of the 10s/30s curve with front end rates. For example, in mid- and late-2008, the flattening in 10s/30s was driven by the hedging of non-inversion notes from the dealer community.4 Additionally, hedging flows from pension funds have frequently been an important driver of the 10s/30s curve. Over the past couple of years, however, 1-way CSA hedging has dominated long end flows due to the peripheral debt crisis.5 We develop a relative value model that captures the sensitivity of the 10s/30s curve to both macro and technical factors using the following variables: 1) 10Y weighted peripheral spread, 2) 2Y swap rate, 3) pension fund coverage ratio, and 4) dummy variables for hedging of non-inversion notes in mid-2008 and early 2009 (Exhibit 15). The betas of the regression model are significant and have the correct signs. The model indicates that 10s/30s tends to flatten:
Exhibit 14: Although flatteners appear attractive on the spot curve, steepeners appear attractive on the forward curve based on valuations and carry, especially when they have little correlation with the spot 2s/10s curve; however, we remain cautious
Risk-adjusted carry* of various forward steepeners** vs. its 5Y z-score of relative steepness*** (reverse axis); unitless
1.0 0.8 0.6 0.4 0.2 0.0 0.5
2Yx 5s/20s
3Yx 5s/20s 2Yx 10s/20s 3Yx 10s/20s 4Yx 5s/20s 5Yx 5s/20s 5Yx 10s/20s
0.0 -0.5 -1.0 Relativ e steepness; z-score
-1.5
* Risk-adjusted carry defined as 1Y carry divided by 2*SD of quarterly changes on the forward curve. ** Black dots indicate curve segments with correlation lower than 20% to 2s/10s; grey dots indicate curve segments with correlation higher than 20% to 2s/10s. *** 5Y z-score; negative numbers (right side of X-axis) indicate a flat curve. We include 1Y to 5Y forward steepeners on 5s/10s, 5s/20s, 5s/30s, 10s/20s, and 10s/30s.
10s/30s EUR swap curve regressed against 1) 10Y weighted peripheral spread*, 2) 2Y EUR swap yield, 3) pension fund coverage ratio, and 4) two dummy variables for the 31 May 2008 to 15 August 2008 and the 15 September 2008 to 15 January 2009 periods; past 5Y; bp
Exhibit 15: as technical factors such as 1-way CSA hedging may drive the 10s/30s swap curve flatter if the crisis gets worse
60 40 20 0 -20
y = -0.03*(Peri sprd)-13.0*(2Y)+42.6*(PF ratio) - 6.8(I1) - 28.3*(I2)+15.7 R 2 = 74%
18-Nov -11 -40 0 200 400 600 10Y w eighted peripheral spread; bp 800
* Weighted peripheral spread computed against Germany for Ireland, Portugal, Italy, Spain and Greece (weighted by the size of their outstanding bond markets).
4 See Global Fixed Income Markets Weekly, 25 April 2008 and 06 June 2008. 5 See Global Fixed Income Markets Weekly, 30 July 2010.
63
Exhibit 16: With the crisis spreading to Italy, the correlation between peripheral spreads and 30Y swap yields has become more negative on the back of technical flows such as 1-way CSA hedging. We have a 10s/30s swap curve flattening bias
2M correlation of daily changes in 10Y weighted peripheral spreads to daily changes in 30Y EUR swap yields; %
Exhibit 17: since we believe that 30Y swap yields could richen further if peripheral sovereign spreads continue widening
Residual of 10s/30s/50s EUR swap fly regressed against 30Y EUR swap yields vs. 10Y weighted peripheral spreads*; past 2Y; bp
10 5 0 -5 -10
0% -20% -40% -60% -80% -100% Jan 10 Apr 10 Jul 10 Oct 10 Jan 11 Apr 11 Jul 11 Nov 11
y = -0.02x + 6.26 R 2 = 54%
18-Nov -11 -15 0 200 400 600 10Y w eighted peripheral spread; bp 800
* See Exhibit 15 for definition of weighted peripheral spread.
3bp for every 100bp widening in the 10Y weighted peripheral spread, 13bp for every 1% increase in 2Y yields, 4.2bp for every 10% decline in pension fund coverage ratio, and 7bp and 28bp in the two episodes of hedging of noninversion notes of mid-2008 and early 2009, respectively. Anecdotal evidence suggests pension fund and exotic desk hedging has declined recently, leaving 1-way CSA hedging as the main driver of this segment of the curve. With peripheral spreads expected to widen significantly from current levels, we see risk of further flattening in 10s/30s. Specifically, we expect the 10Y weighted peripheral spread to widen 250bp, putting around 68bp of flattening pressure on the 10s/30s swap curve. Indeed, the correlation between weighted peripheral spreads and 30Y swap yields has become more negative in 2H11, as the peripheral crisis has spread to Italy (Exhibit 16). Thus, our negative view on peripherals leads us to a 10s/30s swap curve flattening bias. 1-way CSA hedging on the swap curve has richened the 30Y sector. Indeed, 30Y is trading rich on a level neutral 10s/30s/50s fly. However, historical analysis indicates that 30Y richness is a function of peripheral stress, with the residual of the 10s/30s/50s level neutral fly being strongly correlated with peripheral spreads (Exhibit 17). Given our negative view on the crisis, we believe that the
64
Exhibit 18: Implied curve directionality has frequently underestimated delivered directionality in 1s/10s, making conditional structures attractive in the past
Implied* and 3M forward looking delivered directionality** of 1s/10s EUR swap curve; %
100% 80% 60% 40% 20%
Deliv ered directionality
Implied directionality 0% Jan 10 Apr 10 Jul 10 Oct 10 Jan 11 Apr 11 Jul 11 Nov 11
* Implied directionality for swap curve calculated as ratio of implied volatility of 3MxLong leg/3MxShort leg -1. ** Delivered directionality for swap curve calculated as 1 - 3M beta of daily changes in Short leg regressed against daily changes in Long leg.
30Y sector could richen further and we recommend that investors do not fade the move. Implied and delivered curve directionality With the ECB unlikely to cut rates below 0.50%, we believe the curve will remain driven by the intermediate sector, flattening in rallies and steepening in selloffs. During 2011, this directionality was not fully priced in the curve and offered several opportunities for
Exhibit 19: especially since delivered directionality is autocorrelated
Spread of ex post 3M delivered and implied 1s/10s curve directionality* regressed against the spread of ex ante 3M delivered and implied 1s/10s curve directionality; past 2Y; %
Exhibit 20: Implied directionality is below delivered directionality for curve trades anchored at the very front end, making conditional curve trades attractive
Current implied and delivered directionality* for various curve trades; % Trades Implied Deliv ered Implied-Deliv ered
80% 60% 40% 20% 0% -20% -20%
y = 0.76x + 0.08 R 2 = 57%
1s/2s
18-Nov -11
1% 21% 47% 65% 20% 46% 63% 22% 36% 12%
42% 59% 67% 73% 28% 41% 51% 12% 26% 14%
-41% -38% -19% -8% -8% 5% 12% 10% 9% -2%
1s/5s 1s/10s 1s/30s 2s/5s 2s/10s 2s/30s 5s/10s 5s/30s 10s/30s
0% 20% 40% 60% 3M lagged (deliv ered - implied) directionality ; %
80%
* See Exhibit 18 for definition of implied and delivered directionality.
* See Exhibit 18 for definition of implied and delivered directionality. Y variable: Delivered directionality at time t minus implied directionality at time t 3M. X variable: Delivered directionality at time t 3M minus implied directionality at time t 3M.
Exhibit 21: Although the primary driver of swap spreads in 2011 was risk aversion,*
10Y German b/m swap spreads vs. DJ Euro Stoxx 600 index (inverted); since 1 January 2011 18 November 2011; bp points (inverted axis)
80 10Y ASW 70 60 50 40 30 20 Jan 11 Mar 11 May 11 Jul 11 Sep 11 Nov 11 Euro Stox x 600 (inv erted ax is)
200 220 240 260 280 300
conditional curve trades. For example, implied directionality in the 1s/10s swap curve has been, until recently, in a 20-50% range whereas delivered directionality has rarely fallen below 50% (Exhibit 18). Moreover, since delivered directionality is autocorrelated (Exhibit 19), investors may use historical delivered directionality as a measure of future expected delivered directionality, thereby providing them with a way to judge the attractiveness of conditional curve trades at any point in time. For example, implied directionality is significantly below delivered directionality for curve trades anchored at the very front end (1Y), making such conditional curve trades attractive. However, trades anchored further out (2Y+) are unattractive due to recent richening of implied directionality in such structures (Exhibit 20). Therefore, we favour bull flatteners in 1s/2s and 1s/5s, while avoiding other structures.
* The correlation between DJ Euro Stoxx 600 index and 10Y swap spreads since the beginning of the year is around 95%.
Swap spreads
Like most asset classes, swap spreads were held hostage to market risk aversion stemming from the burgeoning peripheral debt crisis (Exhibit 21). After staying in a narrow 20-35bp range in the first half of the year, 10Y swap spreads embarked on a widening trend in the second half in response to a sharp deterioration in risk markets, reaching a peak of 75bp. Subsequently, spreads
declined below 50bp in October as equity markets rallied, before widening sharply in response to the threat of a Greek referendum and significant political/market stress in Italy. Since the end of last year, 2Y, 10Y and 30Y swap spreads have widened 40, 45 and 15bp, respectively. Historically, we have used 5 factors to explain 10Y swap spread movements: German government bond issuance, Amount of liquidity in the banking system, proxied by EONIA rates, Directionality to 10Y Bund yields,
65
Risk aversion, proxied by swaption implied volatility, and Swapped issuance. Much of the widening this year, however, may be explained by sovereign spreads. Indeed, a regression of 10Y swap spreads against 10Y weighted peripheral spreads over the past year shows an R-squared of 80%, up from just over 40% in 2010. This suggests that, much like every other asset class, swap spreads are living in a single-risk factor world. Although swap spreads appear to be driven solely by sovereign risk, in actuality it is impossible to distinguish between this (relatively) new driver of swap spreads versus traditional drivers. This is because sovereign risk is highly correlated with typical high frequency drivers of swap spreads. Indeed, as peripheral spreads have widened, Bund yields have fallen, implied volatility has shot up, and swapped issuance has plummeted. Since the beginning of the year, the correlation between peripheral spreads and these three drivers is 70-90% (Exhibit 22). Given these high correlations, the change in 10Y swap spreads since end-2010 can be explained reasonably well with traditional drivers. Between December 2010 and November 2011, for example, 10Y German b/m swap spreads (averaged over a 1M period) widened 26bp, of which 25bp can be explained by high frequency factors (Exhibit 23).
Exhibit 22: high cross correlations make it is difficult to disentangle the effect on swap spreads of peripheral sovereign spreads versus more traditional drivers
Correlations between traditional high frequency drivers of swap spreads and 10Y weighted peripheral spreads; data since 1 January 2011; %
3Mx 10Y 10Y sw ap 10Y Bund spreads 10Y sw ap spreads 10Y Bund y ield 3Mx 10Y implied v ol Sw apped issuance* 10Y peripheral spread** 100% -95% 91% -77% 90% 100% -96% 72% -93% 100% -71% 86% 100% -68% 100% y ield implied v ol Sw apped issuance* 10Y peripheral spread**
* We use fixed-rate, -denominated issuance by financial institutions (including covered bonds) and supras/agencies, plus one-half of corporate bond issuance, as an indicator of potential swapped issuance. ** Weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).
Going forward, we expect swap spreads to continue to be driven by the peripheral debt crisis which is likely to get worse before it gets better (see Overview). Over the past 2 years, 10Y swap spreads have widened an average of 9bp for each 100bp widening in 10Y weighted peripheral spreads (Exhibit 24). Given our expectation that 10Y weighted peripheral spreads are likely to widen from their current level of 700bp to a peak of 955bp in 1H12, we believe that 10Y swap spreads will widen to a
Exhibit 23: and, indeed, much of the swap spread widening in 2011 may be explained by lower Bund yields and higher implied volatility, which themselves reflected an escalation of the peripheral debt crisis
Return attribution of 10Y German b/m swap spreads between 30 December 2010 and 18 November 2011*; bp 20-day MA as of:
18-Nov -11 10Y sw ap spread; bp Low freq regression; 1 Jan 00 - 1 Jan 08 IMM1 EONIA; % 1Y sum of German gov ie issuance**; bn Exp. chg from low freq factors High freq regression; 1 Jan 08 - Present Yields (10Y Bund); % 3Mx 10Y sw aption v olatility ; bp/day 20-day MA of sw apped issuance*** (bn/day ) Exp. chg from high freq factors; bp Exp. chg from all factors; bp 1.91 7.5 1.3 2.97 6.0 0.5 -1.07 1.5 0.8 -19.0 4.7 -2.6 20 7 -2 25 26 0.56 170 0.73 194 -0.17 -24 11.0 -0.12 -2 3 1 62 30-Dec-10 36 Chg 26 Beta Ex p. Chg
* J.P.Morgans low frequency model regresses 10Y swap spreads against first IMM date EONIA and gross annual issuance of German government bonds over the 8Y period ending 1 January 2008. Residual from this low-frequency regression is computed over the period 1 January 2008 present and regressed against the three high frequency factors. ** We use an exponential model of govie issuance as an explanatory variable because the sensitivity of swap spreads declines with increasing govie issuance. For simplicity, however, we show the overall (linear) sensitivity of spreads to issuance in the table, at current issuance levels. *** We use fixed-rate, -denominated issuance by financial institutions (including covered bonds) and supras/agencies, plus one-half of corporate bond issuance, as an indicator of potential swapped issuance.
66
Exhibit 24: Going forward, we expect swap spreads to continue to be driven by the peripheral debt crisis, which has caused 10Y swap spreads to widen an average of 9bp for each 100bp widening in peripheral sovereign spreads
10Y German b/m swap spreads regressed against 10Y weighted peripheral spreads*; past 2Y; bp
Exhibit 25: and, given our view that the peripheral debt crisis will get worse before it gets better, we believe that swap spread widening has more to go
J.P.Morgan swap spread forecast*; bp
10Y w td peripheral** spread 2Y Sw ap spreads 5Y Sw ap spreads 10Y Sw ap spreads 30Y Sw ap spreads 18 Nov 11 699 110 96 67 19 1Q12 835 130 115 80 25 2Q12 955 145 130 90 30 3Q12 900 135 120 85 25 4Q12 855 130 115 80 25
80 60 40 20 0 0
y = 0.09x + 7.7 R 2 = 82%
* We use the model in Exhibit 24 above and our projections on 10Y weighted peripheral spreads (see Euro Cash) to forecast 10Y swap spreads. Swap spreads at other maturities (such as 2Y) are projected by assuming that they widen proportionally based on their widening relative to 10Y swap spreads over the past 2 years. ** See Exhibit 22 for definition of weighted peripheral spreads.
200 400 600 10Y w eighted peripheral spread*; bp
800
Exhibit 26: Although correlations between various spreads have skyrocketed as the peripheral debt crisis has intensified
10Y weighted peripheral spreads* versus % of variability explained by the first factor in a rolling 6M PCA** on various spreads***; bp %
* See Exhibit 22 for definition of weighted peripheral spreads.
800
% v ariability ex plained by 1st PCA factor
100 90
peak of 90bp. Thereafter, as the peripheral debt crisis starts to wane, we expect 10Y swap spreads to narrow to around 80bp by end-2012. We also expect 2Y, 5Y and 30Y swap spreads to widen concurrently with 10Y spreads, in line with their historical pattern. For example, over the past two years, 2Y spreads have widened around 60bp while 10Y spreads have widened 40bp. This suggests a ratio of 1.5 between 2Y and 10Y spreads. We use this ratio to project 2Y spreads. Exhibit 25 shows our swap spread forecast. Given that we believe the sovereign debt crisis will continue to worsen in 1H12, swap spreads across the curve are likely to widen, with the front end being the most susceptible to spread widening. We therefore recommend that investors position for wider swap spreads in anticipation of further peripheral stress, especially in the front part of the swap spread curve. Correlations between various spread markets have skyrocketed as the peripheral crisis has intensified. Exhibit 26 shows 10Y weighted peripheral spreads versus the % of variability explained by the first factor in a rolling 6M PCA6 performed on various spread
700 600 500 400 300 200 Jan 11 Mar 11 May 11 Jul 11 Sep 11 Nov 11 Weighted peripheral spread
80 70 60 50 40
* See Exhibit 22 for definition of weighted peripheral spreads. ** Principal component analysis (PCA) is a tool that recombines observed variables into factors such that a smaller number of these factors explain a large percentage of variation observed in the original data set. *** We run PCA on the correlation matrix of spread levels shown in Exhibit 27 below (not including bank CDS and Euro Stoxx bank index).
markets.7 The % of variability explained by the first factor has jumped from a low of 50% to its current level of around 90%, suggesting that spreads have increasingly been driven by a single factor sovereign risk. Given this dynamic, investors who are bearish on the peripheral crisis should consider positioning for wider spreads in markets which have lagged the recent widening in peripheral spreads, and are trading far from their peak levels. Such spreads have more
7 We run PCA on the correlation matrix of spread levels shown in Exhibit 27 below (not including bank CDS and Euro Stoxx bank index).
6 Principal component analysis (PCA) is a tool that recombines observed variables into factors such that a smaller number of these factors explain a large percentage of variation observed in the original data set.
67
Exhibit 27: some spread markets remain further away from their local and Lehman peaks than others. Investors who are bearish on the peripheral crisis, should consider risk-aversion trades where levels are far from peak levels achieved during periods of stress. On this basis, OIS swap spread wideners appear attractive, followed by 3s/6s basis wideners
Average distance of various spreads from local and Lehman peaks; %
18 Nov 11 1 2 3 4 5 6 7 8 9 10 IMM2 FRA/OIS IMM6 FRA/OIS 2Y Libor sw ap spreads 10Y Libor Sw ap spreads 2Y OIS sw ap spreads 10Y OIS sw ap spreads 1Y 1s/3s basis 1Y 3s/6s basis 1Y EUR/USD Fx basis 2Y EUR/USD Fx basis Average 11 12 13 10Y w td peri spread Bank CDS*** Euro Stox x Bank Index 699 312 125 353 123 193 739 322 119 322 90 77 60 110 67 18 13 33 29 -87 -72 01 Jun 11 24 28 53 33 10 -2 14 15 -26 -26 Peak level* Since 1-Jun-11 77 60 112 78 25 35 33 35 -88 -73 LEH times 93 62 118 82 41 40 54 44 -123 -87 0% 0% 2% 24% 46% 59% 1% 26% 2% 3% 16% 10% 5% 8% Distance from peak** Since 1-Jun-11 LEH times 24% 8% 11% 30% 74% 63% 53% 51% 38% 25% 38% 5% 34% Av erage 12% 4% 7% 27% 60% 61% 27% 39% 20% 14% 27% 10% 5% 21%
* Peak level between 1 June 2011 and 18 November 2011 and past 5Y to capture Lehman peaks. ** Distance from peak defined as (Peak minus current level) / (Peak minus 1 June 2011 level). *** Average of 5Y CDS on Deutsche, UBS, BNP, Barclays, Santander, and Unicredit.
potential to widen than those that are trading at, or close to, their wides. Exhibit 27 shows how far various market variables are, on average, from their local and Lehman peaks.8 We examine FRA/OIS basis, Libor and OIS swap spreads, interest rate bases, FX bases, sovereign spreads, European bank CDS, and European bank equities. Markets that appear to be farthest away from their local/Lehman peaks are OIS swap spreads and 3s/6s basis. We therefore recommend that investors consider positioning for a widening of these spreads. On the other hand, FRA/OIS, sovereign spreads and European bank CDS spreads are close to their wides; investors who believe that the sovereign debt crisis will subside should consider positioning in these markets. As the sovereign debt crisis has progressed, the options market has frequently underpriced swap spread directionality. As Exhibit 28 shows, the implied directionality of 10Y swap spreads has consistently stayed above delivered directionality. This is because markets have failed to account for the fact that swap spread counter directionality increases significantly during stressed environments as investors rush into the
8 Distance from peak is defined as (Peak minus current level) / (Peak minus 1 June 2011 level). 1 June 2011 was roughly the start of this leg of the sovereign debt crisis.
Exhibit 28: As the sovereign debt crisis has progressed, the options market has consistently underpriced swap spread directionality
2-week MA of implied directionality of front Bund swap spreads versus 3M delivered directionality*; past 1Y; %
10 0 Implied directionality -10 -20 -30 Deliv ered directionality -40 Jan 11 Mar 11 May 11 Jul 11 Sep 11 Nov 11
* Implied directionality computed as (implied vol on maturity-matched swaption / implied vol on Bund) 1. Delivered directionality computed as rolling 3M beta of daily change in swap spread regressed against daily change in CTD spot yield. Contract is rolled 10-days before option expiration.
Bund contract (Exhibit 29). Since market stress has increased almost monotonically in the past few months, swap spread directionality has increased concurrently (that is, become more negative). Given that we expect market stress to increase in 2012, we recommend
68
Exhibit 29: as markets have failed to take into account increasing counter directionality of swap spreads in stressed environments. Going forward, we recommend conditional swap spread trades as a less risky alternative to outright trades
Exhibit 30: Despite heightened uncertainty and peripheral stress, 2011 turned out to be a mediocre year for outright long EUR gamma positions
3M delivered directionality* of 10Y swap spreads versus 20D MA of 10Y weighted peripheral spreads (inverted axis); % bp (inverted axis)
Average 1M returns* and annualised information ratio (IR) from buying EUR 3Mx10Y swaption straddles; bp of notional information ratios
40 20
-5 -10 -15 -20 -25 -30 -35 Jan 11 Mar 11 May 11 Deliv ered directionality
Weighted peripheral spread (inv erted ax is)
200 300
Av erage (left)
4 2 0 -2 -4
0
400
-20
500
-40
600
-60
700 Jul 11 Sep 11 Nov 11
Information ratio (right) 2001 2003 2005 2007 2009 2011
-6
* Delivered directionality computed as rolling 3M beta of daily change in swap spread regressed against daily change in CTD spot yield. Contract is rolled 10days before option expiration.
* Trades entered daily and held for 1M. Returns calculated using J.P.Morgan volatility indices, which assume daily delta-hedging and zero transaction cost. Options are re-struck at the beginning of each month.
Exhibit 31: as positive returns from buying gamma during riskoff periods was offset by losses during risk-on episodes
Average* cumulative return** from buying 3Mx10Y EUR swaption straddles; bp of notional
conditional bull wideners as a less risky alternative to outright swap spread widening trades.
60
Euro volatility
Much like 2010, 2011 was not a year to buy swaption gamma since such a strategy would have yielded flat returns with large P&L swings (Exhibit 30). On average, losses from buying volatility during risk-on episodes offset gains collected during risk-off episodes, resulting in high P&L volatility (Exhibit 31). We analyse returns from buying 3Mx10Y swaption gamma during various risk-on and risk-off episodes from January 2010, a period during which the sovereign debt crisis dominated market sentiment. Exhibit 32 lists the date ranges used for this analysis. We find that, on average, buying gamma during risk-off periods is indeed profitable.9 For instance, buying 3Mx10Y gamma has returned an average of 40bp of notional during these risk9 We note that there are episodes when volatility markets initially do not respond in an expected manner at the onset of a risk-off regime. Nevertheless, on the back of continued stress, they eventually do catch up, resulting in positive returns to long volatility positions. For example, during the risk-off episode in May 2011, volatility markets were unmoved by discussions surrounding the Greek PSI even though equity markets were trended lower. Long gamma positions resulted in losses during this period. However, volatility shot up as soon as concerns surrounding Italy came to the fore in July.
40 20 0 -20 -40 -60 -80 0% 20% 40% 60% 80% % day s into risk on/off period; %
Risk-off
Risk-on
100%
* Averaged over risk-on/off episodes (see Exhibit 32 below for definition of dates). ** See Exhibit 30 for definition of returns.
off episodes. On the other hand, buying volatility during risk-on episodes has resulted in an average loss of 75bp of notional (Exhibit 31). Drivers of volatility in 2011 The sovereign debt crisis and the ECB were the dominant drivers of the volatility surface in 2011 (Exhibit 33). Volatility in the belly of the curve headed lower in 1Q11 as markets expected the peripheral crisis
69
Exhibit 32: Risk-on/off episodes from January 2010 to present
Start and end dates for various risk-on/off episodes since 1 January 2010, and % change in Euro Stoxx50 between start and end dates (%); Risk-on
Exhibit 33: Implieds on longer tails were boosted by the sovereign debt crisis, whereas implieds on shorter tails were negatively impacted by high excess liquidity
EUR 3Mx10Y and 3Mx2Y implied volatility versus 10Y weighted peripheral spreads*; 1 January 2011 18 November 2011; bp/day bp
Date Start 05-Feb-10 End 15-Apr-10 # months 2.3 2.7 1.6 1.6 2.1 Risk-off 04-Jan-10 18-Feb-11 28-Oct-11 05-Feb-10 16-Mar-11 18-Nov -11 Av erage 1.1 1.3 0.9 4.5 0.7 1.7 3018 3013 3068 3009 2462 Start 2632 2651 2721 1995
Euro Stoxx 50 End 3013 3068 3009 2462 % change* 14% 16% 11% 23% 16%
30-Nov -10 18-Feb-11 16-Mar-11 02-May -11 12-Sep-11 28-Oct-11 Av erage
3Mx 10Y
Peripheral spread
800 700
7 600 6 500 400 5 3Mx 2Y 300 200 Mar 11 May 11 Jul 11 Sep 11 Nov 11
2632 2558 2721 1995 2237
-13% -15% -11% -34% -9% -16%

4 Jan 11
15-Apr-10 24-May -10 02-May -11 12-Sep-11
* % Change of Euro Stoxx 50 defined as (End level Start level) / Start level.
* 10Y weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).
to be contained to Greece, Ireland and Portugal. However, as discussions around Greek PSI gained momentum, implied volatility inched higher and finally spiked up as contagion spread to Italy and Spain. The front end of the volatility curve was, however, equally impacted by both economic data and the peripheral crisis. An improving outlook for global growth and rising inflation expectations led the ECB to start hiking policy rates in April. Higher rates and the escalating peripheral crisis drove front end volatility higher. Later on, as the Euro area economy started to deteriorate, the ECB reacted by extending extraordinary liquidity measures and eventually cutting policy rates in November. At that point, front-end volatility decoupled from peripheral stress as it became clear that the ECB was likely to remain in a low-for-long mode. Trade the belly of the volatility curve from the long side Our baseline view for 2012 is that the peripheral crisis will worsen leading to higher implied volatility in the belly of the curve (see 2012 targets below). While implied volatility will likely increase in tandem with peripheral spreads, delivered volatility in the belly of the curve is also expected to remain high for two reasons. First, event risk remains high. Even if policy makers come up with a resolution to the crisis, implementation risks of such measures will remain high. Second, bank funding pressures and thin markets will support delivered
70
Exhibit 34: Long EUR volatility returns have been driven by the level of implieds and peripheral spreads at trade initiation
Rolling 1M return* from buying EUR 3Mx10Y volatility regressed against 1M lagged levels of 1) EUR 3Mx10Y implied volatility, 2)10Y weighted peripheral spread**, and 3) 2s/10s EUR swap curve; 1 February 2011 18 November 2011; bp of notional
-400 -450 -500 -550 -600 -650 -700 4
y = -52.3*(3Mx 10) + 0.67*(Peri sprd) + 2.1*(2s/10s) - 207 R 2 = 68%
5 6 7 1M lagged 3Mx 10Y implied v ol; bp/day
* See Exhibit 30 for definition of returns. ** See Exhibit 33 for definition of weighted peripheral spreads.
volatility. Thus, we recommend trading the belly of the volatility curve from the long side in 2012. The ex-ante level of 1) implied volatility, 2) weighted peripheral spread, and 3) 2s/10s swap curve were the best predictors of volatility returns in 2011. Buying gamma at lower levels of implieds yielded better returns. Similarly, buying gamma when peripheral spreads are high and the 2s/10s curve is steep resulted in better
returns. The intuition behind this is that wide peripheral spreads and a steep curve results in high subsequent delivered volatility, leading to improved returns. Together these three factors explain about 2/3rds of the total variability in returns in 2011 (Exhibit 34). Using the above relationship we develop a framework for identifying trading opportunities in 2012 (Exhibit 35). Over the past few months, the 2s/10s swap curve has exhibited a strong relationship to peripheral spreads (see Exhibit 11). Thus, in our model, we replace the 2s/10s curve with peripheral spreads. We then use the modified 2-factor model to calculate the level of implied volatility that generates flat returns over the next month given current level of peripheral spreads. We find that buying gamma when the combination of implied volatility and peripheral spreads is below the -1 standard deviation line is generally profitable. Strong sell signals are rarely generated because of the overall bias towards higher volatility and wider peripheral spreads over the past few months. For 2012, we therefore recommend initiating gamma trades if the combination of implied volatility and peripheral spread is outside the 1 standard deviation line. With the peripheral debt crisis expected to escalate, and 10Y weighted peripheral spreads targeted to widen by 250bp, we target implied volatility on 3Mx10Y at 9.4bp/day (Exhibit 36).10 Fade volatility flare-ups at the front end in 2012 With excess liquidity high and the ECB expected to be on perma-hold starting in 2H12 (see EONIA curve), front-end yields are likely to stay rangebound, leading to a low volatility environment (Exhibit 37). We, therefore, target 3Mx2Y implied volatility at 4bp/day and recommend short gamma positions on 2Y tails. A risk to this outlook comes from FRA/OIS volatility which may stay elevated due to bank funding pressures (see Swap spreads). Front-end volatility is likely to flare-up during episodes of FRA/OIS widening. We recommend fading such moves.
Exhibit 35: leading us to recommend buying volatility when the combination of implied volatility and peripheral spreads is 1SD below the zero iso-return line
3Mx10Y EUR volatility versus 10Y weighted peripheral spreads* at initiation of long gamma trades in 2011; solid line is the zero iso-return line and light lines are 1SD lines**; bp/day
10 9 8 7 6 5 4 300 Stay neutral
Iso-return line <0 returns >0 returns Current
Sell v olatility
18-Nov -11 Buy v olatility
400 500 Weighted peripheral spread; bp
600
700
* See Exhibit 33 for definition of weighted peripheral spreads. ** Based on regression beta and standard errors in Exhibit 34. Note: we reduced our 3-factor model to a 2-factor model as follows. The 2s/10s swap curve has exhibited strong directionality with peripheral spreads over the past few weeks (2s/10s swap curve = -0.0014*(10Y wtd peripheral spread) + 1.94; 8 August 201118 November 2011; R2 = 71%). We use this relationship to replace the 2s/10s curve with peripheral spreads. We then generate the zero iso-return line based on implied volatility and peripheral spreads alone.
Exhibit 36: EUR implied volatility targets for 2012
EUR implied volatility targets for 1Q12 and 2Q12*; bp/day 18 Nov 11 1Q target 2Q target 2Q target - current
3Mx 1Y 3Mx 2Y 3Mx 5Y 3Mx 10Y 3Mx 30Y 2Yx 2Y 2Yx 5Y 2Yx 10Y 5Yx 5Y 5Yx 10Y 10Yx 10Y
5.4 5.5 6.6 8.0 8.9 5.9 6.1 6.4 5.6 5.6 5.0
4.0 4.5 6.6 8.7 9.8 5.5 6.0 6.6 5.5 5.7 5.0
3.5 4.0 6.7 9.4 10.6 5.2 6.1 6.7 5.6 5.7 5.0
-1.9 -1.5 0.1 1.4 1.7 -0.7 0.0 0.4 0.0 0.0 0.0
* Target for 3Mx10Y is based on its relationship with the 10Y weighted peripheral spread and our forecast of this spread in 2012 (see Euro Cash). For 30Y, we assume that the beta between 30Y and 10Y remains at 120%. 2Y implieds in 2Q is the average implied of 3Mx2Y for 1H10 a period when the ECB was firmly on hold and excess liquidity was high. 5Y is computed as an average of 2Y and 10Y.
10 We use the following model to arrive at our target for 10Y implied volatility: 3Mx10Y = 0.0061*(10Y weighted peripheral spread) + 3.56; R-sqr = 73%; 1 Jan 2010 18 November 2011.
71
Exhibit 37: Look for delivered volatility at the front end to decline due to exceedingly high excess liquidity; fade volatility flare-ups in 2Y tails
1M delivered volatility on 3Mx2Y EUR swaps regressed against 1M delivered volatility of 3Mx3M EONIA; 1 January 2009 18 November 2011; bp/day
Exhibit 38: High excess liquidity is also expected to keep front end rates in a range,
EUR 2Y swap yield; %
1.65 1.60 1.55 1.50 1.45 1.40 1.35 05-Aug 26-Aug 16-Sep 07-Oct 28-Oct 18-Nov
8 7 6 5 4 3 2 1 0
y = 0.71x + 1.78 R = 64%

2
18-Nov -11
2 3 4 5 1M deliv v ol of 3Mx 3M EONIA; bp/day
Exhibit 39: and we recommend enhancing returns of a short volatility position in 2Y tails by delta-hedging infrequently
Statistics on 1M rolling returns* from selling 3Mx2Y EUR swaptions using various time-based and delta-based hedging strategies; 8 August 2011 18 November 2011; bp of notional Av erage SD Information Ratio
With front end swap yields range-bound (Exhibit 38), mean reversion may be an attractive source of return as delta-hedging short volatility positions infrequently could boost performance. Indeed, 2Y swap yields have been about 30% less volatile measured on a periodic basis than on a daily basis.11 Exhibit 39 shows statistics on 1M rolling returns from selling 3Mx2Y EUR gamma using time-based and delta-based hedging strategies. We draw the following conclusions from this analysis. First, delta-hedging less frequently provides higher riskadjusted returns than daily delta-hedging. Second, resetting hedges based on a delta threshold is better than using a time-based rule. Thus, going into 2012, we recommend infrequent delta hedging of short volatility positions at the front end of the volatility curve. Stay neutral on ultra long tails The 30Y sector remains the forbidden fruit of European volatility markets because volatility on 30Y tails has historically been impacted by technical factors such as non-inversion note hedging, pension fund receiving, and 1-way CSA hedging. Due to these technical factors, implied volatility on 30Y tail tends to increase in excess of 10Y tails whenever we encounter a crisis (Exhibit 40).
Days based hedging rule Daily Weekly Fortnightly 9.3 16.3 18.0 8.8 8.6 10.8 3.7 6.5 5.8
Delta threshold based hedging rule 10% 30% 50% 9.5 15.0 20.2 8.8 9.9 9.6 3.7 5.2 7.3
* See Exhibit 30 for definition of returns.
Exhibit 40: Implieds on 30Y tails have lost some of their sensitivity to 10Y tails as technical factors have declined; we expect technicals to remain a driver of 30Y tails, however, and suggest a neutral view in this sector
Rolling 3M beta from regressing daily changes in EUR 3Mx30Y implied volatility against daily changes in EUR 3Mx10Y implied volatility; unitless
2.0 1.8 1.6 1.4 1.2 1.0 0.8 Non-inv ersion note hedging 1-w ay CSA hedging Current
For instance, 1M delivered volatility of 2Y swaps measured on 5D changes is currently at 3.5bp/day (after adjusting by sqrt(5)) while that measured on daily changes is at 5bp/day.
72
11
0.6 Jan 08 Oct 08 Jul 09 Apr 10 Jan 11 Nov 11
The sensitivity of 30Y tail volatility to 10Y tail volatility appears to have declined somewhat over the past few months. This may be because some of the technical flows supporting 30Y tails have declined recently (for example, pension funds refuse to receive at these low levels of yields). Unfortunately, however, other hedging flows such as 1-way CSA flows have continued to drive 30Y yields lower (and volatility higher). Going forward, it is difficult to judge the evolution of these technical flows and therefore we recommend staying away from 30Y tails. Buy Bund volatility vs. swaption volatility as a source of relative value 2011 has been a year of govie volatility. Regular bouts of flight-to-quality have helped long gamma positions in govie options produce high risk-adjusted return, especially relative to swaption gamma. Exhibit 41 shows the average monthly returns and annualised information ratio from several such trades. We draw three main conclusions from this analysis. First, buying govie gamma outright was profitable across the curve, with Bunds outperforming Bobl and Schatz. Second, buying govie gamma vs. maturity-matched swaption gamma was also profitable. Here again a strategy involving Bunds was the clear winner. Finally, returns from buying govie gamma vs. swaption gamma were more stable than buying outright govie gamma, i.e., they yielded superior information ratios. Swap spread directionality was consistently negative in 2011 and generally exceeded implied directionality (see Swap spreads). This disparity, which was most prominent in Bunds, was the primary reason for the outperformance of govie vs. swaption gamma. We expect this disparity to continue in 2012 and recommend buying Bund gamma vs. swaption gamma as an efficient way to position for an escalation of the peripheral crisis. To help generate a trading signal, we develop a framework akin to the one in Exhibit 35. As a first step, we estimate that 60% of the variability in returns is explained by peripheral spreads and implied volatility differential (defined as govie implied minus swaption implied) at trade initiation (Exhibit 42). The regression signs are intuitive; buying govie vs. swaption gamma at low implied volatility differential and wide peripheral spreads is attractive. This is because wide peripheral spreads indicate stress in the system which typically
Exhibit 41: Buying govie volatility vs. maturity matched swaption volatility was a profitable strategy in 2011. Implementing the strategy in Bunds would have resulted in higher P&L and information ratio than buying Bobl or Schatz volatility outright
Average 1M returns* (bp of notional) and annualised information ratio (unitless) from buying govie volatility and selling a gamma-equivalent amount of maturity matched swaption volatility, vs. buying volatility outright; all trades scaled to have the same gamma as Bund; 1 January 2011 18 November 2011; bp of notional Information ratios
18 15 12 9 6 3 0 Bund Bobl Schatz Bund Bobl outright Schatz
3.0 2.5 2.0 1.5 1.0 0.5 0.0
Long gov ie v s. mat matched sw pt

Rolling 1M returns* from buying Bund volatility vs. selling a gamma equivalent amount of maturity matched swaptions regressed against 1M prior levels of 1) 10Y weighted peripheral spreads, and 2) implied volatility differential; 1 January 2011 18 November 2011; bp of notional
Exhibit 42: Outperformance of govie vs. swaption volatility can be explained by the implied volatility differential and peripheral spreads at trade initiation
120 100 80 60 40 20 0 -20 200
y = 0.23*(Peripheral sprd) - 24.4*(Imp v ol spread) - 57.2 R 2 = 63%
250
300 350 400 450 500 550 1M ago 10Y w td peripheral spreads; bp
600
650
results in high subsequent delivered volatility differential as asset swaps become increasingly counter directional.
73
Exhibit 43: suggesting that investors should buy govie vs. swaption volatility when the combination of implied volatility differential and peripheral spreads is 1SD below the zero isoreturn line
Exhibit 44: Vega outperformed gamma in 2011 as low delivered volatility was a drag on the latter
Cumulative returns from buying 100mn notional EUR 3Mx10Y vs. 5Yx10Y swaption straddles; 1 January 2011 18 November 2011; bp of notional
Implied volatility spread* vs. 10Y weighted peripheral spread; solid line shows the zero iso-return line and light lines are 1SD** lines; table shows statistics of trades initiated in 2011***; 1 January 2011 18 November 2011; bp/day
300 5Yx 10Y 200 100 0 3Mx 10Y
5 4 3 2 1 0 -1 200 1SD line
Av erage Iso-return line IR <0 returns % profitable >0 returns Current
22 3 85%
18-Nov -11
-100 -200 Jan 11

300 400 500 600 10Y w eighted peripheral spread; bp 700
Mar 11
May 11
Jul 11
Sep 11
Nov 11
Light (dark) dots represent trades that were profitable (unprofitable) in 2011 over the next month, without considering the trading rule discussed here. * Bund implied maturity-matched swaption implied volatility; futures rolled 20D before option expiry. ** Based on regression beta and standard errors in Exhibit 42. *** 1M statistics of trades entered when the combination of implied volatility differential and peripheral spreads was below the zero iso-return line.
Exhibit 45: Vega has exhibited a concave profile to gamma. Thus, despite our bullish view on gamma, we stay neutral on vega
5Yx5Y EUR implied volatility regressed against 3Mx5Y EUR implied volatility; past 2Y; bp/day
6.0 5.5 5.0 4.5 4.0 3
y = -0.08x 2 + 1.20x + 0.99 R 2 = 73%
18-Nov -11
In Exhibit 43, we present the zero iso-return line and 1 standard deviation lines from buying Bund gamma vs. maturity-matched swaption gamma. Buying Bund volatility vs. swaptions when the combination of implied volatility differential and peripheral spreads is below the zero iso-return line is generally profitable. The table within Exhibit 43 shows statistics on trades initiated following the signal from this model in 2011. In addition to a higher average P&L (overall average is shown in Exhibit 41), this strategy has a success ratio of 85%. Of course, these statistics are based on in-sample data and so need to be viewed accordingly. However, given that the fundamental drivers are not likely to change in 2012, we recommend buying govie vs. swaption gamma when the combination of implied volatility differential and peripheral spreads is below the -1 standard deviation iso-return line. Stay neutral on vega Unlike gamma, buying vega would have been profitable in 2011 (Exhibit 44). Outright long vega positions would have resulted in higher returns with smaller swings. For instance, buying 5Yx10Y swaption straddles would have produced an average P&L of 21bp
74
5 6 3Mx 5Y implied v ol; bp/day
of notional per month compared to flat returns for 3Mx10Y. This outperformance of vega is mainly due to the increase in implied volatility and reduced impact of lower delivered volatility. For instance, 5Yx10Y implied volatility has increased from 5.0bp/day to 5.6bp/day in 2011 while 3Mx10Y implied volatility has increased from 5.7bp/day to 8bp/day. However, subsequent delivered volatility on 10Y swaps has been relatively low compared to implied volatility at inception. For example, average 3Mx10Y implied volatility at inception during 2011 is 5.8bp/day, while subsequent 1M delivered volatility has been around 5.1bp/day. Thus, for short dated options, vega gains (from increasing implieds) have been largely offset by gamma losses.
Exhibit 46: Similar to EUR volatility, buying GBP volatility in 2011 would have generated mixed returns, since positive returns during risk-off periods were offset by losses during risk-on periods
Average 1M returns* and annualised information ratio achieved by buying GBP 3Mx10Y swaption straddles; bp of notional ratios
Exhibit 47: We expect front-end volatility in GBP to decline marginally as BoE remains on hold in 2012 and the policy expectations curve remains flat
3Mx2Y GBP implied volatility regressed against 3M/15Mx3M GBP OIS curve; past 1Y; bp/day
40 30 20 10 0 -10 -20 -30 -40 2008 2009 2010 2011 Information ratio (right) Av erage (left)
3 2 1 0 -1 -2 -3
7 6 5 4 3 -20 0 20 40 60 80 100 3M/15Mx 3M OIS curv e; bp 120 140 18-Nov -11 y = 0.02x + 4.04 R 2 = 76%
* Trades entered daily and held for 1M. See Exhibit 30 for definition of returns.
Exhibit 48: even as volatility on 10Y tails increases in response to an escalation in the peripheral debt crisis
GBP implied volatility targets for 2012 vs. 18 November 2011; bp/day 18 Nov 11 1Q target 2Q target 2Q target - current
Additionally, vega has exhibited a concave profile to its gamma counterparts (Exhibit 45). Thus, despite our bullish view on gamma, we stay neutral on vega. We present our targets for various vega points in Exhibit 36.
3Mx 2Y 3Mx 5Y 3Mx 10Y
4.8 5.9 7.2
4.5 6.0 7.5
4.0 6.1 8.2
-0.8 0.2 0.9
Sterling volatility
The story of Sterling volatility in 2011 has been eerily similar to its EUR counterpart. Despite changing inflation expectations and BoE policy reaction, GBP volatility in the belly of the curve has been largely driven by peripheral concerns. Thus, it is not surprising to see that buying GBP swaption gamma in 2011 would have also yielded mixed results, i.e., low monthly average returns with large swings (Exhibit 46). We do not expect this to change in 2012 and recommend buying GBP swaption gamma in 10Y tails. At the front end of the curve, we recommend fading flare-ups by selling 2Y tail volatility and enhancing returns in a range-bound environment by hedging infrequently. Indeed, the argument for selling volatility in GBP is even stronger despite the recent widening of FRA/OIS at the front end. Our economists forecast the BoE to engage in further QE purchases (see United Kingdom) and the base rate to remain at 0.50% throughout 2012, given the gloomy economic backdrop and declining inflation expectations. A flat SONIA curve will likely keep delivered volatility at the front-end
Exhibit 49: Given our view that the peripheral crisis will get worse, we prefer buying EUR gamma over GBP gamma
Average* cumulative return** from buying 3Mx10Y EUR and GBP swaption straddles; bp of notional
60 40 20 0 -20 -40 -60 -80 0%
Risk-off Risk-on
EUR
GBP
GBP EUR 20% 40% 60% 80% % day s into risk on/off period; % 100%
* Averaged over risk-on/off episodes (see Exhibit 32 for risk-on/off dates). ** See Exhibit 30 for definition of returns.
low (Exhibit 47). We present our target for various points on the GBP volatility surface in Exhibit 48. Given a choice between buying EUR and GBP gamma, what should an investor do? Investors facing this
75
choice should consider buying EUR gamma over GBP gamma. Although GBP volatility has been driven by peripheral spreads, EUR gamma remains at the fore front of the crisis and is more likely to outperform GBP. Indeed, long gamma positions in EUR have, on average, outperformed long GBP gamma positions during various risk-off episodes in the past two years (Exhibit 49). We also note that selling EUR gamma during risk-on episodes is better than selling GBP gamma. One reason for superior returns of EUR gamma is that GBP implieds tend to decline with respect to EUR implieds during periods of stress and vice versa (Exhibit 50). We use this relationship to arrive at our targets for GBP gamma for 2012, listed in Exhibit 48.
Exhibit 50: as GBP implieds will likely underperform EUR implieds during such episodes
3Mx10Y GBP-EUR implied volatility spread regressed against 3Mx10Y EUR implied spread; risk-off periods* from 1 January 2010 18 November 2011; bp/day
2 1 0 -1 -2
y = -0.44x + 2.57 R 2 = 55%
18-Nov -11
Trading themes
We are bullish on EONIA and high excess liquidity; receive 6Mx6M EONIA on the back of ECB policy rate cuts; implement bullish option structures on Dec12 Euribor With the ECB expected to cut refi and deposit rates to 0.50% and 0.25% respectively, we expect EONIA fixings to fall to 30bp by mid-2012. We recommend longs in 6Mx6M EONIA with a target of around 30bp, and bullish option structures on Dec12 Euribor. We see value in bullish option structures on 5Y swaps and recommend 1s/5s outright or conditional swap curve flatteners With peripheral sovereign stress expected to escalate in 2012, we expect the swap curve to flatten and recommend carry efficient flatteners such as 1s/5s, which are a proxy for bullish positions but with better risk profile, along with receiver structures on 6Mx5Y. 10s/30s is likely to remain driven by technical flows; we have a flattening bias The 10s/30s swap curve is likely to remain driven by technical factors and is expected to flatten as peripheral spreads widen. We present a framework to analyse the attractiveness of long-dated forward steepeners but refrain from recommending this trade due to the prevalence of technical factors. We favour 2Y wideners and have a swap spread curve flattening bias Swap spreads will likely widen beyond their Lehman peak in 1H12. We target 2Y and 10Y swap spreads at 145bp and 90bp, respectively, by mid-2012. Conditional swap spread wideners offer better risk/reward than outright swap spread wideners.
76
3.4
4.0
4.6 5.2 5.8 6.4 7.0 EUR 3Mx 10Y implied v ol; bp/day
7.6
8.2
* See Exhibit 32 for dates of risk-off episodes.
Position for further widening in spread markets, such as OIS swap spreads and 3s/6s basis, that have lagged the recent widening in swap spread Although spread markets are being largely driven by a single-factor, i.e. sovereign risk, some spread markets such as OIS swap spreads and 3s/6s basis are still trading far below their local/Lehman peaks, suggesting that investors should express risk-off views in these markets. Trade EUR gamma from the long side in the belly of the curve, while fading spikes in volatility at the front end; buy Bund volatility vs. swaption volatility EUR volatility in the belly of the curve will increase as the peripheral crisis escalates; we target 3Mx10Y at 9.4bp/day by 2Q12. Buy gamma in the belly of the volatility curve and fade flare-ups in volatility at the front end. Favour long positions in govie volatility vs. swaption volatility. On a cross-market basis, buy EUR gamma vs. GBP gamma.
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Francis DiamondAC (44-20) 7325-3541 francis.diamond@jpmorgan.com J.P. Morgan Securities Ltd
United Kingdom
Gilt yields fell during 2011 on the back of peripheral spread widening and reached all-time lows across the curve. The BoE restarted QE gilt purchases despite headline CPI breaking the 5% level The macro picture for the UK is worrying, with 2012 growth expected to be just 0.5%. The BoE will likely keep rates on hold and we expect QE gilt purchases to increase to 425bn in 2012. As a result we estimate the BoE will own just over 40% of the total secondary market of outstanding gilts UK banks have relatively small exposures to peripheral sovereign debt but they could be vulnerable if financing markets freeze up given their exposure to wholesale funding markets and 2012 refinancing needs We expect 2Y gilts to trade in a 50-60bp range and recommend fading any richening move below the 50bp level Wider peripheral spreads will drive 10Y and 30Y gilt yields lower we target 1.50% in 10Y gilts and 2.25% in 30Y gilts by 2Q12. Should the UK fall under the sovereign risk spotlight, we expect ongoing QE purchases, a relatively low proportion of non-domestic ownership and an independent currency to limit any fiscal-driven rise in gilt yields The 2s/10s curve will remain highly directional with 10Y yields. We position for flattening and expect 2s/10s to reach the 100bp level by mid2012 We have a bias for a flatter 10s/30s curve as directionality with 10Y gilts should weaken and increasing QE purchases should drive the curve flatter. We target the 75bp level by 2Q, 15bp flatter than currently implied by the forwards Position for wider 5Y and 10Y swap spreads in 1H12, driven by wider peripheral spreads, lower gilt yields and a 150bn increase in QE gilt purchases
Correlation of UK par yields and curve with 10Y weighted peripheral spreads*, last 6M; %
Exhibit 1: The story of 2011 its a one-factor world
100% 80% 60% 40% 20% 0% 2Y 5Y 10Y 30Y 50Y 2s/5s 5s/10s 2s/10s 10s/30s 30Y 30s/50s
77
74% 72%
78% 79%
73% 62% 37%
66% 45% 36%
Exhibit 2: 2011 was the year of the long bond

Total return performance, 1 January 18 November 2011, government bonds;%
40% 35% 30% 25% 20% 15% 10% 5% 0% 2Y 5Y 10Y UK US Germany
2011 The year of the long bond

2011 saw the European sovereign risk crisis move up another notch. The year started off relatively benignly, with the market pricing in rate-tightening expectations as inflation moved higher and surprised to the upside. The SONIA curve was pricing 100bp of tightening over a 1Y forward horizon in the first couple of months of the year. 10Y gilt yields briefly touched the 4% level in February and traded in a broad range until the sovereign risk crisis reignited in the spring. From then on, gilt yields steadily fell almost in lockstep with Bunds to make new all-time historic lows, with 10Y real yields moving into negative territory. Domestic data took a backseat as peripheral spreads were the main driver of gilt yields and of the 2s/10s
curve (Exhibit 1). The large fall in 30Y yields resulted in long-end bonds outperforming on the curve in total return terms, a theme that was common to both US Treasuries and Bunds (Exhibit 2). The BoE increased QE by 75bn in October on the back of heightened downside growth risks despite headline CPI inflation reaching the highest level since mid-2008 (5.2% oya), although the market impact was more limited than when QE1 first started. The 2s/10s curve was highly directional and flattened almost 70bp whilst 10s/30s steepened from April to September only to flatten back following the QE extension in October. Sterling funding markets reflected the various facets of the sovereign risk crisis as Libor/OIS spreads reached their widest levels since May 2009 whilst the flight-toquality demand for gilts and the increase in QE gilt purchases drove gilt GC below SONIA for short dates1.
Exhibit 3: There are signs of a balance sheet recession in the UK

8% 4% 0% -4% -8% -12% 1990 1995 2000 2005 2010 Financial Gov t Corporates Household
Financial balance by sector (negative is net borrower, positive is net lender); % GDP
Exhibit 4: Private sector loan growth in the UK is in the doldrums

Growth rate in MFI lending to private sector; %oya
2012 looks bleak

The macro outlook for the UK in 2012 is predominantly a function of the global outlook and the ongoing sovereign risk issues in Europe. Our economists expect 0.8% growth in the UK for this year and 0.5% oya in 2012, with the likelihood of two consecutive quarters of mild contraction estimated at close to 50%. We believe that that the UK will avoid a deep recession provided growth outside of Europe holds up. Empirically, the UK PMI has demonstrated larger sensitivity to the global ex-Euro area PMI, but other factors such as bank funding concerns and the ongoing fiscal contraction present downside risks to UK growth. There are worrying signs that UK output could remain at anemic levels for several years. Corporates have been running financial surpluses (i.e. have been net lenders) for several years but since 2007, households have shifted from being net borrowers to net lenders possibly indicating signs of a balance sheet recession (Exhibit 3). As a result, the government stepped in with fiscal supports in order to prevent the economy from shrinking, with total government net borrowing moving from 2.5% of GDP in 2007 to 10% of GDP in 2010. Fiscal contraction has prevented this pace of government support from continuing, but the private sector is in debtreduction mode and appears unlikely to take up the mantle to drive the economy forward for some time (we estimate consumption will not begin recovering until at least 2H12). The growth rate of loans to the household sector has recently fallen to negative levels, and both
60% 50% 40% 30% 20% 10% 0% -10% -20% 1998 2000 2002 2004 2006 2008 2010 Financial Corporates Household
Exhibit 5: Several more years of fiscal correction are needed to bring the deficit down
Fiscal thrust* and Public Sector Net Borrowing (PSNB); % of GDP forecast 10
8 6 4 2 0 -2 -4 -6 1972
Fiscal thrust
Net borrow ing
2% fiscal drag is max imum likely achiev able in one y ear 1982 1992 2002 2012
* Fiscal thrust/drag is defined as year on year change in PSNB.
This has been the case in Euro cash markets for some time.
78
corporate and financial loan growth remain at very depressed levels historically (Exhibit 4). We expect the base rate to be kept on hold at 50bp during 2012, with QE purchases increasing by another 150bn by the end of the next year, taking total QE purchases up to 425bn (around 30% of GDP and 40% of total outstanding of gilts). Headline inflation is forecast to fall to 1.9% oya by the end of 2012, but we expect this to have little impact on either gilt yields or inflation breakevens (see Inflation Linked Markets), as markets will remain focused on the risks around peripheral Europe and the growth outlook. The UK budget deficit remains historically high, and we expect this to come in at 5.6% of GDP for FY11/12, but several more years of fiscal correction of around 1.52% of GDP per year are needed to reduce the deficit to more appropriate levels (Exhibit 5). The Chancellor has stated that the governments self-imposed rules are to achieve a balanced, cyclically adjusted deficit and have a falling trajectory of debt/GDP by 2015/16. The recent downshift in UK growth will likely mean that the 2012 Budget may just show these rules being missed and that the Chancellor will have to pencil in some additional modest tightening in the next few years.
Exhibit 7: Gross gilt sales likely to rise to close to 200bn in FY12/13, but net sales will be low
Gross gilt issuance and net gilt issuance adjusted for redemptions and QE*, fiscal year basis; bn
250 200 150 100 50 0
Net of QE and redemptions
Gross gilt sales
* We expect 150bn of additional QE taking total gilt purchases to 425bn by the end of 2012.
Exhibit 8: Gilt issuance in the belly of the curve is likely to increase in FY12/13
Gilt issuance by sector as proportion of total issuance, FY11/12 and J.P.Morgan forecast for FY 12/13; %
40% 35% 30% 25% 20% 15% 10% 5% 0% Short nominal Medium nominal
* Includes syndicated issuance.
FY 11/12
FY 12/13
Summary themes
The evolution of peripheral spreads will likely be the key driver of gilt yields for 2012. Macro, fiscal and technical factors will all be important, but will likely have a larger impact on the evolution of the curve and swap spreads than the outright level of yields. We describe our trading themes and gilt market views in detail in the following pages, but in summary we think: 1) 2Y gilt yields will remain anchored and will likely trade within a tight range. We believe the BoE is unlikely to cut base rates below the current level and think 2Y gilt yields are floored at close to 50bp.
Current and J.P. Morgan gilt forecasts for 2012; % unless stated
Long nominal*
Linkers*
Exhibit 6: We expect 10Y and 30Y yields to fall substantially in 1H12, with the 2s/10s curve flattening
18-Nov -11 BoE base 1M SONIA 2Y 5Y 10Y 30Y 2s/10s (bp) 10s/30s (bp) 0.50 0.52 0.48 1.13 2.26 3.19 178 93 1Q12 0.50 0.50 0.50 1.00 1.80 2.70 130 90 2Q12 0.50 0.50 0.50 1.00 1.50 2.25 100 75 3Q12 0.50 0.50 0.55 1.05 1.75 2.40 120 65 4Q12 0.50 0.50 0.65 1.15 1.95 2.60 130 65 Q2 v s. fw d Q4 v s. fw d (bp) n/a 4 4 -22 -89 -103 -93 -14 (bp) n/a 0 -1 -17 -61 -75 -60 -14
79
12/13 (JPM)
04/05
05/06
06/07
07/08
08/09
09/10
10/11
11/12
2) Wider peripheral spreads will likely drive 10Y and 30Y gilt yields lower we target 1.50% in 10Y gilts and 2.25% in 30Y gilts by 2Q12 (Exhibit 6). We see limited risk from sovereign concerns in the UK. Indeed, if the UK were to fall under the sovereign risk spotlight, we expect ongoing QE purchases, a relatively low proportion of non-domestic ownership and an independent currency to limit any fiscal-driven rise in gilt yields. 3) The 2s/10s curve will remain highly directional with 10Y yields. Position for flattening as we expect 2s/10s to reach the 100bp level by mid-2012. 4) Bias for a flatter 10s/30s curve. We expect directionality with 10Y gilts to weaken as yields fall further. We think 30Y gilts will modestly outperform as QE gilt purchases increase to the 425bn level. 5) Position for wider swap spreads in the 5Y and 10Y sectors.
Exhibit 9: Total QE BoE gilt holdings are expected to peak at around 40% of the secondary market next year, but the BoE could own almost 60% of the 10-25Y sector
Evolution of BoE holdings as % of gilt secondary market outstanding* based on additional QE** and gilt issuance profile for 2012; %
60% 50% 40% 30% 20% 10% Nov 11
3-10Y
10-25Y
25Y+
Total
Mar 12
Jun12
Dec12
* Excludes DMO collateral holdings. ** We assume a further 150bn of QE gilt purchases taking total QE gilt holdings to 425bn.
Gilt supply, QE purchases and bank funding

We expect gross gilt issuance of around 190bn in FY12/13, some 25bn above planned gilt sales for the current fiscal year. Net borrowing for FY12/13 will likely be some 10bn higher than forecast by the OBR (122bn) given a small degree of fiscal slippage, and we expect any increase in the current year net borrowing requirement to be reflected in higher borrowing needs for FY12/13. Gilt redemptions are also higher next fiscal year at just under 50bn. Issuance net of redemptions and QE purchases is expected to be just below 25bn in the current fiscal year, which will then rise to just under 60bn in FY12/13 (Exhibit 7). In terms of maturity split, we expect a 17bn increase in issuance in the 5-15Y sector, to partially counter the expected increase in QE gilt purchases (Exhibit 8). We expect long nominal gilt issuance to increase around 5bn, but remain roughly unchanged as a proportion of total gilt issuance. On a calendar year basis we expect 185bn of gilt supply in 2012 to be fully offset by 43bn of redemptions and 150bn of QE gilt purchases. We think that the demand impact through QE gilt purchases could be a significant driver of gilt yields. Our central view is for the BoE to announce QE
Exhibit 10: If BoE QE gilt holdings were to reach 450-500bn, then the liquidity of the gilt market could be severely impacted as the BoE would own large segments of the gilt market
Estimates for distribution of BoE QE gilt ownership by sector by end 2012 incorporating our issuance forecast; bn
80% 70% 60% 50% 40% 30% 20% 10% 0% 33%
3-10Y
10-25Y
25Y+ 59%
Total 66% 60%
52% 43% 40% 37%
52% 45%
50% 40%
400 450 500 Hy pothetical total BoE QE gilt holdings; bn
increases of 75bn in both February 20122 and May 2012, taking total QE asset purchases to 425bn. We expect these purchases to be entirely in gilts as no other sterling fixed income markets are large enough to accommodate the scale of QE purchases, although we dont rule out that the BoE could extend purchases into bank debt on a much smaller scale. The BoE currently
2 Note: The 75bn of purchases which began in October 2011 will be completed in early February 2012, taking total QE purchases to 275bn.
80
Exhibit 11: UK bank exposures to peripheral government bonds are relatively small compared with total bank reserves
UK bank exposures to Euro zone sovereigns, public sector claims*; $bn Public sector
owns around 25% of the gilt market and around 38% of the secondary market outstanding of bonds in the 10-25Y sector. Given our QE forecast and supply estimates for the next fiscal year, we estimate that the BoE could own just over 40% of the entire gilt market and close to 60% of the 10-25Y sector (Exhibit 9). In our view, this is close to the limit at which QE gilt purchases will affect gilt market liquidity, and we think that the BoE will be unable to increase QE gilt purchases above the 450500bn level as holdings in some sectors of the curve would become prohibitively large (Exhibit 10). Concerns around the impact of the sovereign risk crisis on banks increased in the past month. UK banks have relatively small exposures to peripheral government bonds (Exhibit 11), but they could be vulnerable if financing markets freeze up given their exposure to wholesale funding markets and 2012 refinancing needs. UK banks have around 100bn of refinancing needs in 2012, which is large in absolute terms compared with other European countries, but is relatively small in terms of total banking sector assets (Exhibit 12). Nonetheless, should conditions in financing markets deteriorate if sovereign risk stresses increase, then this could well pose some problems. We would expect the BoE to respond by increasing the amount of funding available via the 3M and 6M LTROs, which are currently small in size at 5bn and 2.5bn, respectively, as was done in 2009. The BoE is unlikely to reopen the SLS3, which will close in 2012 as other mechanisms such as the discount window facility are now in place. UK banks have raised capital and reduced leverage, compared with European banks, but capital remains below the 10% Basel III requirement which needs to be implemented by 2019. Our equity analysts expect them to meet this target organically, but in the event of a severe recession, they are uncertain as to whether regulators would intervene and force capital-raising at some point over the next few years.
$bn Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Peripheral Ita + Spa 2 6 4 56 62 4 4 17 20 2 8 34 25
% 1% 3% 2% 31% 34% 2% 2% 10% 11% 1% 4% 19% 14%
% bank reserv es 0% 1% 1% 10% 11% 1% 1% 3% 3% 0% 1% 6% 4%
* Public sector claims as proportion of all Euro zone public sector claims. Derived from BIS data 2Q11.
Exhibit 12: UK bank refunding needs for 2012 are large in absolute terms but are comparatively small as a proportion of banking sector assets
Total securitised, secured and unsecured bank refinancing requirements for 2012 bn proportion of total banking sector assets; %
140 120 100 80 60 40 20 0
Amount
Proportion of bank assets
4.0 3.0 2.0 1.0 0.0
Netherlands
Portugal
Austria
Ireland
France
Italy
Finland
Greece
Spain
Belgium
Exhibit 13: We dont expect the spread between O/N SONIA and the base rate to fall to the levels seen during QE1, as access to the BoE reserve facility was expanded in late 2009
O/N SONIA base rate spread and periods of QE purchases; bp
Germany
Front end duration view

2011 was almost a re-run of 2010 in SONIA space, as rate-tightening expectations priced in during the first few months of the year were steadily unwound from April onwards, although the magnitude of the re-pricing was
3 Special Liquidity Scheme (SLS) was launched in April 2008 and allowed institutions to lend a wide range asset-backed bonds to the BoE in exchange for 9M T-bills.
10 5 0 -5 -10 -15 Mar 09
QE1
Sep 09
Mar 10
Sep 10
Mar 11
Sep 11
UK
QE2
81
larger than that seen in 2010. We expect the BoE to keep base rates on hold at 50bp, with additional monetary easing coming from further QE rather than a lower base rate. The BoE views 50bp as an effective floor and has previously stated that its preferred tool for further easing monetary conditions is more QE. To this extent, we think that SONIA is effectively floored at 50bp, and we do not expect it to trade much below this level. Following on from the start of QE in March 2009, O/N SONIA fell around 15bp below the base rate, but this was due to fragmentation in the SONIA market given the limited number of participants who had access to the BoE reserve facilities. In the autumn of 2009, the BoE lowered the requirements for institutions to become reserve scheme members and access to the BoE reserve accounts was expanded; the move saw the O/N SONIA vs. base rate spread gradually increase (Exhibit 13). Given that access to the reserve facilities remains unchanged, we dont think that any additional QE in itself should push O/N SONIA below the level of base rates. Given this, we expect SONIA rates up to 12M forward to hover around 50bp, and we would look to fade any cheapening on a forward basis as we think the probability of rate hikes over the next 12 months is very small. Libor rates have widened to reflect increased funding pressure concerns stemming from the sovereign risk crisis, and we think that Libor/OIS spreads can widen further in the UK as peripheral spreads widen as per our forecast. Historically, the level of 2Y par yields can be well explained by the slope of the SONIA curve (we use 18Mx1M 6Mx1M), and this has been the case over most of the 2011 (Exhibit 14). Given our view that base rates will be on hold for 2012 as QE is increased further, we expect the SONIA curve to remain at very flat levels, with a limited amount of term premia priced in. The 6Mx1M/18Mx1M curve has traded in a range around the 10bp level over the past few months, and we expect this dynamic to continue in 1Q12 as we see little chance of the curve pricing in any BoE rate tightening expectations. We think the forward SONIA curve should not trade below 0bp (as the BoE will not lower base rates further in our view) which effectively means 2Y gilt yields should not fall much below the 50bp level. We expect 2Y gilt yields to range-trade around the 50-60bp level for much of 2012.
Exhibit 14: 2Y gilts are driven by the slope of the SONIA curve and should not trade much below 50bp in our view given that the base rate is floored at 50bp
2Y par gilt yield regressed against level of SONIA curve, November 2010 November 2011; %
1.60 1.40 1.20 1.00 0.80 0.60 0.40 -20 0
y = 0.0004x 2 + 0.005x + 0.58 R 2 =98%
current
20
40 60 SONIA curv e*; bp
80
100
120
* 6Mx1M/18Mx1M SONIA curve.
Exhibit 15: Peripheral spreads are more significant in explaining 10Y gilt yields than macro factors
10Y par gilt yield regressed against 10Y peripheral spread*, UK composite PMI, CPI and QE dummy; monthly data March 2007October 2011 Factor Coefficient T-Stat
Peripheral Spread, bp Composite PMI CPI, %oy a QE dummy R-squared Std. error, bp
-0.01 0.03 0.24 -0.36 83% 32
-12.7 3.7 3.7 -3.0
Exhibit 16: and the model suggests 10Y yields are some 60bp above fair value
Actual and model predicted 10Y par yields, monthly data March 2007November 2011; %
6.00 Actual 5.00 4.00 3.00 2.00 1.00 Mar 07 Mar 08 Mar 09 Mar 10 Mar 11 Predicted
Be long 10Y and 30Y duration.

In the belly of the curve, 5Y and 10Y gilt yields made all-time lows in 2011, and are far below the levels that
82
macro-economic models would suggest. However, the low level of yields is less a function of the UK macro landscape and more a function of the sovereign risk crisis in Europe. We present a simple regression model to highlight this point (Exhibit 15) and whilst the level of the composite PMI and level of headline CPI are important drivers, their statistical significance is much lower than 10Y weighted peripheral sovereign spreads. We also include a QE dummy to control for periods of BoE gilt-buying. This model suggests that 10Y yields are some 60bp above fair value level (Exhibit 16). We expect 10Y yields to gradually fall to the 1.50% level over the first half of 2012 as peripheral spreads widen substantially, with the weighted peripheral spread expected to reach the 955bp level, from the current 700bp (see Euro Cash for our forecasts). A tightening back in peripheral spreads should result in a rise in 10Y gilt yields in 2H12. We note that over recent weeks gilt yields have not fallen as much as the model suggests given the move in peripheral spreads, but we think this is reflection of a lack of investor appetite into year end to enter fresh duration positions. We expect the usual linear relationship between gilt yields and spreads to resume in the New Year. We think that 5Y yields will become sticky in a rally and will fall less than 10Y yields. We expect 30Y yields to outperform, driven by the ongoing QE gilt purchases and weakening directionality (See curve section below). Gilts have acted as a safe haven in recent months as European sovereign risks have increased, and we expect this to continue into 2012. In the event that the situation in Europe escalates far enough so that Germany looses its safe-haven status then Bund yields could rise. In that environment, we think gilt yields will also rise but may well outperform bunds due to flight-to-quality flows.
Exhibit 17: The currency can provide an escape valve when UK fiscal pressures build
Trade-weighted sterling index regressed against PSNB*, yearly 19762011
120 110 100 90 80 70 -2.0 0.0 2.0 4.0 PSNB, % GDP

* Public Sector Net Borrowing.
y = -2.7x + 106.3 R 2 = 41%
2011 est
6.0
8.0
10.0
Exhibit 18: We expect debt/GDP to reach the 75%85% level in five years time
J.P. Morgans debt/GDP forecasts under various economic scenarios* on a fiscal year basis, dashed line is end of government forecast period; %
110 100 90 80 70 60 2011/12
Baseline Sev ere Recession Mild Recession 95% lev el
2012/13
2013/14
2014/15
2015/16
2016/17
* Baseline is our central GDP forecast; mild recession assumes mildly negative GDP in 2012 followed by a modest recovery back to the 2% level; severe recession assumes a repeat of 2008/09, with 2012 GDP at -1.5% and 2013 GDP at -3.5%.
as we dont expect fiscal pressures to drive gilt yields higher

In presenting our duration view, we are assuming that the current sovereign risks do not envelop the UK in the coming months. Although the UK is one of the worst AAA countries in terms of its debt dynamics and still has some way to travel on its fiscal journey, our base case scenario is that UK gilt yields do not rise materially in response to fiscal concerns. Through the large-scale purchases of gilts the BoE appears to be sacrificing longterm inflation certainty for near-term fiscal stability. QE gilt purchases are expected to rise to 425bn by the end of 2012, which should limit the potential for substantial fiscally-driven increases in gilt yields. If fiscal stresses
were to become severe enough then the currency could also act as a pressure valve (Exhibit 17), limiting the magnitude of any rise in gilt yields (note our central view is for relatively stable trade weighted sterling in 2012)4. The impact of weaker growth on the UKs debt/GDP dynamic is important. Exhibit 18 shows our estimates of the evolution of this metric under several different economic scenarios. In our central case of 0.5% GDP in 2012 and a slow recovery to the 2% level over the subsequent few years, debt/GDP (excluding the impact of financial interventions) will peak just under 77% on a
4
See Global FX Strategy 2012

83
five-year horizon, some 10% higher than the estimate for FY11/12. In a mild recession scenario where output is mildly negative for 2012 this peak is predicted to be around 85% with a steeper trajectory. It would take a recession of the severity of 2008/09 to push debt/GDP to close to 100% in five years time, something we think is not likely at the moment. If either the baseline or mild recession scenarios play out then gilt yields will not rise materially due to fiscal concerns as long as the government maintains its fiscal commitments and credibility. However, if the severe recession scenario plays out then we expect rating agencies to act, not only because of the weaker growth outlook but also because of an expected increase in contingent liabilities as the UK government may have to re-introduce its guarantee programmes, and financial markets will likely price an increase in UK sovereign risk. In this environment safe-haven purchases of gilts would abate and non-domestic holdings of gilts would probably fall. In addition, as non-domestic holdings are just less than one third of the total gilt market which should limit an impact from foreign investors taking flight from the UK gilt market. This compares favourably with the main Euro zone countries where non-domestics own between 40 and 65% of the total bond markets and domestic holdings in the UK are not highly concentrated in the banking sector (Exhibit 19). It is worth noting that foreign central bank ownership has been very stable at around 6-8% of the market and that the increase in nonresidents holdings has come from other investors. In fact, the increase in other foreign investors (ex central banks) holdings of gilts has been well correlated with gilt supply, suggesting that non-domestic investor participation has increased as the size of the gilt market has expanded (Exhibit 20). In our view, market pressure is only likely to come to bear if the governments fiscal plans lose credibility. So far the coalition has shown no signs in wavering on its fiscal tightening commitment but the Chancellors selfimposed rules are close to being breached. If output weakened to the point where fiscal contraction needed to be slowed then we think that the Government can back load tightening further into the current parliament term. However, we think that the maximum tightening achievable in any one year would be 2% of GDP. Attempting to do any more than this amount would risk straining the governments fiscal credibility in our view. Our base case view is that the UK retains its AAA rating and stable outlook but rating agency action
84
Exhibit 19: Non-domestic investors holdings of gilts are lower than non-domestic holdings in the main European bond markets
Foreign and domestic bank* holdings of government bonds as % of total market outstanding; %
90 80 70 60 50 40 30 20 10 0
81 66 45
Non-domestic
Domestic bank*
46 38 33 30 12 4 6
44
11
10
13
Italy
UK
France
Spain
US
R 2 = 74% 200
* Excludes Central Bank holdings.
Exhibit 20: Non-domestic investors gilt holdings have increased as gilt issuance and market liquidity have increased
Foreign investor ex-Central Banks gilt holdings regressed against gross gilt issuance, annual data 19762011; %
30% 25% 20% 15% 10% 0 50 100 150 250 Gross gilt issuance, bn y = -0.0001x 2 + 0.002x + 0.10
Exhibit 21: Our central view is that the UKs sovereign rating will not be downgraded, but should rating agencies put the UK on negative outlook, we expect only a modest 20bp rise in 10Y gilt yields
Adjusted 10Y gilt yield* around the time when S&P placed the UK on negative outlook in May 2009; bp
20 15 10 5 0 -5 -10 -15 -20 -25 Mar 09
S&P places UK on negativ e outlook
Apr 09
Germany
May 09
Jun 09
Jul 09
* 10Y par gilt = 0.7* 10Y par Bund yield + 0.76 * 12Mx3M (SONIA EONIA) + 1.2. Rsquared: 82%, std. error: 9bp.
Japan
cannot be ruled out if output weakens substantially. However, if the UK were to be put on review for a potential downgrade then we think the gilt market reaction will be muted. When the UK was placed on negative outlook by S&P on 21 May 2009 the adjusted 10Y gilt yield rose by less than 20bp (Exhibit 21) and the impact dissipated quite quickly. Downgrades of AAA sovereigns have typically had a limited impact on yields historically with the Japanese downgrades in November 1998 and February 2001 resulting in less than a 15bp rise in yields in the week following the move. The US downgrade earlier this year had a similarly limited reaction on the day of the announcement with Treasury yields falling some 30bp in the following week on flight to quality flows.
Exhibit 22: The 2s/5s curve can be well explained by the slope of the SONIA curve, the level of 5Y RPI, and QE purchases
Actual 2s/5s par gilt curve vs. model curve*; March 2009November 2011; %
1.80 1.60 1.40 1.20 1.00 0.80 0.60 Mar 09 Sep 09 Mar 10
Actual
Model
Curve view position for 2s/10s flattening

The 2s/5s curve is flatter since the start of the year but has been trading in a 30bp range since mid-August and has shown little correlation with European peripheral spreads over this period. We can model the par 2s/5s curve as a function of three variables: 1) the forward SONIA curve (24Mx1M 12Mx1M), to capture changing base rate expectations; 2) 5Y RPI to capture changing inflation expectations; and 3) a QE dummy variable to take into account QE gilt purchases (which have been in the 3Y+ sector of the gilt curve). The model fit has been good over the past couple of years (Exhibit 22), and currently the 2s/5s curve is fairly valued. Going forward, we expect this range-trading dynamic to continue. Under the assumption that QE gilt purchases continue for the next 6 months, we outline how the curve could change under different assumptions for the slope of the SONIA curve and the level of 5Y RPI swaps (Exhibit 23). In our view, the SONIA curve is unlikely to trade below zero (as we think the chances of base rate cuts are low), and we dont expect this curve to trade much above 35bp, given our low-for-long view. We think 5Y RPI swaps could modestly fall to the 2.50% level at worst, and with the SONIA curve likely to remain close to current levels, we expect very modest flattening of the 2s/5s curve. Any steepening is likely to be limited, and we think the 2s/5 curve can trade in a 50-70bp range over 1H12. The 2s/10s curve has been highly directional with 10Y yields when the 10Y rate is below 3.20% (Exhibit 24). We expect this to remain the case over 2012 as 2Y yields are anchored, and we find it hard to envisage a scenario in which 10Y gilt yields will rise above this 3.20% level in the next couple of quarters. We dont expect the
Sep 10
Mar 11
Sep 11
* 2s/5s par = 0.25 * 5Y RPI inflation swap + 0.49*OIS curve 0.18 * QE dummy + 0.028. R-squared: 79%, std. error: 10bp.
Exhibit 23: We think the 2s/5s curve cannot flatten much more, and any steepening will struggle to beat the forwards we recommend playing the range
Projected changes* in 2s/5s curve over the next 6 months, given various levels of 5Y RPI swaps and OIS curve**; shaded indicates our expected outcomes; bp 5Y RPI swap, %
OIS curve**; bp 75 50 25 0 -25
2.50 15 5 -10 -20 -30
2.75 25 10 0 -15 -25
3.00 30 15 5 -10 -20
3.25 35 25 10 0 -15
3.50 40 30 15 5 -10
* Calculated using regression equation in Exhibit 22. Assumes QE buying is in place over the next 6 months. Current OIS curve = 17bp, 5Y RPI swap = 3.04%. **OIS curve = SONIA OIS 24Mx1m SONIA OIS 12Mx1M.
Exhibit 24: With the front end anchored, the 2s/10s curve is likely to remain highly directional with 10Y yields during 2012
2s/10s par curve regressed against 10Y par yields, last 12M, %
2.80 2.60 2.40 2.20 2.00 1.80 1.60
10Y par < 3.2% y = 0.77x - 0.03 R 2 =93%
10Y par > 3.2%
y = -0.04x + 2.6 R 2 = 0% current 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 10Y par rate, %
85
increase in QE to generate an increase in forward inflation expectations and we think it is too early for this dynamic to be reflected via a steeper curve. Inflation expectations are only likely to become unanchored to the upside when excess demand has recovered and the excess central bank liquidity is being reflected in strong credit and loan growth. This scenario is several years away yet in our view. We forecast the 2s/10s curve to flatten to the 100bp level by the end of 2Q, substantially more than priced into the forwards. The 2s/5s/10s fly is typically viewed as a directional fly, and this has been the case for most of the past 12 months, although we note that the directionality has weakened substantially when 5Y par yields have been below 1.5% (Exhibit 25). This convexity phenomenon is likely due to investors attributing some degree of risk premia to longer maturity rates when the base rate is anchored (i.e. some upward slope) and has been observed in Japan (see Euro Cash) in the past when yields get to very low levels. We think this convexity phenomenon will remain in place as 5Y and 10Y yields fall further. The 10s/30s curve has also been highly directional with 10Y yields (Exhibit 26), although the curve has persistently traded too flat vs. the level of yields over the past couple of months. We think this is a reflection of the resumption of QE gilt purchases in early October as the 10s/30s curve, adjusted for the level of 10Y yields, has been broadly flattening as the BoE has been buying relatively more 25Y+ gilts compared with 10-25Y gilts, taking into account DMO supply (Exhibit 27). Any implied steepening from lower 10Y gilt yields will likely be gradually offset by the BoE purchasing relatively more 25Y gilts compared with 10-25Y gilts (in 10Y equivalent terms). We note that since QE restarted in October that the BoE has had to pay relatively more compared with preauction levels for 25Y+ gilts and that cover ratios for buybacks in this sector have been much lower than in the other two sectors (Exhibit 28). We expect this dynamic to continue as QE purchases increase. The directionality of the 10s/30s curve with the level of yields is likely to become weaker as 10Y yields fall further as investors roll along the curve in search of higher yields. Overall, these factors should result in modest flattening in the 10s/30s curve and we target the 75bp level by 2Q, some 15bp flatter than currently implied by the forwards.
Exhibit 25: 2s/5s/10s fly directionality has weakened as yields have fallen and is likely to remain low
2s/5s/10s par yield fly (50:50) regressed against 5Y par yield, last 24M; %
0.30 0.20 0.10 0.00 -0.10 -0.20 -0.30 -0.40 0.50 1.00 1.50 2.00 2.50 5Y par y ield; % 3.00 3.50 5Y par < 1.5% y = 0.16x - 0.4 R 2 =6% 5Y par > 1.5% y = 0.26x - 0.61 R 2 = 90%
Exhibit 26: The 10s/30s gilt curve has been driven by 10Y yields
10s/30s par gilt curve regressed against 10Y par yields, January 2011November 2011; bp
140 120 100 80 60 current 40 2.00 2.50 3.00 3.50 10Y par y ield; % 4.00 4.50 y = -35.2x + 196.6 R 2 = 80%
Exhibit 27: and the excess flattening since October can be explained by BoE gilt purchases in the 25Y+ sector
10s/30s curve adjusted for level of 10Y gilt yields* vs. cumulative excess of QE gilt purchases in the 25Y+ sector compared with the 10-25Y sector adjusted for DMO supply **; weekly data 10 October 201118 November 2011 bp bn, 10Y equivalents
-10 -12 -14 -16 -18 -20 -22 -24 10 Oct 17 Oct 24 Oct 31 Oct 07 Nov Adjusted 10s/30s Cum. ex cess of 25Y+ purchases (bn**)
-4 -2 0 2 4 6 8 10 14 Nov
86
* Residual from linear regression of 10s/30s curve vs. 10Y yields over the past 12 months, weekly average. ** Expressed in bn of 10Y equivalents, cumulative purchases from 6 October 2011 when QE restarted minus DMO supply. Scale inverted.
Market technicals curve-trading rule and trading short-dated gilts

The 5s/10s curve has generally been non-directional with the level of yields for most of the past year, although this directionality has increased in the past couple of months. We think this sector of the curve is the most likely to be affected by momentum dynamics as regular DMO supply, QE purchases and reasonable liquidity can result in opportunities for investors wishing to express shortterm tactical views. To this extent, we have developed a momentum-trading rule model for the 5s/10s par gilt curve, based on a moving average indicator5. The intuition is that when the signal is large and positive, then the curve should flatten in subsequent days, and when the signal is large and negative, the curve should steepen in subsequent days. We test a trading rule based on different signal thresholds for initiating and exiting trades since the start of 2010 and since the start of 2011. A rule that initiates curve trades when the signal threshold is +/- 2.5 and exits trades when the threshold is +/- 0.25 has performed the best over the two periods, generating a total P/L since 2010 and 18bp P/L since 2011 (Exhibit 29), with a success ratio around 75% in both periods. The performance of this rule has been lower in 2011 compared with 2010 as general risk appetite has fallen as a result of the worsening sovereign risk crisis. In an environment in which both regular supply and BoE buying continue to take place in the 5Y and 10Y sectors, we expect the 5s/10s curve momentum model to continue delivering positive results. We also look at the behaviour of very short-end gilts from a technical perspective. As bonds fall below 1Y remaining maturity, they become less liquid and cease to become part of most investors portfolios. We would expect gilts to start to cheapen vs. surrounding lines at some point before they fall below the 1Y point. Looking at examples over the past couple of years and using a yield spread adjusted for the level of 2Y gilts we find that this starts to occur, on average, 50 days before the bonds become sub 1Y in maturity (Exhibit 30). Gilt 4.5% Mar13 falls below the 1Y remaining maturity point
5
Exhibit 28: The BoE has had to pay more vs. market levels to buy 25Y+ bonds we think this will continue
Average coverage ratio and average yield of bonds purchased vs. market levels for QE gilt buybacks conducted from 6 October 2011 onwards Ratio bp
3.5 3.0 2.5 2.0 1.5 3-10Y
Av g. cov er ratio
0.4 0.3 0.2 0.1 0.0
Yield v s. market
10-25Y
25Y+
Exhibit 29: A momentum trading rule for the 5s/10s gilt curve has generated positive P/L since 2010
Total P/L and number of signals generated by 5s/10s curve momentum model * trading rule ** since 2010 and 2011; bp
50 P/L 40 30 20 10 0 Since 2011 18 Total no. signals
46
11 4
Since 2010
* We first generate a MACD (Moving Average Convergence Divergence) measure. This is calculated as the expected moving average (EMA) of 10Y yields over a 5-day period minus the EMA of 10Y yields over a 25-day period. We then generate a signal measure which is the 10-day EMA of the MACD measure. The trading signal is defined as the MACD measure minus the signal measure. ** The trading rule is calibrated based upon a threshold for the signal at which to instigate and unwind trades. We use a signal of + (-) 2.5 as the threshold to enter trades and a signal of + (-) 0.25 as the threshold to close open trades.
We first generate an MACD (Moving Average Convergence Divergence) measure. This is calculated as the expected moving average (EMA) of the 5s/10s curve over a 5-day period minus the EMA of the 5s/10s curve over a 25-day period. We then generate a signal measure which is the 10-day EMA of the derived MACD measure. The trading signal is defined as the MACD measure minus the signal measure.
87
towards the end of 1Q12 and we recommend underweighting this line vs. gilts 8% Sep14 and 2.25% Mar14.
Exhibit 30: As bonds approach 1Y remaining maturity they tend to underperform vs. 18M and 2Y gilts
Adjusted yield curve* for selected 1Y vs. 18M and 2Y bond pairs** in the business days around gilts falling below 1Y remaining maturity; bp
0 -5 -10 -15 -20 -90 -60 -30 0 Day s around bond falling sub 1Y 30
Position for wider 5Y and 10Y swap spreads

Sterling swap spreads have widened over 2011, with the bulk of this occurring in the past few months of the year. 10Y swap spreads were relatively stable during 1H11, trading in an 8bp range from January to July, before widening out during 2H11 to reach levels not seen since mid-2009 (Exhibit 31). The move wider occurred across the swap spread curve, with 2Y swap spreads widening the most (53bp up to 18 November 2011), driven by increased European sovereign risk fears and renewed bank funding concerns. The resumption of QE in October resulted in 30Y swap spreads moving out of the trading range that had been in place for most of 2011. Swap spreads in the 2Y sector have broadly mirrored the moves in European peripheral spreads, as Libor/OIS spreads have trebled with the increase in bank funding concerns, and we expect this dynamic to continue in 2012. Further out, 5Y and 10Y swap spreads have generally been directional with the level of yields, although this relationship has started to become non-linear as yields have reached extremely low levels. The resumption of BoE gilt purchases has also affected swap spreads. Over the past three years, 10Y swap spreads can be reasonably well explained by the level of 10Y yields, monthly QE purchases in the 10Y sector net of DMO supply, and implied volatility (Exhibit 32). Given our views of lower 10Y gilt yields and an increase in QE to 425bn, we expect 10Y swap spreads to widen to around the 55bp level in 1H12 (Exhibit 33) before slowly tightening back in 2H12 as yields rise modestly. Our expected increases in net gilt supply for FY12/13 is relatively small, and we think that the market is already expecting a slight worsening in public finances. However, the scope for the government to increase borrowing is limited, and we expect any tightening impact on 10Y swap spreads from increased borrowing forecasts from the OBR this year or increased gilt issuance in FY12/13 to be limited. 30Y swap spreads can be explained by the level of 10Y swap spreads and net QE buying of 30Y gilts (Exhibit 34). As discussed, we dont expect the BoE to alter the maturity split of gilt purchases if QE is extended, and we expect 30Y swap spreads to gradually widen over 1H12,
88
* Yield spread adjusted for 2Y par gilt yield. ** Yield spreads used: Mar11/Mar12 yield spread around 7 March 2010; Dec11/Jun12 yield spread around 7 December 2010; Mar12/Mar13 spread around 7 March 2011; Jun12/Mar13 spread around 6 June 2011.
Exhibit 31: Following a period of relative stability, 2Y and 10Y swap spreads have substantially widened back to mid-2009 levels
2Y and 10Y maturity matched swap spreads; bp
150 100 50 0 -50 Nov 08 May 09 Nov 09 May 10 Nov 10 May 11 Nov 11 QE1 2Y 10Y QE2
Exhibit 32: 10Y swap spreads can be modeled as a function of gilt yields and net QE purchases
10Y swap spread regression model*, monthly data November 2008October 2011 Variable Coefficient T-stat
10Y y ield, % Monthly Net QE purchases, bn 3Mx 10Y v ol, bp R-squared Std error
-9.7 2.3 3.4 59% 11
-2.2 5.2 2.2
* Maturity matched benchmark gilt swap spread.
Exhibit 33: We expect wider swap spreads across the curve, with 30Y spreads breaking the 0bp level
Current maturity matched swap spreads and J.P. Morgan forecasts for 2012 Sw ap spreads 18-Nov -11 1Q12 2Q12 3Q12 4Q12
2Y 5Y 10Y 30Y
96 71 31 -9
110 85 45 0
120 95 55 5
110 85 45 0
100 65 25 -10
reaching and then modestly breaking through the 0bp level. However, we note that 30Y swap spreads globally are much tighter than their pre-crisis levels (Exhibit 35), and whilst we expect 30Y UK swap spreads to widen, we dont expect a move back to the levels seen in 2007, given the current global backdrop. In addition, 30Y swap spreads are more affected by investor flows and rangetrading behaviour, and the relative cheapness of linkers on ASW compared with nominal gilts (see Inflation Linked Markets) could encourage asset swap investors to shift out of holding nominal gilts into inflation-linked gilts on ASW. Given these technical factors, we think that swap spread widening to 0bp will be a slow process interspersed by bouts of tightening.
Exhibit 34: 30Y swap spreads can be broadly explained by the level of 10Y swap spreads and net QE buying in the 25Y+ sector
Actual 30Y maturity matched swap spread and model predicted 30Y swap spread*; weekly data, past 12 months; bp
-5 Actual -10 -15 -20 -25 -30 Predicted
Trading themes
Be long duration in 10Y and 30Y gilts in 1H12 We expect wider peripheral spreads to drive gilt yields lower and we forecast 10Y gilts to reach 1.50% by 2Q. The fiscal environment is challenging but we expect ongoing QE gilt purchases and a relatively low proportion of non-domestic investors to prevent UK yields from rising due to a spill over of sovereign risk fears into the UK Enter 2s/10s curve flatteners The curve will likely remain highly directional with 10Y yields. We forecast almost 90bp of flattening compared to what is priced by the forwards by the middle of 2012 Bias for a flatter 10s/30s curve We think the 10s/30s curve can flatten despite our view of lower nominal yields. As gilt yields fall further we expect the directionality of the 10s/30s curve to weaken as investors move out the curve in search of yield. In addition, further QE gilt purchases should support 30Y gilts relative to 10Y gilts Position for wider 5Y and 10Y swap spreads Swap spreads are expected to widen further driven by wider peripheral spreads, lower nominal yields and a total of 425bn of QE gilt purchases by the end of 2012
Nov 10
Feb 11
May 11
Aug 11
Nov 11
* 30Y ASW = 0.45 * 10Y ASW + 0.62 * net QE buying of 25Y+ gilts 25.3. R-squared: 58%, std error: 4bp.
Exhibit 35: and we expect further widening but not back to the levels seen prior to 2008
Maturity matched 30Y swap spread for the UK and weighted global 30Y swap spread*; bp
75 50 25 0 -25
UK
Global ex UK
-50 -75 Nov 06 Nov 07 Nov 08 Nov 09 Nov 10 Nov 11
* Calculated as the average of US, German and Japanese 30Y maturity matched swap spreads, weighted by total government bond market value.
89
US Fixed Income Strategy US Fixed Income Markets 2012 Outlook November 24, 2011 Srini RamaswamyAC (1-212) 834-4573 Kimberly L. Harano (1-212) 834-4956 J.P. Morgan Securities LLC
US Cross Sector
We look for growth to slow in 1H12 as past fiscal stimulus programs roll off, but accelerate thereafter. With the unemployment rate remaining elevated, core inflation should slow Although our baseline forecast does not yet call for an imminent QE3, we think there is a significant chance that the Fed resumes balance sheet expansion in 2012, most likely by buying MBS The European crisis will likely worsen before it gets better. Key developments in 2012 are likely to include further funding pressures on Italy, a Euro-area recession, and increased risks of a downgrade of France The crisis should pressure Treasury yields lower into 1Q12, but a deteriorating supply/demand balance, rich valuations, and a decoupling from Europe will limit the rally Despite near-term risks from a worsening crisis, spread products are likely to benefit from significant offsets such as solid credit fundamentals, negative net supply (in securitized products), and attractive valuationscautiously overweight B rated high yield bonds, new issue AAA and subordinate CMBS relative to legacy paper, select ABS, and Agencies; stay overweight EMBIG versus CEMBI Overweight mortgages given significant risks of QE3 and attractive valuations Tail risk is not what it used to be: the European crisis is far from the only tail scenario investors will need to consider in 2012. We characterize four distinct risksthe European crisis, an EM Asia hard landing, and deflationary and inflationary scenariosand explore the best hedging strategies We find that synthetic strategies designed to match the duration of liabilities or hedge the funding gap for defined-benefit pension plans significantly lower the tracking error and volatility
Exhibit 1: Another wild ride in markets

QTD chg
11/17/11 level, quarter-to-date change, quarterly changes over 2011, and changes over 2010, 2009, and 2008 for various market variables
Current Global Equities (level) S&P 500 E-STOXX FTSE 100 Nikkei 225 Sovereign par rates (%) 2Y US Treasury 10Y US Treasury 2Y Germany 10Y Germany 2Y JGB 10Y JGB 5Y Sovereign CDS (bp) Greece Spain Portugal Italy Ireland Funding spreads (bp) 2Y EUR par swap - par gov't spd 2Y USD par swap - par gov't spd EUR FRA-OIS spd USD FRA-OIS spd 1Y EUR-USD xccy basis Currencies EUR/USD USD/CHF USD/JPY JPM Trade-weighted USD Spreads (bp) 30Y CC MBS L-OAS 10Y AAA CMBS spd to swaps JULI (ex-EM) Z-spd to Tsy JPM US HY index spd to worst EMBIGLOBAL spd to Tsy MAGGIE (Euro HG spd to govies) US Financials spd to Tsy Euro Financials spd to govies 10Y AAA muni/Tsy ratio (level) 30Y AAA muni/Tsy ratio (level) Commodities Gold futures ($/t oz) Oil futures ($/bbl) 119.6 50.9 81.4 67.7 -88.2 1.352 0.917 76.96 81.27 47.7 315.0 235.6 741.4 416.6 76.2 325.6 316.3 114% 126% 20.4 22.9 18.0 24.3 -15.6 38.0 5.7 34.0 22.6 -46.5 -0.7 3.8 3.1 -0.2 0.1 0.031 -1.73 -1.46 9.6 30.0 14.9 50.1 -10.4 6.3 21.1 20.1 -5.8% -7.4% 86.60 -18.3 -3.8 -7.2 1.7 22.7 28.5 -5.4 3.3 4.6 -22.0 -42.3 -61.2 -39.9 -60.9 22.6 39.6 8.4 27.0 27.6 -36.0 4964 478 1173 570 763 -975 94 -42 96 22 3827 122 417 303 -50 1019 29 206 24 128 86 -116 101 -86 57 721 235 403 126 441 61 4 -7 -58 -5 205 92 77 145 156 0.290 -0.006 -0.175 -0.301 2.015 -0.007 -1.238 -0.277 0.373 -0.135 -1.058 -0.162 1.963 0.024 -1.111 -0.346 0.132 -0.012 -0.026 -0.023 0.895 -0.100 -0.166 -0.143 0.197 0.148 0.939 0.391 0.034 0.138 -0.58 -0.59 -0.55 -0.44 -0.02 -0.17 0.57 1.09 -0.46 0.33 -0.19 0.14 -2.43 -1.32 -2.21 -1.23 -0.33 -0.28 1216.1 2242.8 84.7 -189.2 63.1 -668.9 -5.2 -62.4 68.2 118.1 143 -172 487 -318 212 -565 517 -1952 979 -2023 1687 -6448 3Q11 chg 2Q11 chg 1Q11 chg 2010 chg 2009 chg 2008 chg
5423.1 294.7 -817.2 8479.6 -220.7 -1115.8
36.9 8.8 61.0 -473.8
0.009 -0.107 0.011 0.21 -0.74 -7.0 -50.0 -13.6 -66.6 -48.4 9.8 -8.5 18.5 -4.29 4.41 17.0 145.0 96.7 241.0 176.5 18.1 153.0 127.9
0.092 -0.103 0.016 -0.056 1.30 -10.81 -1.71 -2.4 -50.0 -17.4 -66.0 10.3 -6.8 -19.2 -29.4 -3.0% -0.9% 17.90 15.34 -3.20 35 -260 -11 -6 14 -11 47 18% 1.49 -23.00 -4.66 -66 -375 -363 -430 -27 -359 -164 -47% 6.38 59 793 333 1128 470 38 336 208 36% 97% 27 -51.4
0.065 -0.073 -0.023 -0.103 -0.012 -0.077
-74 -1068
3.8% 25.4% 6.0% 21.5%
18% -104% 313 12.0 223 34.8
1774.30 157.00 106.90 98.82
19.49 -16.09 -11.30
The year that was

The past year was an unforgettable one, even if we might wish we could forget it. It is impossible to forget the human tragedy in Japan from the earthquake, tsunami and the nuclear reactor crisis in the first quarter. The earth appeared to shift under market participants feet as well, with a number of crises coming out of left field. The year started off strongly, with forecasts for robust growth pushing equities and bond yields higher and
90
Exhibit 2: This crisis-filled year produced trading ranges in yields and credit spreads that were only surpassed by the 2008-09 crisis and recovery
Yearly range in 10-year Treasury yields and JULI ex-EM I-spread to Treasury; bp
Exhibit 3: In 2011, investors were generally not rewarded for taking on more risk in high-beta sectors
400
Rolling 3-month total returns (not duration adjusted), averaged over 3/3110/31/11 versus the standard deviation of those returns for various asset classes; % 6 US 7-10Y
DEM 7-10Y ABS MBS Agy US 13Y JULI DEM 1-3Y EM CMBS HY
300
10Y UST yield
JULI
2 0 -2
200
100
S&P -4 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Standard deviation of 3M returns; %
causing credit spreads to narrow (Exhibit 1). By the middle of the year, however, expectations for US growth plunged sharply as the run-up in commodity prices posed a threat to growth and the supply shocks from Japan appeared more severe than anticipated. At the same time, the sovereign crisis in Europe ballooned from impacting only smaller, peripheral countries into a much larger issue, spreading to Spain and Italy, and threatening the banking system. Third, political deadlock in the US nearly produced a crisis in markets as Congress played a game of chicken with the vote to increase the debt ceiling. In part as a result of that political maneuvering, S&P downgraded the US sovereign rating from AAA to AA+ with a negative outlook. All of these events led to a sharp underperformance of risky assets in the second and third quarter. With the economy looking vulnerable, the Fed was forced to ease again, this time launching Operation Twist instead of another round of balance sheet expansion. Finally, in the fourth quarter, economic data turned more upbeat, and with some positive developments coming out of Europe after the October 26 summit, US risky asset markets have managed to rally this quarter. All in all, given crises erupting left and right, 2011 was a year with trading ranges in yields and credit spreads that were only surpassed by the financial crisis periods in the US (Exhibit 2). Given the high volatility, it is unsurprising that investors were generally not rewarded for taking on additional risk this year, particularly in high-beta sectors. Unlike years
Exhibit 4: Risky asset spreads have been very well correlated to measures of European exogenous risk
Cross-sector spread index* versus average CDS for French banks**; Level bp
0.0 -0.2 -0.4 -0.6 -0.8 French bank CDS Cross-sector spread index
400 350 300 250 200 150 100
-1.0 50 Nov 10 Jan 11 Feb 11 Apr 11 Jun 11 Jul 11 Sep 11 Nov 11 * Average of 3-year z-scores for 5-year swap spreads, JULI spread to Tsy, JPM HY index spread to worst, 10Y AAA CMBS spread to swaps, 2Y AAA card ABS spread to swaps, and EMBIGLOBAL strip spread to Tsy. ** Average of 5-year CDS spreads for Societe Generale, Credit Agricole and BNP Paribas.
past, when the riskiest/most illiquid sectors such as emerging markets and CMBS produced the best returns, this year CMBS, high yield, and especially equities sharply underperformed, while longer-maturity US and German government bonds posted the best returns (Exhibit 3).
91
Lifting the covers off 2012: What lies beneath

Although it has been nearly two years since Greeces fiscal problems entered the spotlight, the European crisis continues to impact US markets in highly significant ways. On a fundamental basis, Europe now risks falling back into recession, posing a threat to global growth as well. In addition, the flight-to-quality premium has been a factor keeping US Treasury yields well below fair value. Indeed, we have often noted that yields, adjusted for their usual long-term drivers as well as short-term technical factors such as positions, have been very well correlated to measures of European exogenous risk. This has been true for risky assets as well: as Exhibit 4 shows, our cross-sector spread index has been very well correlated to French bank CDS spreads. As a result, proxy variables for the European crisis have become part of our frameworks for thinking about valuations in various sectors. Amidst this backdrop of weak fundamentals and high geopolitical risk, regulatory and litigation risks have also been growing. In general, the pace of macro-prudential regulation has been slower than we anticipated, causing regulatory uncertainty to persist. For example, the implementation of Dodd-Frank has been slow, and the rulemaking process continues to suffer from considerable delays, perpetuating uncertainty. However, several critical changes have come to pass: US bank ratings have been downgraded as a result of lower systemic support, and the FDIC changed its deposit insurance assessment to shift more of the burden onto larger banks. One area where some progress has been made is money fund reform. Recently, SEC Chairman Mary Shapiro indicated that the Financial Stability Oversight Council is close to proposing rules for money fund reform, which we expect to be implemented by the end of 2012. At this point, we think the preferred route for regulation will be a capital buffer requirement, likely in combination with gating restrictions. However, we caution that in the current environment of very low yields it will likely be difficult for money funds to raise capital (see Short-Term Fixed Income). Litigation risk around banks has also taken a toll on asset markets, particularly in securitized product markets where banks have a large footprint. To see this, we note that in an environment where banks are very well capitalized, litigation risk should negatively impact bank equity prices more than bank credit. Thus, we define our legal risk metric as the residual of the log of bank equity
92
Exhibit 5: Litigation risk for banks has taken a toll on securitized products
Legal risk index* versus 1-week average of 30-year MBS Libor OAS; bp
0.15 0.10 0.05 0.00 -0.05 -0.10 -0.15 -0.20 May 10
Legal risk index
80 70 60 50 40 30 MBS OAS 20 10 Sep 11
Aug 10
Nov 10
Mar 11
Jun 11
* Legal risk index calculated as the residual of the log of the KBW bank stock index regressed against the average 5-year CDS spread for USD Libor panel banks** over the past two years (1-week averages are used). The residual is multiplied by -1, such that increasing values of the index indicate heightened risk. ** Excludes Bank of Nova Scotia.
prices regressed against the average CDS level for the USD Libor bank panel. (We multiply the residual by -1 such that higher values of our index indicate heightened legal risk.) As Exhibit 5 shows, our index of legal risk has been well-correlated to MBS spreads, with MBS cheapening when legal risk rises. All of these factors will continue to drive markets in 2012. In particular, Europe will very much continue to be a factor. To help gauge where we are in the crisis, we find it is interesting to compare the European crisis with the 2007-08 financial crisis in the US and in Japan in the early 1990s. Although the events that may have sparked the crises are not exactly the same (the bursting of real estate bubbles in the case of Japan and the US, and the bursting of a sovereign credit bubble in the case of Europe), the subsequent propagation and contagion, as well as the eventual measures that will be necessary to end the crisis, are similar. In Exhibit 6, we present a combined crisis timeline, which compares events in the Japan banking crisis to similar events in the US financial crisis and European sovereign crisis. As in the US and Japan, the European crisis began with localized losses (on subprime mortgages in the US and on peripheral sovereign bonds in Europe). As bank capital positions weakened in an environment of declining risky asset prices, banks began deleveraging. These actions now appear likely in Europe, as several banks have indicated that they expect to delever. In fact,
Exhibit 6: Side-by-side comparison of Japans banking crisis, the US financial crisis, and the European sovereign crisis
Timeline of key events in Japans banking crisis and comparable events in the US financial crisis and ongoing European sovereign crisis
JAPAN 1991-92 Peak in land prices Diet passes six laws and establishes Housing Loan Administration Corporation and Resolution Collection Bank to cope with liquidation and recovery of assets of failed jusen 1996 and cooperatives. Deposit Insurance Scheme strengthened Government to guarantee full amount of deposits in yen and other currencies, bank debentures and certain trusts offered Nov-97 by trust banks Two new laws passed including a provision of 30tn to Feb-98 recapitalize banks Mar-98 Government injects 1.8tn into large banks US Jul-06 Peak in housing prices Euro area Aug-09 Trough in sovereign CDS EU and IMF launch a 750bn stabilization scheme with Congress passes TARP/EESA to buy troubled assets from 10/03/08 financial institutions Under TARP/EESA, deposit limit increased to $250K for individuals; government still working on plan for unlimited 10/03/08 guarantee for small businesses government-backed loan guarantees and a commitment to 05/10/10 buy European sovereign bonds (Guarantee schemes from the 2008 financial crisis are still in place) EU officials call for a 106bn recapitalization of banks and plan to leverage the EFSF via a monoline insurance scheme 10/26/11 or via a leveraged CDO-style investment vehicle Jun-12 Recapitalization of banks to be completed
10/14/08 TARP modified to provide up to $250bn of capital for banks Government injects $125bn into 9 banks; offers another 10/14/08 $125bn to others FDIC to temporarily guarantee the senior debt of all FDICinsured institutions and their holding companies, as well as deposits in non-interest bearing deposit transaction accounts; Treasury establishes guarantee program for money market funds; CPFF created to provide a liquidity backstop for CP Oct-08 issuers 12/16/08 The Fed introduces a zero interest rate policy The Fed announces Treasury purchases and increased 03/18/09 purchases of Agency debt and MBS (QE1) Oct-Nov MMIFF created to purchase assets from money market 08 funds; TALF created to provide liquidity to ABS investors
Late Government offers 20tn for credit guarantee schemes, and 1998 and then introduces scheme to guarantee bonds of small and Sep-99 medium-sized compnaies Feb-99Aug-00 BoJ employs zero interest rate policy
EU officials discussed exploring a coordinated guarantee 10/26/11 scheme, but nothing concrete has been proposed ECB cut the refi rate by 25bp to 1.25% on 11/3/11, and we expect the refi rate to be cut to 0.5% by June 2012 ECB has been purchasing sovereign debt, but it is not "quantitative easing" since they have been sterilizing purchases
Mar-01 BoJ introduces quantitative easing Jun-03 BoJ starts puchasing ABS (including ABCP)
Source: Akihiro Kanaya and David Woo, The Japanese Banking Crisis of the 1990s: Sources and Lessons, IMF Working Paper, January 2000. Masaaki Kanno, How Japans banking crisis ended in a lost decade, J.P. Morgan Economic Research note, 10/31/08.
our European bank credit analysts estimate that the 28 large banks they cover will reduce assets by anywhere from around 830bn to almost 2tn over the next twelve months in order to raise capital ratios (see The Great Bank Deleveraging: Banking Sector Outlook 2012, Roberto Henriques et al, 11/4/11). Eventually, forced recapitalization through public capital was employed in the US and Japan. Although the EBA has already mandated that some banks raise 106bn of capital (rather than de-lever) to improve their capital ratios, we think another round of capital raising may be needed given that the EBAs estimate of capital needs is nearly 200bn below our estimates. Moreover, definitive public programs to enforce recapitalization (rather than deleveraging) have yet to be announced in Europe. However, in the US case, bank recapitalization alone was not enough to reopen broken primary markets and shortterm funding markets. The Fed established the CPFF to provide liquidity for CP issuers, and supported securitized product markets through TALF. In addition,
Treasury established a guarantee program for money market funds, and the FDIC established the TLGP to temporarily guarantee the senior debt of FDIC-insured institutions. Although some guarantee and deposit insurance schemes from the 2007-08 financial crisis are still in play in Europe, comprehensive programs spanning the Euro area have yet to emerge, and policymakers are currently focused on other mechanisms (such as providing monoline insurance) to allow for primary issuance by affected sovereigns. Of course, the rest is historyZIRP and quantitative easing followed suit in the US as well as in Japan. It is as yet unclear if and when this might occur in Europe, but the ECB did cut rates this month to 1.25%, and we look for the ECB to cut rates to 0.5% by June 2012. One last thing to note from our three-crisis comparison is the timescales. The pace with which policy measures are unfolding in Europe is clearly slower than was the case in the US, but hopefully faster than in the Japanese experience.
93
Look for further deterioration in Europe over the near term Our global strategists look for the following developments in Europe in 2012. First, we expect funding spreads on the semi-peripherals to worsen early next year. Second, Greece is likely to continue to fail to meet austerity targets, which could eventually lead to a more severe debt restructuring. Third, we think the Euro area will dip into recession, with the peripheral area hit hardest. This could make it difficult for EU sovereigns to meet their fiscal and structural reform targets. Fourth, we expect increasing risks of Portugal and Ireland needing to renegotiate their funding packages in mid- to late-2012, which will likely require private sector involvement. Finally, France is at risk of losing its AAA rating, especially if the Euro area economy weakens more than expected. Now that sovereign debt of many European countries is too low-rated to remain in portfolios seeking risk-free assets, these bonds will slowly but surely need to find new homes in risk-seeking investors portfolios. The US experience with GSE debt in 2008-09 (when foreign investors sharply scaled back on holdings of this asset class) makes for a useful comparison, and highlights the staying power of bearish technicals when such shifts are underway. Moreover, although the Fed was willing to intermediate the transfer, by stepping in to buy Agency debt, the ECB has thus far been reluctant to intermediate in the size required to stabilize spreads. Thus, we remain bearish on intra-EMU spreads and expect them to widen going into 1H12, before concerted policy action pushes them narrower in 2H12. Separate from our baseline view, it is interesting to note that our J.P. Morgan Fixed Income Markets Investor Survey indicates that 33% of investors believe Greece will exit the EMU by the end of 2012, while 25% of investors believe a country other than Greece will exit by the end of 2013 (see Appendix at end of piece for detailed results). Thus, with all these risks on the horizon, we expect European developments to remain a major driver for markets in 2012. The outlook for the economy and Fed policy Beyond European factors, the US growth outlook will also be an important driver. Although the super committee failed to agree on a deficit reduction proposal, given that the mandatory cuts will not begin until 2013, there is considerable uncertainty around the nature of fiscal policy adjustments (both potential stimulus
94
Exhibit 7: We expect growth to slow substantially in 1H12 due to fiscal drags and then recover in 2H12
4Q11 Real GDP (% q/q, saar) Unemployment rate (% ) Core CPI (% q/q, saar) 3.0 9.0 1.2 1Q12 0.5 9.0 1.2 2Q12 1.5 9.0 1.0 3Q12 2.5 9.0 1.2 4Q12 2.5 9.0 1.2
J.P. Morgan forecast for real GDP growth, the unemployment rate and core CPI
packages and potential deficit reductions) that may occur over the next year. In our baseline economic forecast, we assume no additional fiscal stimulus will be passed by the end of 2011. This is roughly in line with the results of our investor survey, which shows that 49% of investors believe no additional stimulus will be passed. However, if any fiscal stimulus is passed, we think it is unlikely to be large enough to completely offset the drag from past stimulus programs rolling off, which we estimate will reduce GDP by 1.7% in 2012 on a year-over-year basis. As a result, we look for growth to slow substantially in the first half of 2012 and then recover in the second half (Exhibit 7). With growth likely to weaken, we expect the unemployment rate to remain elevated, and high resource slack should push core and headline inflation lower. Thus, with the economy likely to weaken further, what will the Fed do? The results of our investor survey show that 80% of investors believe the Fed will embark on another round of balance sheet expansion via asset purchases (QE3) in 2012. We think the next policy action the Fed will undertake will most likely be the so-called Evans plan, which calls for specifying the economic conditions that need to be met for the Fed to tighten. That said, we also think there is a significant chance that QE3 will be deployed, especially in the form of MBS purchases, if inflation expectations fall enough. Thus, QE3 remains on the horizon, even if not yet a force in markets.
Implications for US fixed income markets

As mentioned above, Europe is likely to remain the biggest driver of US markets in 2012. In Treasuries, we think the worsening crisis will sustain the flight-toquality bid, pushing yields lower in early 2012 despite already-rich valuations. Some evidence of these flows can already be seen in data on foreign purchases: Europeans have increased purchases of Treasuries, likely reflecting asset allocation shifts out of riskier EU
sovereign debt. However, such effects are unlikely to be permanent. Correlations between US rates and EU sovereign spreads will decline eventually, and such portfolio rebalancing will eventually wane, likely later in 1H12. As this European demand subsides, the supplydemand imbalance in US Treasuries is likely to shift in an unfavorable direction, and we expect poor technicals and rich valuations to eventually cause yields to rise over the course of 2012 (see Treasuries). At the other end of the risk spectrum, emerging markets will also remain hostage to European developments thanks to heightened correlations. However, we think a number of factors make EM a relatively defensive asset class. In contrast to 2008, EMBIG yields and spreads are well above those ofsayhigh grade corporates, making relative valuations attractive. In addition, stable growth, greater policy flexibility, low refinancing needs, a diversified investor base and an upward ratings trajectory are all supportive factors. Thus, we look for 510% total returns in 2012. High yield investors face a similar backdrop. On one hand, Europe remains a risk, but on the other hand, we think default rates will remain low over the next few years due to strong liquidity and the health of corporate balance sheets. We project default rates of 1.5% and 2.0% for high yield bonds and loans, respectively, in 2012; even in a US recessionary scenario, we see default rates peaking near 5-6%, versus 10-15% peaks in prior recessions. High carry and modest spread tightening should produce attractive total returnwe expect returns to reach 9.6% in 2012. Within the sector, however, we remain biased towards greater tiering, and recommend underweighting CCCs versus higher rated credits. In particular, we find Bs to be the most attractive rating category as they offer the best balance between heightened macro risks, sound credit fundamentals and low default risk. In higher quality spread product, we recommend overweights for various reasons. In CMBS, we are cautiously optimistic given cheap fundamental credit risk and negative net issuance of private label CMBS. We favor new issue AAAs and subordinates relative to legacy CMBS. Although early 2012 will very likely experience ongoing spread volatility, over a 12-month horizon, we look for legacy benchmark AAAs to outperform corporates, tightening by 70bp by mid-year.
Similarly, in ABS, we think credit fundamentals will remain solid despite stubbornly-high unemployment and look for technicals to remain favorable due to a persistent supply shortage. Although we believe ABS spreads have limited incremental tightening potential from here, we do think some sectors offer value. Specifically, our top picks in ABS heading into next year in the AAA space are non-prime auto, retail cards and cross-border ABS. In addition, we like subordinate credit card and auto ABS. In Agencies, we are overweightalbeit cautiously so. Agency spreads remain vulnerable to a worsening of the European crisis, and our investor survey points to a desire to reduce exposure to Agencies; however, supply technicals are strong, as net issuance in 2012 will once again be significantly negative. In MBS, the prospect of the Fed buying MBS under a QE3 program is a powerful wildcard, and should limit the downside in the asset class. Given attractive spreads currently, we recommend heading into 2012 with an overweight. In high grade, the broad outlines of our view are similar to our view on rates; prospects for a near-term worsening in Europe keep us underweight in the near term, but the significant divergence between strong corporate credit metrics and wide spread valuations will eventually cause a decoupling, and likely cause spreads to narrow strongly at some point in the future. We forecast high grade bond spreads at 175bp at year-end 2012, versus spreads at 235bp today. Within the high grade market, we recommend underweighting Financials versus NonFinancials and underweighting high beta sectors and those most closely tied to Europe.
Contemplating risks in 2012

Although we are generally positive on risky assets over a medium-term horizon in 2012, given that markets face unusually high risks, we think it is useful to quantify how much various spreads might move in an environment of weakening growth and a worsening European crisis. To do this, we present a stylized model for spreads in various sectors. We use four factors to model spreads over the past two years. First, we use 4-week average jobless claims as a proxy for economic conditions. Second, we use the size of the Feds securities holdings as a proxy for QE. Third, we use the average CDS spread for French banks to account for the impact of the European crisis. Finally, we use an interaction term defined as the product of French bank CDS and our
95
Exhibit 8: Quantifying the risks for spreads
Current level of spreads; our strategists targets for mid-year 2012*; statistics for spreads regressed against 4-week average of jobless claims (000s), size of the Feds securities holdings ($bn), average French bank CDS** (bp), and average French bank CDS multiplied by flight-to-liquidity index*** using weekly data over past 2 years; projection for spreads if $500bn of QE3 occurs; 2-year standard deviation of 3-month changes, and normalized risks****
Current Mid-year Jobless claims Fed sec holdings French bank CDS FTLIxFrench bank CDS Projection SD Econ EU QE3 T-stat Beta T-stat Beta T-stat Beta T-stat R-sq with QE3 (bp) (bp) risk risk level (bp) target (bp) Beta risk 10Y AAA CMBS spd to swaps 320 250 3.09 8.8 -0.103 -2.4 0.24 1.2 0.365 3.5 71% 317 81 0.95 0.30 -0.64 JULI (ex-EM) spd to Tsy 235 213 0.46 6.4 -0.025 -3.0 0.35 8.1 0.085 3.9 80% 237 31 0.37 1.11 -0.40 JPM US HY index spd to worst 743 700 2.29 10.7 -0.085 -3.3 1.02 7.9 0.270 3.9 82% 789 95 0.60 1.07 -0.45 EMBIGLOBAL spd to Tsy 415 375 0.69 5.0 0.025 1.5 0.54 6.4 0.074 1.8 74% 446 48 0.36 1.12 0.26 US Financials spd to Tsy 346 300 0.77 7.3 -0.018 -1.5 0.63 10.0 0.139 4.3 86% 357 50 0.39 1.26 -0.18 5Y CDX.IG spd 5Y CDX.HY spd S&P 500 (points) 136 772 1216 118 650 1475 0.28 2.92 -2.13 6.0 9.6 -8.6 0.004 -0.008 0.182 0.6 -0.2 6.0 0.28 1.52 -1.16 9.6 8.3 -7.6 -0.005 0.228 0.044 -0.3 2.3 0.6 76% 78% 81% 142 811 1245 18 0.38 1.49 150 0.49 1.01 86 -0.62 -1.35 0.10 -0.03 1.06
* For JULI, CDX.IG, and CDX.HY, mid-year targets are interpolated based on year-end 2011 targets or current levels and year-end 2012 targets. For US Financials, target is based on the JULI target and the historical relationship between the two spreads. For the S&P 500, this is the year-end 2011 target. ** Average 5-year CDS spread for Societe Generale, Credit Agricole, and BNP Paribas. *** Average of the 2-year Z-scores of (i) spread between 3-month Fed funds strip and 3-month T-bill, (ii) market depth, which is calculated as the half the size of the sum of the top three bids and offers for the 5-year hot run Treasury note, averaged between 8:30am and 10:30am daily, and (iii) 5-year muni pre-refunded bond yields spread to 5year on-the-run Treasury yield. **** Econ risk assumes a 25K rise in the 4-week average of jobless claims. EU risk assumes a 100bp widening in French bank CDS. QE3 risk assumes the Feds balance sheet increases by $500bn. Normalized risks are calculated as the projected change in each of these scenarios is divided by the 2-year standard deviation of 3-month changes.
flight-to-liquidity index to account for the negative feedback loop between crisis-induced de-risking and declining liquidity. Exhibit 8 presents the statistics for our model. Although our model does not account for the idiosyncratic drivers of each market, it does allow us to make broad projections for spreads based on projections for the underlying drivers, andeven more importantlyassess the risk exposure of each sector to the underlying macroeconomic undercurrents. Exhibit 8 shows our models fair value estimate, based on current levels of the independent variables and also assuming the Feds security holdings increase by $500bn under QE3. In addition, we use the betas from our model to characterize the risk exposure in each sector to economic deterioration (proxied by jobless claims worsening by, say, 25K), the exposure to a worsening crisis in Europe (proxied by a 100bp rise in French bank CDS spreads), and the exposure to QE3 (assuming a $500bn increase in the Feds security holdings). To put the different asset classes on a comparable footing, we divide the projected changes by the 2-year standard deviation of 3-month changes in spreads. Based on these risk exposure metrics, we find that CMBS, high yield, and equities are most exposed to a weakening of the economy. In contrast, CMBS seems to be least exposed to European risks, whereas CDX.IG, US Financials, and equities are most exposed. Finally,
96
CMBS and equities appear poised to benefit the most if QE3 occurs.
and tail risks

As investors contemplate the tail risks to their portfolios in the year ahead, the European crisis perhaps looms largest in their minds; however, we think there are three other key risk scenarios that also merit focus. These include a hard landing in China and emerging Asia, a return of deflation fears in the US, and (in the opposite direction) the potential for high future inflation (or inflation expectations) as a result of easy monetary policy now. In order to assess the impact of these four scenarios on markets, we first pick proxy variables that would be expected to move significantly in these crisesvariables that we could view as almost definitional in characterizing each of the risk scenarios that are to be hedged. For example, we use an average of French bank CDS as a metric for the EU crisis and define our tail risk scenario with respect to this metric. We use a basket of EM currencies to proxy the health of EM Asian economies and in specifying the risk associated with that scenario. Third, we use 10-year Treasury yields as a reference variable for our deflation and high inflation scenarios.
Exhibit 9: The tail wagging the dog? Projected market moves in four tail risk scenarios
European crisis Stress beta 10Y Tsy yields (bp) 2s/10s Tsy curve (bp) 10s/30s Tsy curve (bp) S&P 500 (points) MSCI G7 Index (points) MSCI EM Index (points) Nikkei Index (points) EUR/USD USD/JPY AUD/USD Gold futures ($/t. oz) WTI oil futures ($/bbl) 3M fwd copper price ($/m ton) 5Y CDX.IG spread (bp) 5Y iTraxx Main spread (bp) 5Y CDX.HY price (points) -0.847 -0.644 0.021 -2.22 -1.95 -1.93 -18.1 Normal beta -0.501 -0.171 0.070 -1.19 -1.15 -1.30 -10.3 Stress move -105.8 -80.5 2.6 -278 -244 -241 -2264 0.019 -1.47 -0.077 244.69 -18.55 40.27 54.87 -12.26 Normal move -62.5 1.69 -21.3 3.78 8.8 -148 1.87 -144 1.70 -163 1.48 -1288 1.76 -0.055 -1.44 -0.047 116.52 2.10 -6.57 2.83 -514.84 3.76 60.72 70.70 -8.65 1.42 Stress Ratio beta -20.3 -6.22 -1.64 -99.7 -96.4 -127.5 -1024.7 -0.042 -2.487 -0.062 -49.40 -9.59 -714.5 31.6 29.2 -2.70 EM Asia hard landing Normal Stress beta -15.4 -6.02 -2.07 -83.6 -82.0 -109.6 -827.7 -0.046 -1.349 -0.054 -37.87 -7.02 25.2 23.4 -4.23 move -52.2 -16.0 -4.2 -256 -248 -328 -2636 -0.107 -6.40 -0.160 -127.1 -24.7 81.2 75.2 -6.95 Normal move -39.6 1.32 -15.5 -5.3 -215 1.19 -211 1.18 -282 1.16 -2129 1.24 -0.118 -3.47 1.84 -0.139 1.15 -97.4 1.30 -18.1 1.37 64.9 1.25 60.2 1.25 -10.89
Statistics for various market variables regressed against our proxy variables, in normal periods and stress periods*; projected moves in our risk scenarios**, and ratio of stress move to normal move
High inflation Stress Ratio beta 46.2 -14.3 46.9 42.3 48.2 337.7 2.72 Normal Stress beta 41.2 -14.9 96.1 83.2 88.1 894.2 4.08 move 110 50.8 -15.8 51.6 46.6 53.0 371.4 0.017 3.00 0.031 28.4 10.4 416.7 -21.0 -22.9 4.63 Normal move Stress Ratio beta 74.4 4.2 241 206 188 1235 1.48 Deflation/low yields Normal Stress Normal beta 41.2 -14.9 96 83 88 894 4.08 move -80 -59.5 -3.4 -193 -164 -150 -988 -1.19 move Ratio
45.3 1.12 -16.4 105.7 91.5 96.9 983.6 0.004 4.49 0.033 -13.8 -2.07 8.8 1.18 675.4 -28.9 -31.3 5.83
-33.0 1.81 11.9 -77 2.51 -67 2.47 -70 2.13 -715 1.38 -3.26
0.00015 -0.00044 -0.01176 -0.01156 -0.00062 -0.00038 1.959 0.933 -0.149 -15.51 0.32 0.44 -0.098 -0.053 0.49 0.57 -0.069
0.0157 0.0036 0.0286 0.0303 25.9 -12.5 9.47 378.8 -19.1 -20.9 4.21 8.01 614.0 -26.3 -28.4 5.30
0.0563 0.0036 -0.045 -0.003 0.0899 0.0303 -0.072 -0.024 -113.0 -12.5 90.4 10.0 19.01 1131.9 -53.3 -43.6 4.05 8.01 -26.3 -28.4 5.30 -15.2 42.6 34.9 -3.24 -6.4 2.37 21.0 2.03 22.7 1.53 -4.24 614.0 -905.5 -491.2 1.84
-4.12 -1937.45
-566.0 -1838.1 -1456.1 1.26
* For the European crisis, the stress beta is calculated as monthly changes in market variables regressed against monthly changes in average French bank CDS*** over 7/1/119/10/11, while the normal beta is calculated over the past 5 years. For the EM Asia hard landing scenario, the stress beta is calculated as weekly changes in market variables regressed against weekly changes in an EM currency basket**** over 8/1/08-11/30/08 and 8/1/11-10/31/11, while the normal beta is calculated over the past 5 years. For the high inflation scenario, the stress beta is calculated as weekly changes in market variables regressed against weekly changes in 10-year Treasury yields (%) over 5/16/07-6/22/07, 5/5/09-8/7/09, and 6/10/11-6/29/11, while the normal beta is calculated over the past 5 years. For the deflation/low yield scenario, the stress beta is calculated as weekly changes in market variables regressed against weekly changes in 10-year Treasury yields (%) over 4/29/10-9/1/10 and 7/25/11-10/3/11, while the normal beta is calculated over the past 5 years. ** For the European crisis scenario, we assume the French bank CDS average rises 125bp. For the EM Asia hard landing scenario, we assume the EM FX basket cheapens 2.6 points. For the high inflation scenario, we assume 10-year Treasury yields rise 110bp, and for the deflation scenario, we assume 10-year yields fall 80bp. Then we calculate projected moves using both the stress beta and normal beta. *** Average 5-year CDS spread for Societe Generale, Credit Agricole, and BNP Paribas; in bp. **** Average of USD/BRL, USD/INR and USD/MXN.
Next, we define our tail risk scenarios based on changes in these variables. For the Europe and EM scenarios, we define the tail risk as a two standard deviation move in the reference variables, which works out to a 125bp widening of French bank CDS and a 2.6 point rise in our EM currency index. For the deflation risk scenario, we looked at the beta between 10-year yields and 5Yx5Y inflation swap rates during periods when both declined sharply. Then, if we assume that forward inflation swap rates fall 1.25%, we project that 10-year yields fall 80bp; this is our specification for the deflationary scenario. Finally, for the high inflation scenario, we looked at the relationship between 10-year yields and headline CPI in 1971-74 and 1977-80, two periods when CPI surged over 8%-points. We then estimate that if headline inflation doubles from current levels, 10-year Treasury yields should rise 110bp. In general, the best ways to position for tail risk scenarios involve taking advantage of heightened correlation in these crisis periods. Therefore, in order to identify the best hedges for these four tail risk scenarios, we
calculated empirical betas between a number of market variables and our proxy variables, both in normal periods and stress periods, which are periods when our proxy variables moved significantly. Exhibit 9 presents the results of our analysis, as well as the projected moves for the various market variables in each risk scenario. We also show the ratio of the stress move to the normal move for the market variables that show substantial increases in correlation in crises. We can make several observations based on this table. First, we find that in general, currencies are poor hedges for crises since stress betas were not meaningfully higher than normal period betas, perhaps reflecting the impact of FX intervention. Second, for the European crisis scenario, we find that longs in 10-year Treasuries and 2s/10s curve flatteners offer attractive beta pickup. Equities, commodities, and high yield also show increased correlation in an intensifying European crisis, but interestingly, high grade credit does not.
97
Exhibit 10: Indicative prices for digital and dual digital options that we expect to pay out under our stress scenarios
EUROPEAN CRISIS Instrument 10Y Tsy yields (bp) 2s/10s Tsy curve (bp) S&P 500 (pts) Gold futures ($/t. oz) WTI futures ($/bbl) Copper fut. ($/m ton) Instrument #1 Gold futures ($/t. oz) WTI futures ($/bbl) Copper fut. ($/m ton) Strike Payoff if -105 Below -80 Below -275 Below 245 Above -18.5 Below -1950 Below Strike Payoff if Instrument #2 115 Above S&P 500 (pts) -6.5 Below S&P 500 (pts) -515 Below S&P 500 (pts) Premium 17.6% 11.4% 22.0% 30.0% 35.3% 33.5% Strike Payoff if Premium -145 Below 15.6% -145 Below -145 Below 23.8% 23.5% Premium 42.4% 10.7% 19.0% 50.0% 29.1% 34.1% Strike Payoff if Premium -3.5 Below -95 Below -40 Below -3.5 Below 12.5% 11.3% 25.0% 23.0%
HIGH INFLATION Instrument Gold futures ($/t. oz) WTI oil futures ($/bbl) 10Y Tsy yields (bp) Instrument #1 Gold futures ($/t. oz) DEFLATION/LOW YIELDS Instrument Strike Payoff if 2s/10s Tsy curve (bp) -60 Below S&P 500 (pts) WTI futures ($/bbl) Copper fut. ($/m ton) CDX.IG spd (bp) 10Y Tsy yields (bp) Instrument #1 S&P 500 (pts) S&P 500 (pts) S&P 500 (pts) -190 Below -15 Below -900 Below 40 Above -80 Below
Instrument, strike*, payoff if market is below or above the strike, and indicative premium for 1-year digital or dual digital option, with USD payouts (where applicable); the most attractive trades for a given panel are highlighted
Strike Payoff if 30 Above 10.5 Above 110 Above Strike Payoff if Instrument #2 15 Above WTI futures ($/bbl) Premium 46.0% 44.7% 13.3% Strike Payoff if Premium 9 Above 34.6% Premium 15.4% 28.1% 37.5% 39.3% 42.0% 27.4%
EM ASIA HARD LANDING Instrument Strike Payoff if 10Y Tsy yields (bp) Nikkei Index (points) USD/JPY Gold futures ($/t. oz) WTI futures ($/bbl) Copper fut. ($/m ton) Instrument #1 Nikkei Index (points) Nikkei Index (points) Gold futures ($/t. oz) Copper price ($/m ton) -50 Below -2600 Below -6.5 Below -125 Below -25 Below -1840 Below Strike Payoff if Instrument #2 -2100 Below USD/JPY -2100 Below Gold futures ($/t. oz) -95 Below 10Y Tsy yields (bp) -1450 Below USD/JPY
Strike Payoff if Instrument #2 Strike Payoff if Premium -75 Below WTI futures ($/bbl) -6.5 Below 27.8% -75 Below Copper fut. ($/m ton) -490 Below -75 Below CDX.IG spd (bp) 20 Above 28.9% 42.0%
* Strike is ATM spot + indicated value for equity indices, and ATM forward + indicated value for all others. Note: Swaptions are used instead of Treasury options.
Third, in the EM Asia hard landing scenario, we find increased correlation for 10-year Treasuries, equities and commodities, but not for high yield. This scenario is the only one in which currencies serve as effective hedges: the yen typically strengthens as Japanese investors unwind long-EM trades, and the Aussie dollar typically weakens along with other EM/commodity currencies. Fourth, we find that hedging against high inflation is difficult, with very few market variables displaying increased correlation in this scenario. Only 2s/10s curve steepeners and longs in gold or oil show up as attractive hedge trades. In contrast, several market variables show increased correlation in a deflation scenario, with selling equities and commodities among the most attractive trades. It is also interesting to note that the beta between equities and yields in a deflation scenario is higher than the beta in inflation scenarios. This is likely because a weak growth/deflation scenario is clearly negative for equities, whereas the effect of high inflation on equities is mixedon one hand, equities are real assets, so equity
prices should rise with inflation, but on the other hand, increased risks of Fed tightening should constrain prices. Now that we have defined our scenarios and our projected market moves, in order to evaluate which hedges are currently most attractive, we look at digital and dual digital option prices for these market variables. Digital options, which give a payoff of 1 if a certain condition is met, are insightful because they may be interpreted as the market-implied probability of a given scenario. In Exhibit 10, we show indicative prices for several digital options and dual digital options that are potential hedge trades in our four scenarios. We have chosen the strikes for the single digital options based on the expected stress moves of market variables, while the strikes for the dual digital options are based on the normal moves as implied by Exhibit 9 and reflect moves that are highly likely to occur in a given risk scenario. The rationale for choosing less out-of-themoney strikes in dual-digitals is this: although each move may be more probable individually, as well as jointly in a given tail risk scenario, the dual digital can often trade
98
Exhibit 11: Summary of the best tail risk hedges

Summary of the best trading themes for each risk scenario based on Exhibit 10 Scenario Best trading themes European crisis 2s/10s curve floors, yield floors, S&P puts EM Asia hard landing Nikkei puts, USD puts/JPY calls High inflation Payer swaptions Deflation/low yields 2s/10s curve floors, yield floors, S&P and commodity puts
Exhibit 12: We expect total net issuance across fixed income markets to decline in 2012
Net ($ bn) IG corporates HY corporates EM corporates Municipals
Historical data and J.P. Morgan forecast for long-term (>1Y) net issuance in 2011 and 2012; $bn
2004 2005 2006 2007 2008 2009 2010 2011 E 2012 E 215 -17 54 166 423 -24 79 1 25 116 296 169 -25 76 156 616 97 97 50 43 54 263 326 15 102 173 539 269 165 38 86 86 177 408 6 117 38 201 510 175 47 76 -64 135 103 -25 17 29 -359 519 -40 -50 5 -39 396 556 95 76 80 108 -379 524 -47 -25 -17 -19 247 151 150 34 -274 -80 -57 -96 -15 -119 353 100 141 -71 -218 0 -46 -50 -8 -162 303 85 129 3 -186 -55 -17 -50 -13 -185
cheap because the correlation is underpriced in normal markets. As such, cheap dual digitalsbased on less extreme strikeswill give an indication of which pairwise correlations are underpriced currently in markets, relative to what can be expected in a tail risk scenario. For example, our analysis projects that the S&P 500 will fall 278 points if the European crisis worsens significantly, compared to a 148-point fall using normal betas. One way to hedge for such a move in the S&P 500 is to buy a 1-year digital option that pays out if the S&P 500 falls 275 points. We estimate that the premium on this option is approximately 22.0%. Alternatively, one can buy a dual digital option, which pays out only if two conditions are met. In this case, we use two weaker conditions, both of which would be very likely to occur in the risk scenario. For example, one can buy an option that pays out only if the S&P 500 falls to a certain level and gold futures rise to a certain level. Since we project that the S&P 500 will fall 148 points and that gold futures will rise $116.5 based on typical betas, we consider a dual digital option that pays out only if the S&P 500 falls at least 145 points and gold futures rise as least $115. Interestingly, even though the dual digital has strikes that are closer to at-the-money, the premium on the dual digital option is cheaper15.6% versus 22.0% for the single S&P 500 option. This suggests that the correlation between S&P 500 and gold in a European crisis scenario as predicted by our analysis is not priced into markets. Put differently, tail risk is priced in to a greater extent in the S&P implied distribution, but less priced into the S&P/gold correlation market. Indeed, we find that the correlations between commodities and other variables appear underpriced versus what our model predicts for all four scenarios. In each case, a dual digital structure that pairs a commodity with another variable is cheaper than a single digital option on that commodity. We can draw several other conclusions from the table. First, we find that floors on the 2s/10s Treasury curve
Non-Agency MBS Agency MBS CMBS ABS CLOs Agency Debt Treasuries Total
1,549 1,611 1,364 1,036 1,945 1,552 1,403 1,050
1,334 1,596 1,976 1,649
appear to be the most cost-efficient way to position for a deterioration of the European crisis. The dual digital with gold and S&P 500 is the second cheapest way to hedge the European crisis, and is cheaper than all the other single digitals we considered. Second, the best ways to position for economic weakness in EM Asia tend to involve exploiting correlations with other liquid markets, particularly the Nikkei Index. We find that puts on the Nikkei are the most cost-effective way to hedge for EM Asia weakness, followed closely by dual digitals involving the Nikkei and gold futures or the Nikkei and USD/JPY. Third, for the two inflation scenarios, rate hedges are the most attractivefor the high inflation scenario, outright payer swaptions are most attractive, and for the deflation scenario, floors on the 2s/10s Treasury curve are cheapest. This is likely because changes in inflation expectations have the greatest impact on rates as opposed to other asset classes. Exhibit 11 presents a summary of the best trading themes for each risk scenario.
Trends in supply and demand in fixed income markets

Net issuance across fixed income markets averaged around $1.5tn over 2010 and 2011, down from the nearly
99
$2tn peak reached in 2009. Next year, we expect net issuance to take another leg down, driven primarily by a sharp decline in net issuance of long-term Treasuries (Exhibit 12). Although we expect gross issuance of coupon Treasuries to be virtually unchanged in 2012, the net number will likely fall about $328bn due to increased redemptions. Muni issuance, on the other hand, is projected to rebound from -$71 to $3bn. Finally, we expect net issuance across corporates and securitized products to remain roughly the same at $520bn and -$320bn, respectively. How does our supply projection compare to expected demand? In Exhibit 13, we present our projections for the sources of organic demand for fixed income assets in 2012 from six major classes of investors: mutual funds, foreigners, insurance companies, commercial banks, pension and retirement funds, and the Federal Reserve. Clearly, prices in markets will adjust so that demand will meet supply, but we think it is useful to estimate the demand flows that might be expected given the current state of markets. First, we can think of mutual fund purchases of fixed income assets as being driven by two factors: on one hand, as the economy weakens, mutual funds will likely increase their purchases of bonds, and do the opposite if economic prospects improve. Second, as front-end rates fall, investors will likely shift out of money market funds and into bond mutual funds in order to pick up yield. Thus, we find that using initial jobless claims as a barometer for the economy and 3-month T-bill yields as a proxy for money market rates does a good job of explaining changes in mutual fund holdings of fixed income over the past 25 years. If we assume that initial jobless claims rise slightly as the economy weakens in 1H12 and that T-bill yields remain near current levels, then we project that mutual funds will add about $230bn of fixed income assets in 2012, close to the average annual purchase over 2006-10. Second, we model changes in foreign holdings of fixed income assets by regressing annual changes in holdings versus the annual trade deficit of the US. Given that we expect the trade deficit to widen modestly next year, to about $582bn, we project $585bn of net purchases by foreigners in 2012. This is slightly below the net purchases we have seen over 2006-10. Third, we model flows from insurance companies using the slope of the 10s/30s curve (with a steeper curve leading to greater purchases) and the backward-looking
100
Exhibit 13: Expected 2012 fixed income purchases by various investor classes
Statistics for regression models for quarterly changes (1-year average) or annual changes in fixed income holdings for various investor classes, estimated purchases based on forecast for variables, expected purchases in 2012, and average actual annual purchase over 2006-10 ; $bn MUTUAL FUNDS Period 3Q86 -2Q11 R-sq 58% Factor Beta T-stat Forecast Initial jobless claims (1Y avg; 000s) 0.12 3.7 420 3-month T-bill yield (1Y avg; % ) -5.82 -6.5 0.11 Intercept 8.22 0.6 Estimated 2012 purchases ($bn) 228 Actual avg purchase over 2006-10 255 FOREIGNERS Period 1992-2010 R-sq 66% Factor Beta T-stat Forecast Annual trade balance ($bn) -0.94 -5.7 -582 Intercept 39.67 0.6 Estimated 2012 purchases ($bn) 585 Actual avg purchase over 2006-10 647 INSURANCE COMPANIES Period 3Q86-2Q11 R-sq 55% Factor Beta T-stat Forecast 1Y lag 10s/30s Tsy curve (1Y avg; % ) 23.46 5.2 1.10 1Y lag 1Y chg in S&P 500 (1Y avg; pts) -0.06 -6.8 140 Intercept 23.35 11.1 Estimated 2012 purchases ($bn) 163 Actual avg purchase over 2006-10 86 COMMERCIAL BANKS Expected 2012 purchases ($bn) 80 Actual avg purchase over 2006-10 113 PENSION AND RETIREMENT FUNDS Expected 2012 purchases ($bn) 135 Actual avg purchase over 2006-10 138 FEDERAL RESERVE Expected 2012 purchases ($bn) 0 But some chance of QE3 Actual avg purchase over 2006-10 282
TOTAL EXPECTED PURCHASES ($bn) 1191 Note: Insurance companies consist of life insurance and property & casualty insurance companies. Pension and retirement funds include private pension funds, as well as state, local, and federal government retirement funds. Source: Federal Reserve Flow of Funds
performance of the S&P 500 (with positive past performance leading to lower fixed income purchases). Given that the 10s/30s curve has averaged around 1.15% over the past year, and given that quarterly changes in the S&P have averaged around 140 points, we project that insurance companies will purchase around $165bn fixed income assets in 2012. This is substantially higher than the average over 2006-10, but slightly below the $200bn
and $188bn of purchases seen in 2009 and 2010, respectively. For banks and pension funds, we assume that they maintain their pace of purchases over recent years. Banks added $83bn of fixed income securities in 2010 and are on track to add a similar amount in 2011, so we estimate that banks will purchase around $80bn of fixed income assets in 2012. Similarly, pension funds have been steadily adding around $135bn of fixed income assets each year, so that is our estimate for 2012 purchases. Finally, our baseline view is that the Federal Reserve will not (yet) embark on QE3, but there is a significant risk that it will sometime in 2012, particularly involving mortgages. All in all, we estimate net organic demand for fixed-income assets based on current conditions to be about $1200bn in 2012, with additional upside risk due to the Fed. This is comfortably above our net supply forecast. Aggregate measures of net demand and net supply do not tell the whole story, however. The majority of fixed income investor classes discussed in our table are natural buyers of fixed income spread product; in contrast, net supply in 2012 will come almost entirely from Treasuries. With Fed purchases of Treasuries likely to be zero on a net basis in 2012 (based on our assumption that if QE3 occurs, it will be in MBS), the Treasury market will likely witness a deteriorating supply/demand imbalance as we move through 2012 (see Treasuries).
Exhibit 14: Investors plan to reduce exposure to European assets and add exposure to US and EM local currency assets
Weighted expected change in exposure for assets in various currencies according to our J.P. Morgan US Fixed Income Markets Investor Survey
0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 -0.07 EUR assets Other DM 0.07 0.08
0.16
0.17
EM USD USD assets EM local FX assets assets * To calculate the weighted expected change in exposure for each asset class, we assigned values of 1, 0, and -1 to responses of add, maintain, and reduce exposure, respectively, and then we summed the values and divided by the number of responses.
Exhibit 15: Investors plan to add exposure to corporate credit risk and reduce exposure to Agency debt and cash
Weighted expected change in exposure* for various asset classes according to our J.P. Morgan US Fixed Income Markets Investor Survey
0.30 0.16 0.04 0.07 0.09 0.09 0.19 0.20 0.20 0.24
0.3 0.2 0.1 0.0 -0.1 -0.2

-0.19 -0.19 -0.04
Commodities
Non-Agy MBS
Dur'n risk
CMBS
TIPS
Equities
Agy MBS
ABS
In our J.P. Morgan US Fixed Income Investor Survey, we also asked investors about the changes they plan to make to their positions over the next year. In Exhibit 14, we show the weighted average response for expected changes in portfolio currency allocations. Unsurprisingly, investors generally plan to reduce exposure to European assets and add exposure elsewhere. In particular, most investors indicated that they planned to add exposure to emerging market assets (denominated in local currencies) and US assets. Exhibit 15 shows how investors plan to change their asset class exposures over the next year. According to our survey, investors generally plan to reduce exposure to Agency debt and cash, with a smaller fraction planning to reduce duration risk on net. At the opposite end of the spectrum, investors generally plan to increase
* To calculate the weighted expected change in exposure for each asset class, we assigned values of 1, 0, and -1 to responses of add, maintain, and reduce exposure, respectively, and then we summed the values and divided by the number of responses.
exposure to high yield and investment grade corporates. They also plan to add exposure to Agency MBS, perhaps in anticipation of outperformance due to additional Fed purchases of mortgages. It is interesting to note that only a small fraction of investors plan to add exposure to equities and commodities this year, since real assets like equities and TIPS were investors top picks last year (see US Fixed Income Markets 2011 Outlook, 11/24/10).
Agy debt
Cash
EM
HY
101
IG
Investors asset allocation plans for 2012
Exhibit 16: The defined benefit pension funding gap has nearly doubled during 2011
Assets minus liabilities for top 100 defined benefit pension plans; $bn
Exhibit 17: Pension fund asset breakdown

Fixed income Tsys GSE-backed securities Corporates/foreign bonds Equities Other Total assets 353 121 374 1095 272 2215 848
Assets of private defined benefit pension funds as of 2Q 2011; $bn
0 -100 -200 -300 -400 -500 Oct 08
Source: Federal Reserve Flow of Funds Table L.118.b. Note Equities includes equities and mutual fund holdings.
Exhibit 18: Estimated empirical duration of pension fund liabilities

Monthly changes in top 100 D.B. pension fund liabilities regressed against monthly changes in 30-year swap rates (%) and monthly changes in JULI AA spread to Libor (bp); monthly data over 11/08-10/11; $bn
May 09
Nov 09
Jun 10
Dec 10
Jun 11
Source: Milliman 100 monthly Pension Funding Index
150 100
Y = -127.8(swap rate chg) - 1.07(JULI spd chg) + 0.59 R-sq = 75%
ALM strategies for pension funds

Over the first half of 2011, the funding gap for defined benefit pension funds fluctuated around -$200bn, then it deteriorated sharply as equities and rates both tumbled in the second half of the year. As of the end of October, the funding gap stood at -$398bn, nearly double its value at the start of the year (Exhibit 16). This dramatic deterioration highlights the risks of running a significant asset-liability duration mismatch. Indeed, the most recent Flow of Funds data on defined benefit pension fund assets showed that funds had 49% of their assets invested in equities and mutual funds and only 38% invested in fixed income securities (Exhibit 17). If we assume that fixed income assets mirror the composition of an aggregate index like the J.P. Morgan Global Aggregate Index (GABI), then given the 38% fixed income allocation, the estimated duration of assets would be only 1.9 years as of the end of October. On the other hand, if we regress monthly changes in D.B. pension fund liabilities against monthly changes in 30year swap rates and AA rated corporate bond spreads, we estimate an empirical duration of $128bn per 1%-pt change in 30-year swap rates (Exhibit 18). Given aggregate liabilities of about $1615bn currently, this relationship implies the duration of liabilities is around 7.9 years, substantially greater than the estimated asset duration of 1.9 years. One way to address this asset-liability duration gap is to increase allocations to fixed income, and specifically long-duration assets, since the duration of pension fund
102
50 0 -50 -100 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1-mo chg in 30Y swap rate; %
Exhibit 19: Issuance of long-term debt will likely increase in 2012, but remain below historical averages
Gross issuance of long-term (>10Y) debt, actual and forecast for 2011 and 2012, and debt currently outstanding with >10Y to maturity*; $bn
Gross issuance of long-term debt Current 2005 2006 2007 2008 2009 2010 2011 E 2012 E outstanding 14 41 42 49 143 183 191 194 830 67 104 162 95 118 120 85 115 1080 334 415 320 465 373 578 332 477 303 564 301 605 135 411 169 478 1678 3588
Treasuries Corporates Municipals TOTAL
* For Treasuries, we exclude the amounts held by the Fed. Corporate issuance is fixed-rate only.
liabilities is longer than the aggregate duration of the US fixed income universe. However, the prospects for better ALM-matching using cash assets may be difficult given our expectation of low long-end fixed income supply. As Exhibit 19 shows, long-end gross issuance fell sharply in 2011 due to declines in municipal and high grade corporate supply. Although we expect long-end gross issuance to increase in 2012, it will likely remain below
Exhibit 20: Strategies to match the duration of liabilities had significantly lower tracking error than the actual assets
Standard deviation of the net P/L of three strategies* and of the actual change in assets minus liabilities**; 1/10 10/11; $bn 55 50.1 50
Exhibit 21: The pension funding gap is exposed to rates, credit spreads, and equities
Monthly changes in top 100 D.B. pension fund funding gap regressed against monthly changes in 30-year swap rates (%),S&P 500 (points), and JULI AA spread to Libor (bp); monthly data over 11/08-10/11; $bn
100 50 0
45 40 35 30 25 20 Futures Swaps Swaps + CDX Actual assets

* Net P/L defined as monthly P/L on assets plus any overlays, net of monthly change in value of liabilities. Strategies match the duration and spread exposure as given by rolling 3-year regressions of monthly changes in liabilities versus monthly changes in 30-year swap rates and AA corporate spreads to Libor. They involve 1) buying 30-year (classic bond) Treasury futures, 2) receiving fixed in 2Mx30Y swaps, and 3) receiving fixed in 2Mx30Y and buying risk (selling protection) in 5-year CDX.IG. We assume monthly rebalancing. ** Change in assets minus change in liabilities for top 100 defined benefit plans as given by Milliman 100 monthly Pension Funding Index.
Y = 134.9(swap chg) + 0.61*(S&P chg) + 1.36(JULI spd chg)+3.95 R-sq = 85%
29.5 26.5 24.5
-50 -100 -150 -1.5 -1.0 -0.5 0.0 1-mo chg in 30Y swap rate; % 0.5 1.0
Exhibit 22: Strategies to hedge the funding gap substantially lowered the volatility of the gap
Standard deviation of monthly changes in gap plus P/L on hedge strategy* and of actual monthly change in funding gap**; 1/10-10/11; $bn
55 50 45 40 35 30 25 20 Swaps+ CDX+ S&P Futures+ CDX+ S&P Chg in gap

* Strategies hedge the duration, equity and spread exposure as given by rolling 3-year regressions of monthly changes in the funding gap versus monthly changes in 30-year swap rates, the S&P 500, and AA corporate spreads to Libor. They involve 1) buying 30-year (classic bond) Treasury futures or receiving fixed in 2Mx30Y swaps, 2) buying risk (selling protection) in 5-year CDX.IG, and 3) selling S&P 500 futures. We assume monthly rebalancing. ** Change in assets minus change in liabilities for top 100 defined benefit plans as given by Milliman 100 monthly Pension Funding Index.
50.1
the average levels seen over 2006-10. Furthermore, the Feds Operation Twist will likely take another $100bn of long-end Treasuries out of the market in December through June. In such an environment of constrained long-term debt, we think synthetic overlay strategies to increase duration and spread exposure are attractive. We consider two sets of strategies. The first set of strategies is designed to match the duration exposure of the liabilities. This can be thought of as the end game position that pension funds undertaking ALM strategies aim to reach, and represents the benchmark for any new investments that a pension fund manager might use. The second set of strategies is designed to hedge the funding gap itself, and could be useful to a fund manager seeking to hedge all or part of a pension funds current funding gap. For the liability-matching strategies, we examine three variations of overlay strategies: 1) buying 30-year (classic bond) Treasury futures, 2) receiving fixed in 30year swaps, and 3) receiving fixed in 30-year swaps and buying risk (selling protection) in 5-year CDX.IG. We calculated the monthly returns on these three strategies assuming monthly rebalancing, whereby the duration and credit exposures were given by the rolling 3-year betas of
28.1 25.1
changes in liabilities regressed against changes in 30year swap rates and AA rated corporate bond spreads (as shown in Exhibit 17). We then compared the returns on these strategies to the actual changes in liabilities, and in Exhibit 20, we show the monthly tracking error over January 2010-October 2011. As the exhibit shows, all three synthetic strategies showed substantially lower magnitudes of tracking error. Unsurprisingly, the strategy
103
involving swaps and CDX.IG had the lowest tracking error, since it hedged both the duration and credit exposure of the liabilities. The outperformance of the strategy using swaps versus the strategy using futures may be in part explained by the fact that we used 30-year swap rates in our models to determine the duration exposure, while the bond futures contract (currently the most liquid long-end Treasury futures instrument) creates exposure to the 15-year point. To hedge the funding gap, we first have to identify the drivers of the gap. Exhibit 21 shows that changes in the funding gap have been well explained by changes in 30year swap rates, AA rated corporate bond spreads, and the S&P 500, with increases in these factors leading to an improvement in the gap, and decreases leading to a worsening of the gap. Thus, the strategies to hedge the funding gap involve positions with the opposite exposures, i.e., a long duration position (either via 30year swaps or bond futures), a long risk position in CDX.IG, and a short in S&P futures. We calculated the returns on these strategies assuming monthly rebalancing based on the betas from rolling 3-year regressions, and we combined them with the actual changes in funding gap. As Exhibit 22 shows, using these hedge strategies significantly lowered the monthly volatility of the funding gap. As with the first set of strategies, the outperformance of the strategy using swaps is likely partly because 30-swap rates are used in our empirical estimation of the duration exposures.
nominal rates rise. QE3 also poses an upside risk to our targets for breakevens. Position for wider 10-year maturity matched swap spreads in the near term, but look to initiate narrowers towards the end of 1Q12 In the near term, FRA-OIS spreads will likely continue to widen, and banking stock valuations will likely remain under pressure, both of which should pressure 10-year swap spreads wider. We expect 10year swap spreads to hover near 27bp towards the end of this year and in early 1Q12. Beyond that point, issuance-related swapping is likely to pick up, and the growing likelihood of concerted action by central banks should help spreads narrow. Look for narrowing to 18bp by mid-year. Stay long gamma going into year end and in 1H12 A persistent crisis in Europe and poor risk appetite should cause market depth to stay depressed for much of 4Q11 and 1H12, creating conditions favorable for long gamma positions. Modestly overweight Agency bullets tactically hedged with swap spread wideners With 5-year Agency spreads wider than our T+33bp fair value target for 2Q12, we recommend an overweight position. However, given the significant impact of the European crisis on Agency spreads, we recommend tactically overlaying swap spread wideners as a hedge. In particular, we like adding to front-end bullets (2- to 3-year) on dislocations. We also recommend that investors use short-dated callables to enhance excess returns versus lockoutmatched bullets. Given greater clarity around shortterm rates due to the Feds low-for-long policy, we would look at structures with lockouts as far out as 1215 months. Modestly overweight MBS going into the new year We recommend beginning the new year with an overweight given solid fundamentals and the risk of QE3 in mortgages. Technicals should also be supportive given negative net supply. However, we only look for 10bp of tightening in the basis due to lower leverage and higher capital requirements. Add CMBS versus corporates We position relatively defensively and favor new issue AAAs and subordinates relative to legacy CMBS. For investors looking for yield in legacy paper, add exposure to higher-quality 2005/2006 AJs but avoid
Cross sector trading themes

Stay long duration early in 2012 targeting 10-year yields to reach 1.70% in 1Q12 A worsening of the European sovereign debt crisis should accelerate the flight-to-quality bid into Treasuries early next year pushing yields lower; rich valuations, a deteriorating supply/demand imbalance, and poor technicals will limit the upside, however, and keep rates range bound later in 2012. Look for TIPS breakevens to narrow sharply over the next few months and then widen modestly over 2012 Breakevens will hit a number of headwinds in early 2012: we expect nominal yields to plummet, fiscal policy to tighten, and headline inflation to fall. Thus, our fair value model projects that breakevens should narrow over the next few months, though breakevens should widen over the remainder of the year as
104
2007 AJs given higher downside in worse-thanexpected economic scenarios. While early 2012 will experience ongoing spread volatility, over a 12-month horizon, look for legacy benchmark AAAs to outperform corporates, tightening by 70bp by midyear. We are neutral on ABS ABS spreads likely have limited incremental tightening from here and tiering will be pronounced. Within ABS, our top picks in ABS heading into next year in the AAA space are non-prime auto, retail cards and cross-border ABS. In addition, we like subordinate credit card and auto ABS. Stay underweight high grade corporates but look for spreads to tighten later in 2012 Although we are currently underweight, we look for spreads to tighten sharply at some point in the future given the divergence between strong corporate credit metrics and wide spreads. We forecast high grade bond spreads at 175bp at year-end 2012, versus spreads at 235bp today. On a sector basis, we recommend underweighting Financials versus NonFinancials and underweighting high beta sectors and those most closely tied to Europe. Remain bearish on CDX in the near term but look for spreads to tighten over the longer term Similar to high grade cash markets, we expect strong credit fundamentals and technical forces to drive spreads tighter. At year-end 2012, we expect CDX.IG to trade at 100bp (versus 136bp today) and CDX.HY at 550bp/$98.0 (versus 772bp/$90.0 today), in line with our 175bp year-end HG bonds forecast. In addition, we believe that investors should take advantage of opportunities where CDS-bond basis is more negative than -75bp in HG and -100bp in HY. Overweight Bs, be neutral on BBs, and underweight CCCs in high yield We think default risk is likely to be negligible over the next few years due to strong liquidity and the health of corporate balance sheets. Thus, we forecast high-yield bond and loan spreads will tighten to T+705bp and L+660bp by year-end 2012. In particular, we find Bs to be the most attractive rating category as they offer the best balance between heightened macro risks, sound credit fundamentals and low default risk.
Stay overweight Senior CLOs (AAA/AA) but neutral CLO Mezzanine to Subordinates (A/BBB/BB) Given the macro risks and lack of liquidity, we stay neutral CLO Mezzanine to Subordinates (A/BBB/BB) but overweight Seniors (AAA/AA). Trading themes include contraction upside as reinvestment periods end, spread tiering in mezzanine bonds, and some opportunities to take advantage of the high current carry in equity Look for muni ratios to fall and for spreads to tighten in early 2012 We look for ratios to fall in early 2012, but remain elevated versus historic norms due to strong reinvestment capital, high volatility, modest inflows, and some flight-to-quality bid. Over the first half, spreads may also tighten as yields remain ultra low and both market returns and the credit environment remain unremarkable. As a relative value trade, we like overweighting A and BBB rated spreads given investors grab for yield and wide A and BBB spreads. Remain marketweight EMBIG but underweight CEMBI EMBIG remains the more defensive asset class with cash flows of $64bn next year versus projected sovereign issuance of $60bn. Within the EMBIG, underweight EMEA EM sovereigns (through underweights in Croatia and Ukraine) as we see the much lower growth environment due to a European recession contributing to a higher level of sovereign stress in this region. We are also underweight idiosyncratic downside risks in the Middle East, where we see continued political instability and worsening fundamentals as underpriced; we are underweight Egypt (through quasi-sovereign Nile Finance) and Lebanon.
105
Appendix: Results of the J.P. Morgan Fixed Income Markets Investor Survey
1) Do you think the Fed will expand its balance sheet via large-scale asset purchases (QE3) in 2012? Yes No 80% 20%
7) How do you expect to change your portfolio currency allocations over the next year? Reduce Maintain Add USD assets 2% 51% 18% EUR assets 16% 27% 9% Other developed markets EM local currency assets EM USD denominated assets 2% 4% 2% 35% 17% 21% 9% 20% 9%
N/A 29% 47% 54% 59% 68%
2) How much new fiscal stimulus do you think Congress will pass by the end of the 2011? None $1-100bn $100-200bn >$200bn 49% 19% 19% 14%
8) Are you looking to reduce, maintain or add exposure to the N/A 13% 42% 47% 41% 46% 46% 48% 39% 50% 56% 60% 76% 43%
following asset classes over the next year? Reduce Maintain Add 3) Do you think any ratings agencies will lower the US Duration risk 27% 38% 23% sovereign rating by one or more notches by the end of 2012? TIPS 7% 27% 24% Yes 52% Agency debt 21% 30% 2% No 48% Agency MBS 11% 17% 31%
4) Do you think Greece will exit the EMU by the end of 2012? Yes No 33% 67%
Non-agency MBS CMBS ABS IG Corporates High Yield EM Equities Commodities Cash
7% 6% 7% 9% 2% 4% 7% 4% 22%
30% 24% 28% 19% 17% 17% 18% 13% 31%
17% 24% 17% 33% 31% 24% 15% 7% 4%
5) Do you think any country other than Greece will exit the EMU by the end of 2013? Yes No 25% 75%
9) In 2012, are you planning to modify your benchmark to 6) Where do you expect 10-year yields to be at the end of 2Q exclude Fed holdings of Treasuries or MBS? 2012? I already have 11% <1.5% 6% Yes 4%
1.5% - 2.0% 2.0% - 2.5% 2.5% -3.0% 3.0% -3.5% >3.5% 22% 48% 20% 5% 0%
No N/A
37% 48%
106
US Fixed Income Strategy US Fixed Income Markets 2012 Outlook November 24, 2011 Terry BeltonAC (1-212) 834-4650 Meera Chandan (1-212) 834-4924 Kimberly L. Harano (1-212) 834-4956 J.P. Morgan Securities LLC
US Treasuries
A worsening of the European sovereign debt crisis will accelerate the flight-to-quality bid into Treasuries early next year; rich valuations, a deteriorating supply/demand imbalance, and poor technicals will limit the upside, however, and keep Treasury rates broadly range bound in 2012 Foreign investors will need to double their Treasury purchases next year to clear the market; while the bid from the EU has been strong, demand from EM has weakened as they have slowed their accumulation of USD reserves Disintermediation of money funds, limited supply, and Fed policy support lower front-end yields With poor liquidity, whipsaw risk in the Treasury market will be elevated next year; we discuss the importance of avoiding consensus trades in 2012 and highlight a measure of investor positioning that helps identify turns in Treasury rates Positive carry yield curve trades in the front end are likely to perform poorly given negative curve convexity; we favor negative carry curve trades anchored in the front end and positive carry trades anchored in intermediates Synthetic Treasuries created by asset swapping cheap foreign bonds should outperform in 2012 Reconstitutions of STRIPS should be strong in 2012, causing short maturity Cs richening versus similar maturity Ps The US fiscal outlook improved in 2011, although debt levels are still on an upward trajectory. We expect ratings agencies to maintain their existing ratings through 2012; another $1.8tn of deficit reduction is still needed for the US to achieve fiscal sustainability
Exhibit 1: Yields tracked US economic data in 1H11 before decoupling in 2H11 as the European debt crisis moved to the core
Average Italy and France CDS spreads (bp) versus 10-year Treasury yields (%) and 3-month rolling average of J.P. Morgan EASI* bp %
50 100 150 200 250 300 350 400 450 Feb 11 10-year Treasury yields Italy / France CDS spreads (inverted) EASI May 11 Aug 11 Nov 11
4.0 3.5 3.0 2.5 2.0 1.5
* EASI is the J.P. Morgan Economic Activity Surprise Index and measures the number of positive surprises minus the number of negative surprises divided by the total number of data releases over the past six weeks. This index has been scaled to fit on the chart as 3.2+J.P. Morgan Economic Surprise Index/10.
Exhibit 2: The Feds biggest impact this year was arguably on 30year rates; the long end flattened 50bp after the Fed began signaling Operation Twist in August
10s/30s Treasury yield curve; %
1.5 1.4 1.3 1.2 1.1 1.0 0.9 Feb 11 May 11 Aug 11 Nov 11
Fall from grace

Treasury rates fell sharply in 2011 with 10-year yields reaching an all-time low of 1.70% in 3Q11. With great irony, the record low occurred in the aftermath of a major fall from grace; for the first time ever, the US lost its AAA sovereign rating as S&P downgraded the US in recognition of the sharp deterioration in US debt metrics over the last few years. Despite a worsening sovereign
credit outlook, Treasury rates moved sharply lower, with 10-year yields falling 135bp since the start of the year to their current level of 1.96%. With front-end yields approaching the zero-rate boundary, the rally also caused the curve to flatten. Year to date, 2-year yields have declined 30bp, 5-year yields have declined 110bp, and 30-year yields have declined 140bp.
107
The downtrend in rates in 2011 masks two very different dynamics driving the Treasury market this year. In 1H11, rates were comparatively range-bound, with yields largely driven by changing perceptions around the US economic outlook. Increased optimism on better growth following the payroll tax cut helped push 10-year yields to the high of the year (3.72%) in 1Q11; these moves were reversed in 2Q11 as economic data disappointed to the downside. During both quarters, volatility was comparatively low, with rates mostly trading within 30bp of their 1H11 average of 3.30%. The dynamic in 2H11 was quite different, as the European debt crisis spread from the smaller peripherals to the much larger bond markets of Italy and core Europe. In response, volatility spiked, Treasury rates plunged, and correlations with European sovereign spreads moved to extreme levels (Exhibit 1). Despite some analysts arguing that the US was in recession, the 2H rally actually occurred as economic data surprised to the upside, highlighting that the rally primarily reflected a strong flight-to-quality bid into Treasuries, as well as unwinds of bad positions. The Treasury market was also supported by Fed policy this year, with stimulus coming from both $640bn of asset purchases under QE2 and another $175bn (10-year equivalents) of duration buying in 2H11 as part of Operation Twist. The overall impact on yields from all that buying was small, however; most estimates including our own suggest that QE2 lowered intermediate rates by 15-30bp (see US Fixed Income Markets Weekly, July 15, 2011). Surprisingly, Fed policy this year likely had its biggest impact on the long end of the curve; after steepening nearly 50bp during the first seven months of the year, the 10s/30s yield curve flattened all that and more after the Fed began signaling Operation Twist in August (Exhibit 2). Fed policy was also modestly effective in helping reduce term premium in the front end of the curve. Two-year rates fell sharply after the August FOMC meeting as the Fed adopted a new communication strategy and clarified its intention to keep the funds rate near zero until mid2013. Since then, 1-year forward 3-month OIS has averaged 11bp compared to a 24bp average in the 1-month period leading up to the August FOMC meeting (Exhibit 3). To be sure, low front-end yields are also being supported by a collateral shortage in the Treasury bill market, and by the European-led flight to quality into Treasuries.
Exhibit 3: Term premium in the front end declined sharply this year in part due to the Feds new guidance on policy rates
2-year Treasury yields versus 1Yx3M OIS rate; %
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Feb 11 May 11
2-year Treasury yields 1yx3m OIS rate
Aug 11
Nov 11
Exhibit 4: The 2011 rally has pushed US 10-year real rates 140 bp below those in Japan
10-year government bond yield minus 10-year inflation swap rate; %
3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 2009 2010 2011
US Japan
Finally, we note that the large decline in Treasury rates in 2011 occurred in an environment where long-term inflation expectations remained relatively stable. For the first time since 1997, 10-year Treasury yields moved significantly below long-term inflation expectations, thus implying negative 10-year real returns. With the exception of the UK, these real yields compare unfavorably to virtually every other developed bond market in the world, including Japan, where real rates are currently 140bp above those in the US (Exhibit 4). Such low real rates highlight both the strength of demand for Treasuries during this years debt crisis as well as the significant challenges Treasury investors face to earn reasonable returns going forward.
108
The outlook for Treasury yields

Looking into 2012, our outlook is for Treasury volatility to remain high but with rates ultimately remaining rangebound as further deterioration in Europe competes with rich valuations. Within this range, we look for rates to move lower early in the year as the crisis in Europe deteriorates further and investors add to the risk-off trade. At the same time, we see the flight-to-quality bid eventually fading, as rich valuations, crowded investor positions, and still-heavy supply allow the Treasury market to decouple from the ongoing stress in Europe. In the discussion below, we assess the major factors likely to drive Treasury yields in 2012, present our interest rate forecast, and highlight the many risks around that forecast. The case for lower Treasury yields in 2012 really rests on only two factors, but they are important ones that, at least early next year, are likely to cause Treasury yields to trade at even richer levels than those reached in 2011. The first is Europe. Our baseline view is that the sovereign debt crisis in Europe is now set to take another turn for the worse and is likely to fuel a renewed flight-to-quality bid into Treasuries early next year. While the details are more fully discussed in our Global Fixed Income Markets Outlook publication (see Global Fixed Income Markets 2012 Outlook, November 24, 2011), we would highlight three of the most worrisome aspects of that outlook with implications for the Treasury market. We look for intra-EMU spreads to widen significantly early next year. The widening will likely be driven by increased selling from unlevered investors who thought they owned risk-free government bonds but are now overweight high yielding risky assets. With yields elevated, Italys access to private market funding may become less secure, creating some risk that they are forced to seek external funding support. Italy is the worlds third largest bond market, however, and with 250bn of funding needs in 2012 alone (including 200bn of redemptions), they will stretch the limits of the EUs and IMFs ability to meet their funding needs. Policymakers will likely continue to move too slowly to get ahead of the deepening crisis. The ECB will likely remain a reluctant lender of last resort and, with heavy investor selling likely to continue, is unlikely to buy enough to stabilize spreads near current levels. While we do not
1-month rolling average of 10-year Treasury yields (%) versus the 1-month rolling correlation between weekly changes in 10-year Treasury yields and semiperipheral* Europe sovereign CDS spreads % correlation; inverted
Exhibit 5: Correlations between 10-year Treasuries and European sovereign spreads remain elevated and were especially high when 10-year yields hit record lows
3.0 2.9 2.8 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 Sep 11 Oct 11
Treasury yields Correlation; inverted
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2
Nov 11
* Average 5-year CDS spreads of France, Spain and Italy.
expect a EU breakup, statements from officials in Germany, the Netherlands, and other creditor countries about EU exit strategies are likely to increase and provide ongoing support for investors to short semi-peripheral bond markets. The crisis is likely to broaden further beyond Italy, especially if the Euro area recession proves severe. In that case, risks would rise of a France sovereign downgrade; this would lower the EFSF funding capacity and further increase the resources required from the creditor countries.
We note that these views represent our base case outlook; the headline risk around Europe is extraordinarily high and a permanent solution to the crisis could certainly be found just when the stress reaches its deepest point. How low could intermediate Treasury yields go if Europe deteriorates further? Correlations between Treasury yields and European sovereign spreads remain high and should stay that way into the early part of next year. Using our fair value model (see box), we estimate that for every additional widening of 100bp in 10-year Italy spreads to Germany, 10-year Treasury yields should fall by 16bp, all else equal. If 10-year Italy yields reach 9% next year, we would expect 10-year Treasury yields to move below the record 1.70% reached last year. To be sure, US rates will eventually decouple from Europe as rich valuations bring
109
Exhibit 6: Buy the rumor, sell the fact; 10-year Treasuries have rallied an average of 50bp leading into the last four Fed QE announcements
10-year Treasury yields (%) and J.P. Morgan index of levered investor bond positions* % positions index; inverted axis
3.0 2.9 2.8 2.7 2.6 2.5 2.4 2.3 -30 -20 -10 0 10 20 Business days around quantitative easing 30 10-year Treasury yields Positions index; inverted
-0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4
* Levered investor positions index is the weighted average of aggregate net longs held by non-commercial investors as provided by CFTC and the partial beta of hedge fund returns versus 10-year Treasury yields, converted to a z-score. QE announcement dates are 11/25/08, 3/18/09, 9/21/10, and 9/21/11.
Flight to quality and fair value of 10-year Treasuries In February 2011, we introduced a new short-term fair value model for 10-year Treasury yields that combines both economic and technical factors likely to drive yields in the near term. Since then, we have modified the model somewhat by introducing better measures of investor positioning, and added European sovereign debt spreads as a proxy for flight-to-quality flows from Europe. The model now explains 10-year Treasury yields as a function of five variables including 5-year forward inflation expectations derived from the inflation swap market, the 1-year ahead J.P. Morgan forecast for real GDP growth, near-term expectations on policy rates measured by the 3-month forward on 3-month OIS, the average 5-year CDS spreads of France, Italy, and Spain, and a measure of levered investor positions (see box below on measuring investor positions). The model is estimated using weekly averages for data since June 2008. All five variables are highly significant and explain about 80% of the variability in 10-year rates during this period. The coefficient on the flight-to-quality variable indicates that every 100bp increase in semi-peripheral spreads lowers 10-year yields by 33bp on average. Alternatively, because average semiperipheral CDS spreads have been moving only half as much as spreads on Italian government bonds recently, we estimate 10year yields should fall by roughly 16bp for every 100bp increase in spreads on Italian government bonds. This estimate holds other variables constant; we expect the actual change in 10-year rates to be larger early next year as increased sovereign stress will also be accompanied by an increase in levered investor duration longs. 10-year Treasury yield model parameters:* Variable Current level Beta T-stat Intercept 1.983 14.2 Inflation expectations; % 2.72 0.328 7.3 GDP growth forecast; % 1.75 0.184 24.9 3Mx3M OIS; % 0.10 0.476 24.9 Sovereign CDS; % 4.01 -0.326 -26.5 Levered investor positions -0.64 -0.296 -19.1
*Estimated from 6/1/08-11/10/11. N=864, SER=0.229, R2=0.824
sellers. But for now, correlations (and betas) remain quite elevated and were especially high when rates made their record lows back in September (Exhibit 5). After Europe, the second most important factor supporting lower Treasury yields in 2012 is likely to be Fed sponsorship. With growth likely to remain sluggish in 1H12 as fiscal drags come into play and unemployment remains high, we look for the Fed to announce another round of asset purchases in MBS. While the eventual impact on Treasury yields may be modest, the initial signal by the Fed that more QE is coming is likely to lead to an outsized rally as investors rebalance their portfolios. In each of the previous four instances where the Fed announced large scale asset purchases, Fed signaling in advance of the formal announcement caused 10-year Treasury yields to fall sharply (by an average of 50bp) as levered investors significantly increased duration longs (Exhibit 6). Beyond Europe and the Fed, other factors supportive of lower yields in 2012 are small by comparison. They include a worsening of the front-end collateral shortage helping push GC rates lower (see GC repo discussion below), a modest weakening in growth and inflation expectations early next year, and further changes in the
10-year Treasury yields versus model fair value; %

4.0 3.5 3.0 2.5 2.0 1.5 2009 2010 2011 Actual Model
110
Exhibit 7: 10-year Treasury rates are rich by 80bp
10-year fair value* excluding the impact of the flight to quality bid into Treasuries versus actual 10-year Treasury yields; %
Exhibit 8: The largest foreign holders of Treasuries slowed their accumulation of USD reserves this year with Fed custody holdings of Treasuries flat since June
Treasuries held at custody at the Fed excluding repo versus foreign exchange reserves of the largest Treasury holders*; $bn $bn
5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 2009 2010 2011
Model Actual
2680 2660 2640 2620 2600
Treasuries held in custody at the Fed
5700 5600 5500 5400 5300 FX reserves 5200 5100
* 10-year Treasury model yield = 0.34 + 0.66 * 5yx5y inflation swap rates (%) + 0.85 * 3mx3m OIS rate (%) + 0.2 * 1-year ahead J.P. Morgan real GDP growth forecast (%) 0.33 * J.P. Morgan investor positions index in bonds; Estimated over 6/1/08-11/11/11 period.
2580 2560 Feb 11
Apr 11
Jun 11
Aug 11
Oct 11
Feds policy rate guidance that should help lower term premium in the yield curve next year. Against these bullish factors for the Treasury market, there are also some important offsets that we believe will ultimately push Treasury yields higher in 2012. The first is valuations. Ignoring the impact of the (presumably temporary) flight-to-quality bid into Treasuries, our fair value model for 10-year yields suggests Treasuries are 80bp expensive (Exhibit 7). While this richness can persist when Europe remains under stress, the magnitude of the mispricing highlights the significant cheapening that is likely to occur in the Treasury market as soon as the situation in Europe stabilizes. Second, we view the supply/demand outlook in the Treasury market as unfavorable next year, with duration supply rising and Fed buying likely to be significantly weaker than in 2011. As discussed in more detail below, we expect overall duration supply (net of Fed purchases) across US fixed income markets to increase 11% next year, assuming a moderately sized QE3 program in MBS. Treasury supply net of Fed purchases is projected to more than double to $1tn, requiring a significant increase in buying from foreign investors. This increased supply is occurring at the same time that the largest foreign holders of Treasuries have slowed their accumulation of USD reserves with Fed custody holdings of Treasuries essentially flat since June (Exhibit 8). Third, with Treasury rates very rich and most investors already extremely pessimistic about Europe, we would
* Includes FX reserves of South Korea, Brazil, Japan, China, Russia and Taiwan. Dollar is held constant at March 2009 levels.
Exhibit 9: Correlations between US rates and spreads in Greece/Portugal/Spain had fallen to zero this summer despite continued spread widening
Average Greece, Portugal and Spain CDS spreads versus the 3-month rolling correlation between 1-month changes in 10-year rates and average Greece, Portugal and Spain CDS spreads bp correlation; inverted axis
1200 1100 1000 900 800 700 600 500 400 Feb 11 Apr 11 Average CDS spreads Correlation; inverted Jun 11
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 Aug 11
expect Treasury yields to begin decoupling from European headlines at some point next year. In part, this reflects investor positioning: correlations tend to rise early on in the crisis as the news forces investors to rebalance but fall once those rebalancing trades have occurred. Indeed, prior to the broadening of the crisis in August to include Italy and France, US rates had already begun decoupling from peripheral spreads; correlations between 10-year Treasury yields and the average spreads of Greece, Portugal and Spain had fallen to zero this summer, even though
111
peripheral spreads continued to move significantly wider (Exhibit 9). To be sure, rebalancing flows out of Italy are far from done, suggesting the Treasury rally is not over. But at some point next year, we expect the short risk trade to become crowded again limiting the upside in Treasuries even if Italy continues to widen. The offsetting factors of further deterioration in Europe and already rich valuations in the Treasury market will make it difficult for Treasuries to trend for an extended period in either direction next year. Along with our outlook for near-trend growth and for policy rates to stay anchored near zero at least until 2014, we expect rangebound Treasury rates in 2012. Within that range, we look for rates to move lower early next year as Europe deteriorates and investors reset short risk positions. Our 1Q12 target for 10-year rates is 1.70% (Exhibit 10), which we view as consistent with another 100-200bp increase in Italian 10-year yields combined with an increase in duration buying from levered investors early next year as they reset the risk-off trade. Our year-end target for 10-year yields is 2.5%. This assumes only a partial retracement back to fair value as we expect Europe to remain unsettled, resulting in some lingering flight to quality-related demand. At the front end of the curve, our forecast tracks the outlook for 10-year rates (adjusted for the lower volatility of short rates) but incorporates some additional richening driven by an anticipated decline in GC repo and OIS (see The front end and the outlook for GC below). We expect the Fed to further strengthen its forward guidance on policy rates early in 2012 helping further flatten the OIS curve. In addition, front-end Treasuries are cheap to OIS and we expect 3-5bp of richening in 2-year yields even if OIS remains unchanged. We are biased to a modestly steeper 10s/30s yield curve in 2012 targeting 110bp by mid-year. Following the announcement of Operation Twist in September, the long end of the curve had significantly flattened relative to fair value before retracing much of the mispricing over the past month. Our 10s/30s yield curve model, which is based on the level of front-end rates, inflation expectations, variable annuity (VA) hedging needs, and long-end supply (see U.S Fixed Income Markets Weekly, September 23, 2011) currently shows the curve as 7 bp too flat (Exhibit 11). While we could see some modest flattening in 1Q12 as stresses in Europe increase long-end VA-related duration buying, increased prospects that the Fed will expand its balance sheet as we move through the year provides an offset. On balance, with Treasury rates largely range bound into mid-year, we look for a modestly
112
Exhibit 10: J.P. Morgan interest rate forecast

%
Current 21 Nov 11 2-yr Treasury 5-yr Treasury 10-yr Treasury 30-yr Treasury 0.28 0.91 1.96 2.94 4Q11 31 Dec 11 0.22 0.90 2.00 3.00 1Q12 31 Mar 12 0.17 0.75 1.70 2.70 2Q12 30 Jun 12 0.30 1.25 2.50 3.60 4Q12 31 Dec 12 0.30 1.25 2.50 3.60
Exhibit 11: After a significant overshoot following Operation Twist, the 10s/30s curve has moved back to fair value
Actual versus model* for the 10s/30s Treasury yield curve; %
1.6 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 May 10 Nov 10 May 11 Nov 11 Actual Model
* 10s/30s curve modeled as 1.147 - 0.003292 * index of variable annuity hedging ($bn 20-year equivalents) + 0.1169 * 5yx5y inflation swap rates (%) 0.2446 * 3year yields (%) + 0.001028 * Amount of Treasuries outstanding in the 10- to 30year sector net of Fed purchases via Operation Twist ($bn)
steeper 10s/30s curve and target 30-year yields to reach 3.60 by mid-year. Finally, we would note that this is a year where volatility is likely to be extraordinarily high even if rates ultimately prove range-bound. With poor liquidity and balance sheets of many financial institutions becoming more constrained next year, the market is likely to continue to experience frequent whipsaws in Treasury rates driven as much by technicals and investor rebalancing flows as by fundamentals. In the next section, we discuss why technicals are likely to be so important next year and make the case that contrarian duration trading strategies are likely to be especially profitable in 2012. We also discuss improved metrics we have developed this year for tracking investor technicals that have proven to be a good leading indicator of turns in the Treasury market.
Avoiding crowds in 2012

While our baseline forecast is for yields to move lower early in 2012 as contagion spreads in Europe, markets are very unlikely to move in a straight line; we expect volatility to remain extremely high with frequent turns in US interest rates during the course of the year driven by poor liquidity and technicals around investor positioning. While technicals have always been an important factor driving yields, a combination of cyclical and structural factors have amplified their impact since the end of the 2008 financial crisis, causing intermediate Treasury rates to be extraordinarily volatile despite stable policy rates (Exhibit 12). These factors include both declining market depth and a growing disparity between the size of dealer and end-user balance sheets. The importance of technicals on Treasury rates in the last couple of years is highlighted in Exhibits 13 and 14 which compare 10-year Treasury yields to a measure of investor duration positioning we have developed (see grey box). The exhibits highlight four major turns (whipsaws) in the Treasury market since 2Q10 that were triggered or at least amplified by investor unwinds of crowded consensus trades. This includes: 1) the Treasury rally in 2Q10 following a period where improving growth and Fed exit strategy discussions resulted in significant duration shorts by levered investors; 2) the sell-off in Treasuries following the start of QE2 in October 2010; 3) the 3Q11 rally in Treasuries following the end of QE2; and, 4) the 4Q11 sell-off in Treasuries following the EU driven flightto-quality bid in Treasuries. While the initial catalyst for the unwinds differed, the whipsaw in Treasury rates occurred in each case just after our investor position measure reached a fairly extreme level with unwinds amplifying the market move. Over the four episodes, the change in 10-year Treasury yields (with Treasury yields moving in the opposite direction of the consensus view in all four cases) averaged 70bp. The largest move (97bp sell-off) followed the start of QE2 when our position measure indicated levered investors held their largest long duration position since 2007.
Exhibit 12: Going nowhere fast: a structural decline in liquidity has caused Treasury volatility to spike despite stable policy rates
Rolling 1-year standard deviation of daily changes in 10-year Treasury yields divided by the rolling 1-year standard deviation of daily changes in the 3Mx3M OIS rate
12 10 8 6 4 2 0 1995 1999 2003 2007 2011
Exhibit 13: The importance of avoiding crowds

Consensus View Peak positons by Duration Rationale Short Growth/Fed exit 5/18/10 Long QE2 coming 10/21/10 Short QE2 ending 4/12/11 Long EU crisis 9/23/11 Average move (abs value) Positions Time 10y yield reversed by (months) change; bp 10/21/10 4/12/11 6/23/11 10/18/11 5.0 5.6 2.3 0.8 3.4 -85 97 -59 34 69
Exhibit 14: Crowded trades have led to four major rate whipsaws since 2Q10
10-year Treasury yields (%) versus the J.P. Morgan levered investor position index in bonds*
1.5 2.0 2.5 3.0 3.5 4.0 May 10 Nov 10
10-year Treasury yields; inverted 1.5 Positions index 1.0 0.5 0.0 -0.5 -1.0 -1.5 May 11 Nov 11
* Positions index is the weighted average of aggregate net longs held by noncommercial investors as provided by CFTC and the partial beta of hedge fund returns versus 10-year Treasury yields, converted to a z-score.
113
Measuring investor positions One of the more active areas of research for us in 2011 has been developing alternative measures of investor positioning that better explain Treasury yield movements. While survey data (e.g., the J.P. Morgan Treasury Client Survey) are often used as a barometer of investor positioning, we have found it to be less useful for explaining and predicting Treasury yields than actual position measures available from futures markets. In addition, position measures calculated by reverse engineering return data (i.e., return betas) appear more reliable to us than survey measures. The table below provides a comparison of alternative measures of investor positioning based on how well they explain yield movements when added to our short-term fair value model for 10-year Treasury yields. The investor position measures we compare include (duration weighted) CFTC data on net speculative positions in interest rate futures, survey data from our weekly duration survey, and return betas calculated using daily returns published by the largest core bond funds as well as hedge funds. The fair value model (see grey box above), which is estimated using weekly data since June 2008, explains 10-year Treasury yields as a function of inflation expectations, the J.P. Morgan 1-year ahead growth forecast, the level of 3-month forward OIS, a flight-to-quality measure proxied by European semi-peripheral spreads, as well as a measure of investor positions. While most position indicators are statistically significant and improve the overall fit of the model, we find levered investor position measures explain Treasury yield movements much better than unlevered investor positions. The best-fitting models are the ones that use CFTC data on speculative positions followed by a position measure based on macro hedge fund return betas. These models have the lowest regression standard error, highest t-statistic on the position variable, and highest Rsquared (see Table). By contrast, the J.P. Morgan duration survey measure and the return betas for core bond fund managers have the lowest explanatory power with the bond fund return beta statistically insignificant. We also find that the 10-year model can be improved by combining some of these position measures. In particular, we have created a levered investor index that converts the CFTC and hedge fund return betas to a z-score, and takes a weighted average based on regression weights. This index has the best fit in our 10-year yield model when compared to other position measures and has been stable through time; the Rsquared increases to 0.82, and the standard error of the regression falls to 0.23. This index is currently at -0.5 suggesting levered investors are now modestly short duration. Beyond helping understand the factors driving yield changes, we also find position indicators to be useful contrarian indicators that help us forecast future yield changes. Reflecting the fact that levered investor positions tend to be mean-reverting, we find our position index to be a useful leading indicator of yield changes. This is highlighted in the last exhibit below which compares monthly changes in 10-year yields to the ex ante level of our J.P. Morgan positions index. 10-year Treasury yield model summary statistics*
Model includes: Current value Coefficient T-stat RSQ SER CFTC spec longs -1.1 -0.175 -11.8 78.3 0.25 Hedge fund beta -0.4 -0.090 -11.3 78.0 0.26 JPM duration survey 0.6 -0.059 -6.2 75.9 0.27 Mutual fund beta 0.3 -0.006 -0.5 74.8 0.27 CFTC/HF Index -0.8 -0.295 -21.5 82.3 0.23 *All models estimated from 6/1/08-11/8/11. In addition to a measure of investor positions, the model includes 5Yx5Y forward inflation expectations, the JPM 1-year ahead growth forecast, the level of 3-month forward OIS, and an average of France/Italy/Spain 5-year CDS.
JPM index of levered investor positions:*

1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 May 10 Nov 10 May 11 Nov 11
1M change in 10Y UST yields vs. one-month ago JPM index of levered positions; 5/10-11/11
1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 Nov 10 May 11 Nov 11 1m ago Positions Index 0.8 1m change in yields; % 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0
* Defined as 0.66 * CFTC net spec longs z score +.33 * hedge fund return beta z-score. Hedge fund betas are calculated using daily returns on the IQ Global Macro Beta Index.
114
Exhibit 15: Market depth has declined as dealers continue to lower VAR and shrink their balance sheets
Reported average quarterly VAR for the 9 largest investment banks* versus market depth**; $mn $mn $mn
Exhibit 16: while end-user demand for liquidity has increased

AUM for the largest asset managers*; $tn
18.0 17.5 17.0 16.5 16.0 15.5 15.0 14.5 1Q10 2Q10 3Q10 4Q10 1Q11 2Q11 3Q11 *Includes BlackRock, Allianz Group, State Street Global, Deutsche Bank, Vanguard Group, Fidelity Investments, J.P. Morgan Chase, BNP Paribas, AXA Group and BNY Mellon. Source: Bloomberg, Company websites, P&I/Towers Watson Global 2010 ranking
125 120 115 110 105 100 95 90 May 10 Nov 10
VAR Market depth
240 220 200 180 160 140 120 100 80
May 11
Nov 11
* Average VAR reported by JPM, GS, MS, BAC, C, UBS, CS, Soc Gen and DB. 3Q11 VAR is an estimate, projected from the results reported to date (which excludes C). ** Market depth is calculated as the 3-month moving average of average size of the top three bids and offers, in $mn, for the on-the-run 10-year Treasury note, averaged between 8:30 a.m. and 10:30 a.m. daily.
Exhibit 17: A simple contrarian trading strategy based on our index of investor positioning was highly profitable in 2011
While markets often zigzag, the magnitude of the recent whipsaws appear unusually large, given that the economic outlook has been comparatively stable (slow but positive growth) and the Fed is expected to keep rates near zero for a prolonged period of time. In our view, the volatility reflects a secular decline in market liquidity as a declining pool of dealers face increasingly larger capital requirements and stricter leverage ratios. Some evidence of this can be seen in the decline of dealer reported VAR levels since 2010 and the corresponding decline in market depth (Exhibit 15). We would note that this reduced ability of dealers to provide liquidity is occurring at the same time that the balance sheets of the largest asset managers are growing, creating an imbalance in the supply and demand for liquidity (Exhibit 16). Because poor market depth is unlikely to improve next year as US and especially European banks de-lever further, technicals are likely to again be a dominant driver of Treasury yields with investor rebalancing flows having an outsized impact on yields. Thus, despite our 1Q12 outlook for lower yields driven by renewed stress in Europe, our bias in 2012 will be to trade duration tactically with an emphasis on fading crowded trades and positioning for mean reversion. This strategy proved especially profitable in 2011; a simple rule of going long duration each day our levered investor positions index was below -0.75 and short duration each
Number of trades, success rate and average P/L from back-testing a trading rule* based on the J.P. Morgan positions index
Trade
Number of trades
Success rate (%) Average P/L (bp)
Long 10s 39 95 22.9 Short 10s 32 100 28.3 Total 71 97 25.3 * The rule involved going long duration each day our levered investor positions index was below -0.75 and short duration each day our index was above 0.75. Positions were held for a one month period.
day our index was above 0.75 had a success rate of over 95% (based on 71 trades) and average P/L for 1-month holding periods of 25bp (Exhibit 17). With the index currently at -0.5, we are currently biased to holding long duration trades expecting investors to build long duration positions as the risk-off trade gains momentum early in 2012.
The front end and the outlook for GC

The past year has been characterized by increased demand for front-end Treasuries from money market funds. As the European crisis escalated and risk aversion increased, the assets under management (AUM) in prime money market funds fell dramatically and investors instead favored the relative safety of government money market funds. Since the end of May, the AUM in prime money market funds have fallen $200bn, while the AUM in government money market funds have increased by $130bn. As a result, Treasury holdings of money market funds (including repo)
115
Treasury bills outstanding ($bn) versus Treasuries held by money market funds ($bn)
Exhibit 18: The perfect storm: high money market demand in a period of falling supply resulted in lower bill yields
Exhibit 19: We expect net issuance in the front end to be negative in 2012
Projected 2012 net issuance in short-term fixed income markets by sector; $bn
1800 1750 1700 1650 1600 1550 1500 1450
T-bills outstanding Treasuries held by money market funds
520 510 500 490 480 470 460 450 440 430
50 0 -50 -100 -150 -200 Non-fin CP T-bills ABCP Agency discos Financial CP Yankee CDs
Feb 11
May 11
Aug 11
Nov 11
Source: Treasury, J.P. Morgan, iMoneyNet
have increased dramatically (Exhibit 18). This increased demand has come in a period of falling T-bill supply. The combination of these two factors proved to be very supportive of valuations: T-bills across the curve richened to historical levels, with yields at or near-zero levels. Supportive technicals notwithstanding, GC repo rates have stayed stubbornly high, with 1-month term GC currently at around 11bp. Looking ahead, we expect GC repo rates to head lower targeting 1-week GC to average 5bp in 1H12 for the following reasons. First, supply technicals in the front-end are very supportive. Our short-term fixed income strategists expect issuance in most front-end markets to be negative in 2012, with Non-Financial CP being a notable exception (Exhibit 19). Moreover, net T-bill issuance for FY2012 will be close to zero or negative, depending on the magnitude of fiscal stimulus that is passed. We expect the budget deficit to total $1.1tn in FY2012, assuming that a $100bn stimulus is passed. With coupon net issuance expected to be $1.09tn, the net issuance of T-bills will total a mere $10bn over that period. However, if no stimulus is approved, then T-bill issuance will total -$90bn. Second, demand for T-bills in particular is likely to stay elevated despite shrinking supply. Investors will likely stay risk-averse since headline risk from Europe will stay elevated in 1H12, and demand for government paper will persist as assets continue to migrate out of prime and into government money market funds. Moreover, unattractive net yields in prime funds (a mere 3bp) relative to government funds (1bp) will also be supportive of this trend. Finally, possible regulatory
116
Exhibit 20: We look for 1-week GC repo rates to average near 5bp in 1H given our outlook for supply and money market fund demand
1-week Treasury GC repo rate (bp) modelled as a function of outstanding T-bills net of money market fund holdings of Treasuries ($bn)
25 20 15 10 5 0 950
Y = 0.0421 X1 - 36.6723 R = 65.92% standard error = 3.9300 period = Nov 19,09 - Nov 19,11
1000 1050
1100 1150 1200 1250 1300 1350 1400 Effective T-bills outstanding; $bn
changes in money market fund capital requirements and the constant NAV model in late 2012 could further accelerate this trend (see Short-term Fixed Income). The combination of these factors is likely to put downward pressure on GC repo rates as we head through the year. Historically, the outstanding balance of T-bills net of money market holdings of Treasuries has been a significant driver of repo rates, with every $100bn of additional supply biasing 1-week GC repo rates 4bp higher (Exhibit 20). Given our expectation of close-tozero net bill issuance in 1H12, combined with our expectation that money market funds will purchase $20bn of Treasuries per quarter, we expect that 1-week GC repo
rates will eventually average around 5bp in 1H12, or 6bp lower than current levels. This environment will be broadly supportive of lower 2-year yields as well, and will push them closer to 17bp in 1Q. In addition, to lower financing rates, our positive outlook on 2-year Treasuries is driven by the following factors. First, the expansion of the Feds balance sheet under QE3 will increase the supply of reserves and help lower the effective funds rate. Historically, every $100bn of additional reserves in the system has lowered 3-month OIS rates by 1.5bp, as shown in Exhibit 21. Thus, if the Fed announces quantitative easing of $500bn, it could significantly lower front-end OIS rates, bringing them close to zero. Second, changes in the Feds communication policy will be supportive as well, since it will further reduce term premium from the front end of the curve. We expect the Fed to modify its communications on policy rates during 1Q12 by linking future interest rate hikes to specific economic conditions along the lines suggested by Chicago Fed President Evans. This should lengthen the markets expectation of the Fed-on-hold horizon beyond mid-2013. Assuming that this outcome causes forward OIS rates to narrow back to levels reached when the Fed first introduced the mid-2013 language on August 9 suggests 1Yx1Y OIS rates could fall by almost 10bp (Exhibit 22). We estimate that this would lower 2-year Treasury yields by another 8bp. Finally, we expect 2-year Treasuries to richen relative to OIS as we approach mid-year when the Fed is scheduled to complete Operation Twist. Fed selling of the front end has helped Treasuries cheapen relative to OIS with the spread between 2-year Treasury yields and 2-year OIS currently at 13bp, or 6bp wider than its 1-year average. In sum, given our outlook for GC, term premium, and an expansion of the Feds balance sheet, we expect 2-year Treasuries to richen on an outright basis and relative to OIS. We look for 2-year rates to trade back to 17bp, or around 10bp richer than current levels. Finally, we also note that changes to the Feds communication policy are also likely to remove term premium even further out the curve, such as in the 3-year sector. As shown in Exhibit 22, at close to 35bp, the 1Yx1Y/2Yx1Y OIS curve is quite steep and has the
Exhibit 21: The expansion of the Feds balance sheet under QE3 will increase the supply of reserves and help lower the effective funds rate
1-month rolling average of 3-month OIS rate (bp) regressed against US bank reserves ($bn)
24 22 20 18 16 14 12 10 8 6 1000 1100 1200 1300 1400 Reserves; $bn 1500 1600 1700 Y = -0.0145 X1 + 33.8150 R = 78.67% standard error = 1.9268 period = Nov 21,09 - Nov 21,11
Exhibit 22: OIS forwards in the front end of the curve have substantial room to fall if the Fed adopts the Evans plan
Minimum 1-year spot, 1Yx1Y and 2Yx1Y OIS rates since the FOMC meeting on August 9 versus current levels; %
0.6 0.5 0.4 0.3 0.2 0.1 0.0
Current Min since 8/9
0.55 0.46
0.19 0.12 0.06 0.08
1y spot
1yx1y
2yx1y
potential to flatten significantly further if the market starts to price in a longer period of Fed-on-hold.
117
Exhibit 23: Gross duration supply fell by 10% in 2011, but is poised to increase by 8% in 2012
Annual duration supply via MBS, Municipal, investment grade corporates, Treasuries and Agency debt markets, gross and net of Fed purchases; $bn 10year equivalents
Exhibit 24: Look for net issuance of coupon Treasuries to fall further in 2012
Net issuance in Treasuries by sector and by calendar year; $bn
3500
Gross duration supply Supply ex-Fed
3000
Sector Bills Nominal coupons TIPS Total Total ex-SFP bills Nominal coupons+TIPS
CY 2010 -19 1563 48 1592 1397 1611
CY 2011* -274 1245 120 1090 1290 1364
CY 2012* -8 936 100 1028 1028 1036
2500
* 2011 and 2012 are J.P. Morgan projections.
2000
Exhibit 25: The share of T-bills in the Treasury market fell again this year, further lengthening the weighted average maturity of the Treasury market
Weighted average maturity of outstanding Treasuries (months) versus the share of T-bills in the Treasury market (%); dashed lines are projections* Months % 35% 75 Share of T-bills 70 30%
1500 2007 2008 2009 2010 2011 2012

* 2011 and 2012 are J.P. Morgan forecasts. We assume that the Fed purchases a total of $500bn MBS 2012 via QE3.
Supply Outlook
Supply technicals improved during 2011 as gross duration supply across fixed-income markets fell 10% and duration supply net of Fed purchases fell 24% (Exhibit 23). The decline in gross issuance was fairly broad-based and came primarily from spread products led by Municipals (down 29%), Agency debt (down 28%), MBS (down 14%), and high-grade credit (down 9%). Issuance within the Treasury market also fell in 2011 driven by a combination of lower deficits and the suspension of SFP bills issuance. For calendar year (CY) 2011, we project net Treasury issuance to equal $1.09tn versus $1.59tn in CY2010 (Exhibit 24). Notably, even though nominal coupon sizes were left unchanged during the year, net issuance of nominal coupon Treasuries fell by $318bn as redemptions were higher (Exhibit 24). 2011 also became the third consecutive year of negative bill issuance, causing the share of T-bills in the Treasury market to fall further to 15.2% currently from 19.6% at the beginning of the year. As a result, the weighted average maturity of outstanding debt increased to 63 months, its longest since August 2002, and up from 55 months at the end of 2009 (Exhibit 25). Looking ahead, we expect overall duration supply in fixed-income markets to increase 8% in 2012 approaching the record levels set in 2009-2010 (Exhibit 23). This increase is primarily driven by higher supply in the MBS and Municipal markets with mortgage refinancing
118
65 60 55 50 45 Jan 00 Weighted average maturity Aug 02 Feb 05 Aug 07 Feb 10
25% 20% 15% 10% Sep 12
* Assumes unchanged coupon sizes and FY budget deficit of $1.1tn
Exhibit 26: We look for the budget deficit to fall 15% in FY2012
Budget deficit by fiscal year; $bn
1500
1416
1294 1298 1100
1000 377 413 319 158 0 -237 2000 -127 2003 2006 2009 2012* 455 248 162
500
-500
* 2012 is J.P. Morgan forecast.
Exhibit 27: Nominal coupon sizes should be stable in 2012 while TIPS issuance increases modestly
J.P. Morgan gross issuance forecast; $bn
2s Jan 12 Feb 12 Mar 12 Apr 12 May 12 Jun 12 Jul 12 Aug 12 Sep 12 Oct 12 Nov 12 Dec 12 Chg from 2011 35 35 35 35 35 35 35 35 35 35 35 35 0 3s 32 32 32 32 32 32 32 32 32 32 32 32 0 5s 35 35 35 35 35 35 35 35 35 35 35 35 0 5y 7s 10s 30s TIPS 29 29 29 29 29 29 29 29 29 29 29 29 0 21 24 21 21 24 21 21 24 21 21 24 21 0 13 16 13 13 16 13 13 16 13 13 16 13 0 14 44 6 76 6 26 3 12 535 2150 15 14 12 8 541 14 16 12 8 537 12 10y 30y TIPS TIPS 14 10 537 Subtotal
Exhibit 28: While the magnitude of buying by foreign investors was lower in 2011 than that in recent years, they took down a larger portion of effective supply
Foreign purchases of Treasuries ($bn) versus foreign purchases as a percentage of Treasury net issuance excluding Fed purchases (%) % $bn
140% 120% 100% 80% 60%
%ge of net issuance bought by foreigners
Amount of foreign Treasury purchases
800
600
400
200 40% 0 2001 2003 2005 2007 2009 2011* Source: Federal Reserve Z.1, TIC Note: 2011 is an estimate of the total foreign purchases reported by the Federal Reserve in Flow of Funds data. 1H11 uses actual Flow of Funds data, Q3 models the Flow of funds purchases as -3.7 + 0.975 * Monthly net purchases reported by Treasury in the TIC report, while Q4 is estimated by a combination of: (a) Flow of funds modeled as 3.8 + 1.34 * Monthly change in Treasury custody holdings and (b) the YTD monthly average purchases of Treasuries by foreigners. 20%
Total 420 384 420 348 264 168
projected to increase and municipal finances staying weak. Within the Treasury market, the outlook is more favorable, with the deficit expected to be on a declining trajectory. Our forecast assumes a FY2012 budget deficit of $1.1tn, lower than the $1.3tn in FY 2011 (Exhibit 26). Thus, net issuance will be lower, too, totaling $1.04tn (Exhibit 24). The composition of issuance will likely be similar to that in recent years. We look for nominal coupon sizes to stay unchanged in 2012, but look for TIPS issuance to increase modestly (Exhibit 27). Given our deficit projection, this implies that Treasury bill net issuance will be close-tozero in 2012, with the share of bills in the Treasury market falling modestly to 14.6% (Exhibit 25). This projection assumes Congress passes $100 bn of fiscal stimulus; if no stimulus is passed, net issuance of bills is projected to fall to -$90bn, bringing the share of bills to 13.7% of the market by the end of FY2012. In this case, there is some possibility that Treasury will cut coupon sizes later in the year in order to stabilize the size of the bill market.
Exhibit 29: A flight-to-quality bid: EU countries were net buyers of Treasuries in 3Q11 as the crisis in Europe escalated, more than offsetting selling by EM countries
Average monthly purchases of coupon Treasuries by foreigners in August and September according to TIC data versus its prior 6-month average; $bn
Region EU UK France Japan China EM ex China*
Aug-Sep average 6m average 44.4 33.7 10.9 21.8 -9.8 -9.4 10.9 15.1 -1.6 4.2 4.4 2.0
* Includes Brazil, Russia, African countries, Taiwan, South Korea and Mexico. Countries in grey are also included in the EU total. Source: TIC
Demand Outlook
At the same time that duration supply is set to increase across fixed income markets, demand technicals are likely to be more challenging for the Treasury market in 2012. The primary reason is a decline in Fed sponsorship. While the Fed purchased 60% of net Treasury issuance in 2011, we expect limited net buying of Treasuries by the Fed next year with any increase in its balance sheet likely to be targeted in MBS. As a result, private investors and
foreign central banks will need to more than double (from $454bn in 2011 to $1,000bn in 2012) the amount of Treasuries they buy in order for the Treasury market to clear in 2012. After the Fed, the single largest buyer of Treasuries in 2011 was foreign investors. Net purchases in 2011 are on track to reach $290bn with most of the buying occurring in 3Q11 after the end of QE2. While this is significantly lower than the magnitude of purchases in the last three years of $640bn per year, it is higher as a percentage of effective supply. Foreigners took down 64% of effective
119
Exhibit 30: The share of USD in FX reserves has stabilized at around 65%
Estimated* share of USD in foreign exchange reserves; $bn
Exhibit 31: Despite our expectation of increased foreign demand, the overall supply/demand imbalance will worsen next year given the lack of Fed buying
Net issuance of Treasuries per calendar year versus purchases by different investor types; $bn
2009 Treasury net issuance; $bn 1476 2010 1592 2011* 1093
74% 72% 70% 68% 66% 64% Jun 00
2012* 1028
Jun 02
Jun 04
Jun 06
Jun 08
Jun 10
Source: IMF, Bloomberg * Assumes that the allocation of Japans and Chinas FX reserves mirror the distribution of FX reserve currency allocation of advanced economies and Emerging economies, respectively.
Investor Foreign Investors 555 Commercial banks 92 Pension funds/ Insurance companies 211 Broker dealers -68 Mutual Funds 68 Federal Reserve 300 Other 318 Source: J.P. Morgan, Federal Reserve * 2011 and 2012 are J.P. Morgan estimates.
654 113 224 -29 42 244 344
290 -75 130 100 45 639 -36
540 -35 150 -20 60 0 333
Exhibit 32: Models for estimating Treasury buying by pension funds, insurance companies and commercial banks
Empirical models for quarterly purchases by pension funds/ insurance companies and commercial banks; $bn
supply (i.e., net issuance excluding Fed purchases), higher than the 3-year average of 51% (Exhibit 28). The composition of demand was favorable and foreigners heavily favored coupon securities to bills. While overall foreign demand for Treasuries has been strong since the end of QE2, Treasury data for the third quarter shows a significant shift in the composition of foreign buying. Demand from EM reserve managers weakened during the quarter as they slowed their building of USD FX reserves, while demand from the EU increased in response to the sovereign debt crisis (Exhibit 29). While we view the recent slowing in EM demand as mostly a cyclical phenomenon, it is also likely a sneak preview of a longer-term trend underway for many EM countries to gradually slow their reserve accumulation. Officials in China have become increasingly vocal about the need to slow FX reserve building and diversify away from the US dollar (see US Fixed Income Markets Weekly, September 19, 2011). At the same time, this diversification has been slow-moving and is likely to remain so (Exhibit 30). On balance, we view weaker sponsorship by EM reserve managers as a risk to the Treasury market rather than our base case. In the meantime, softer demand from the EM is likely to be offset by the flight-to-quality bid from the EU. Exhibit 31 shows our 2012 projections for Treasury demand by investor type. The estimates are based on a set of simple empirical models discussed in more detail below (see box and Exhibit 32). For foreign investors, we
120
Pension funds/ insurance companies Variable Coefficient Intercept 35.9 30-year Treasury yields; % -7.8 10s/30s curve; % 18.5 Quarterly effective net issuance of Treasuries; $bn 0.04 Projected purchases in 2012: Commercial banks Variable Intercept 30-year MBS CC Libor OAS; bp 2s/5s Treasury curve; % Quarterly effective net issuance of Treasuries; $bn Projected purchases in 2012:
are for the full year 2012.
T-stat 1.7 -2.1 2.7 2.6
Projection 1 3.25 1 260 150
Quarterly data regressed over 15-years; R2=54%; Standard error = 62
Coefficient -5.4 -0.6 7.0 0.06
T-stat -1.23 -4.7 1.4 5.4
Projection 43 0.8 260 -35
Quarterly data regressed over 5-years; R2=74%; Standard error = 10. Projections
Source: J.P. Morgan, Federal Reserve H.Z.1
project total purchases of $540bn or 53% of net Treasury issuance; by comparison, foreign purchases amounted to 64% of net issuance (net of Fed purchases) in 2011 and 48% in 2009-2010 (Exhibit 31). With the curve flat and rates low, we expect demand from pension funds and insurance companies to be modestly below average at around $150bn (Exhibit 32). We also expect commercial
banks to be net sellers of coupon Treasuries next year, likely preferring MBS given its cheapness and the relatively flat front end of the Treasury yield curve. Finally, we expect mutual fund demand and demand from broker-dealers to be in line with their recent averages. In total, demand across the six largest investor types sums to $695bn. This suggests that a comparatively large amount must be funded by other investors, highlighting a significantly worse supply/demand imbalance than in 2011. One potential offset is the proceeds of maturing FDIC guaranteed bank debt: $45bn is due to mature in December 2011, and another $165bn is due to mature in 2012 (Exhibit 33). While little data is available on the buyers of this product, our best guess is that they were primarily Treasury/Agency debt buyers and that the proceeds will likely make their way back into the front end of the either the Treasury or the Agency debt market.
Exhibit 33: Redemptions of FDIC guaranteed bank debt will be heavy in 2012 providing some support for Treasuries
FDIC guaranteed bank debt redemptions; $bn
60 50 40 30 20 10 0 Nov 11 Jan 12 Mar 12 May 12 Jul 12

Source: Bloomberg
Sep 12 Nov 12
Forecasting foreign demand for Treasuries

In order to forecast foreign demand for Treasuries, we model average monthly foreign Treasury purchases (available from the Federal Reserve Flow of Funds release) as a function of the US trade deficit with the rest of the world and net issuance of Treasuries (i.e., Treasury net issuance minus Fed Treasury purchases via quantitative easing). The trade deficit of the US with the rest of the world is a significant driver of foreign purchases of US securities, since it leads to an accumulation of USD reserves abroad. In addition, the magnitude of foreign purchases of Treasuries also depends on available supply as foreign investors have shown a willingness to increase their buying in response to larger US deficits. Exhibit A1 shows the details of this model, and Exhibit A2 shows that the model has tracked actual purchases reasonably well since 2002. The model shows that each additional $10 bn increase in the monthly trade deficit results in an additional $2bn of foreign buying. Similarly, each additional $10 bn increase in monthly Treasury supply results in an additional $4.3 bn of foreign buying. Given our expectation that the trade deficit and effective net issuance will average $48bn and $85bn per month in 2012, respectively, we look for foreign buying of Treasuries to average $45bn per month or $540bn for the year, representing 53% of Treasury net issuance next year.
Exhibit A1: A model for forecasting foreign purchases of Treasuries

Net purchases by foreigners modeled as: Variable Coeff Intercept -0.3 Trade surplus; $bn -0.20
Exhibit A2: Actual versus predicted monthly average of Treasury purchases by foreigners; $bn
70 60 50 40 30 20 10 0 2002 2005 2008 2011 Actual Model
T-stats -0.1 -2.5
Proj. 2012 -48.0
Net issuance of Treasuries, net of Fed purchases; $bn 0.43 19.1 85.0 Projected foreign purchases of Treasuries; $bn/ month 45 R Square: 91% Std. Error: 5.6 Regression uses rolling 1-year monthly averages and is fitted on quarterly data over 10-years.
Source: Federal Reserve Z.1, US Census Bureau, Treasury
121
With the Agency debt market shrinking every year, Treasuries are likely to be the beneficiary of this demand.
Ratings Risk and the US Fiscal Outlook

One of the great ironies in the Treasury market this year is the fact that 10-year yields reached all-time lows in the immediate aftermath of S&Ps downgrade of the US sovereign rating. The loss of the US AAA rating followed three consecutive years of record deficits that caused net federal debt to increase to 73% in 2011 from 40% of GDP in 2008. In August, with a divided Congress unable to reach a consensus on how to stabilize US fiscal metrics and debt levels projected to continue to grow, S&P fired a warning shot to the US about the need to get its fiscal house in order. While a lot more work lies ahead for the US to stabilize its finances, the good news is that the debate changed dramatically in 2011, and a down payment on spending cuts was made with the passage of the Budget Control Act (BCA) of 2011. These cuts include $900bn agreed to in August and another $1.2tn in automatic spending cuts required by the bill, with the Deficit Reduction Committee failing to reach agreement. With these cuts, federal net debt is projected to grow from 73% of GDP currently to 84% by 2021; while still on an uptrend, the trajectory is much flatter than estimates prior to the passage of BCA, with July CBO projections for 2021 equal to 95% of GDP (Exhibit 34). In order to stabilize debt levels, we estimate Congress still needs to reduce the structural deficit by an additional 1%-pt of GDP per year, or $1.8tn over the 9year period. These additional cuts are not likely in an election year but are certainly possible in 2013. Exhibit 35 presents a short menu of alternatives that could be used to reduce the structural deficit further; any combination that totaled an additional 1% (beyond the BCA) would stabilize the ratio of US debt to GDP. To be sure, these projections are sensitive to a number of assumptions including economic growth. CBOs baseline projections, upon which Exhibit 34 is based, assume real GDP growth will average 2.8% from 2012-2021. If growth averaged 1% per year lower, we estimate debt levels would deteriorate by 5%-pts, reaching 89% of GDP by 2021 (Exhibit 36). In order to stabilize debt-to-GDP under this weaker growth outlook, Congress would need to find an additional $2.3tn in savings beyond the $1.2tn in automatic cuts triggered by the Budget Control Act. With some, albeit modest, progress made on budget reform this year, we expect S&P and Moodys to maintain their existing US sovereign ratings in 2012, although Fitch
122
Exhibit 34: Half way home: Debt to GDP in the US is still on an upward trajectory; another $1.8tn in cuts are needed to stabilize debt levels
Projected Net Federal debt to GDP ratio*; %
100 95 90 85 80 75 BCA+ 1.8 tn Pre BCA baseline
BCA baseline
70 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 * CBOs August alternate baseline and J.P. Morgan estimates
Exhibit 35: Alternatives to reduce the structural deficit: any combination that totals 1% of GDP will stabilize the ratio of US debt to GDP
Various alternatives to reduce the structural deficit; as percentage of GDP and $bn Cumulative Options % of GDP deficit reduction* Sunset all Bush tax cuts 1.4 2459 Sunset tax cuts on wealthy (per Obama**) 0.4 700 Feldstein 2% itemized deduction limit 0.2 350 Cap itemized deductions at 28% 0.2 293 Asset and spectrum sales; user fees 0.1 200 Tax international income 0.1 133 *Cumulative deficit reduction relative to BCA baseline for 2013-2021 ** Extend cuts for families with income under $250,000
Exhibit 36: 2021 debt levels will deteriorate 5%-pts if growth slips by 1% per year
Projected Net Federal debt to GDP ratio in different growth scenarios; %
90
85
1.8% real GDP growth
80
Baseline 2.8% growth

75
70 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Exhibit 37: Comments by rating agencies on further downgrades

Current US Rating Agency sovereign rating Comment on Super Committee
Exhibit 38: US sovereign CDS spreads are below those of countries with AAA ratings
5-year sovereign CDS spreads; bp
Country Norway United States Sweden Finland Australia Britain Germany Netherlands Denmark Austria France S&P Rating Outlook AAA Stable AA+ Neg AAA Stable AAA Stable AAA Stable AAA Stable AAA Stable AAA Stable AAA Stable AAA Stable AAA Stable Moody's Rating Outlook Aaa Stable Aaa Neg Aaa Stable Aaa Stable Aaa Stable Aaa Stable Aaa Stable Aaa Stable Aaa Stable Aaa Stable Aaa Stable Rating AAA AAA AAA AAA AAA AAA AAA AAA AAA AAA AAA Fitch Outlook Stable Stable Stable Stable Stable Stable Stable Stable Stable Stable Stable 5y CDS 45 52 66 70 90 92 97 117 123 215 230
S&P
"The Fiscal Committee's inability to agree on fiscal measures that would stabilize U.S. government debt as a share of GDP is AA+; consistent with our Aug. 5 decision to lower our rating to 'AA+'. Negative However, we expect the caps on discretionary spending as laid outlook out in the Budget Control Act of 2011 to remain in force. If these limits are eased, downward pressure on the ratings could build."
"The Aaa government bond rating for the United States is unaffected by the lack of a deficit reduction agreement by the Aaa; Joint Select Committee on Deficit Reduction...While a change in Moody's Negative the composition of the spending cuts would not be a major rating outlook consideration, a reduction in the total amount that would increase the projected increase in federal debt over the coming decade could have negative rating implications" "Failure by the Super Committee to reach agreement would likely AAA; result in a negative rating action -- most likely a revision of the Stable rating Outlook to Negative...Less likely would be a one-notch outlook downgrade. Fitch now expects to conclude its review of the US sovereign rating by the end of November."
Source: Bloomberg
Exhibit 39: Location, location, location: carry was a contrarian indicator of performance for front-end curve trades in 2011
3-month holding period P/L from going long the belly (5s) versus a level and curve neutral combination of wings (2s/10s and 2s/7s) in 2011 vs. the 3-month carry and roll in the trade at time of initiation between 1/1/11 and 8/12/11; bp
Fitch
30 20 10 0 -10 -20 -30 -6 -4 -2 0 2 3m carry and roll at initiation; bp 4
2s5s10s 2s5s7s
is likely to move to a negative outlook. In their recent statements, S&P and Moodys both noted that the US rating would not be lowered in response to the failure of the Joint Select Committee, as long as the automatic spending cuts of $1.2tn are implemented (Exhibit 37). Fitch has been more negative in its comments around the failure of the Super Committee and we expect them to place the US on negative outlook while maintaining the AAA rating. Finally, we note that derivative markets also appear sanguine about the US fiscal outlook and the near-term risk of further downgrades. CDS on US Treasuries have been remarkably stable since August, outperforming most other AAA-rated sovereigns. With the exception of Norway, US CDS is now tighter than every other AAArated sovereign; spreads are currently 45bp tighter than Germany and 180bp tighter than France (Exhibit 38).
operations, and cross-currency arbitrage that involves creating synthetic Treasuries with cheap foreign bonds. Yield curve trading While yield curve carry trades typically perform well when the Fed is on hold and rates are range-bound, the current environment will likely prove more challenging with front-end rates so low. In 2012, our approach to yield curve trading will generally be driven by carry considerations. But, in contrast to conventional wisdom, our bias will be to favor negative carry curve trades in the front end of the curve where convexity is attractive. In intermediates, however, we expect positive carry trades to outperform and will generally concentrate long carry exposure in the intermediate to long end of the curve. Some support in favor of avoiding positive carry yield curve trades in the front end of the curve is highlighted in
123
Trading themes for 2012

While Treasury yields are expected to remain low in 2012, volatility will still be high, creating regular opportunities for active investors to enhance the return on their investment portfolios. In the discussion below, we discuss four broad areas of opportunities for Treasury investors in 2012 including yield curve carry trading, supply cyclicals, relative value opportunities around Fed purchase
Exhibit 40: A long position in 5s versus 2s and 10s increasingly becomes a steepener in a rally and a flattener in a sell-off
3-month holding period P/L from initiating a long position in a level and curve neutral 2s/5s/10s butterfly when it carries positively (bp), regressed against the change in the 2s/10s curve during the holding period (%)
Exhibit 41: The performance of intermediate butterflies is well correlated with carry at inception
3-month holding period P/L of a long position in 7s versus 5s and 10s versus 3month carry and roll in the trade at inception; bp
15 10 5 0 -5 -10 -15 -20 -25 -30 -35 -1.2 -1.0
10 8 6 4 2 0 -2 -4 -6 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 3m carry and roll in weighted trade at time of initiation; bp 1.0 Y = 3.0385 X1 + 4.1926 R = 63.35% standard error = 2.0730 period = Jan 01,11 - Aug 12,11
Y = -4.03 X1 - 19.43 X1^2 - 5.34 R = 29.0% standard error = 8.26 period = Oct 12,10 - Aug 12,11 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 Change in 2s/10s curve over the holding period; % 0.8
Exhibit 39, which compares the performance of 2s/5s/7s, and 2s/5s/10s weighted butterfly trades this year to ex ante carry and roll on the trade. The weights on each trade were designed to make the butterfly curve and level neutral and returns are calculated over 3-month holding periods. The Exhibit highlights that while carry is a good leading indicator of P/L on these trades, it is a contrarian indicator; the more negative the carry, the larger the subsequent P/L. Initiating these weighted curve trades when carry was positive in 2011 has resulted in a loss 80% of the time, with an average 3-month P/L of -8bp. The poor performance of front-end butterflies during 2011 arises from the negative curve convexity on these trades. As shown in Exhibit 40, positive carry 2s/5s/10s weighted butterflies that are long the belly tend to become steepeners as the curve flattens and become flatteners as the curve steepens. This behavior reflects the changing volatility of 2-year yields with 2s becoming sticky, as rates fall and the curve flattens, and more volatile as rates rise and the curve steepens. To offset these changing volatilities, rebalancing the position would involve adding flattening risk (i.e., selling more 2s versus 5s) when the curve flattens to offset the falling volatility of 2s. Similarly, rebalancing involves adding steepening risk (i.e., scaling back the short in 2s versus 5s) when the curve steepens to offset the increasing volatility of 2s. In contrast to carry trades anchored in the front end, we favor positive carry yield curve trades anchored in the intermediate sector where convexity is less of an issue. Exhibit 41 shows that the 3-month holding period returns from being long 7s versus a risk-weighted 5s/10s barbell
124
Exhibit 42: Butterflies in the long end of the curve are more likely to be mean-reverting than front-end butterflies
1-month holding period returns from initiating level and curve neutral butterflies when the belly is mispriced by at least 1.5 standard deviations in 2011; bp unless otherwise specified
Bfly 10y15y20y 10y12y15y 5y10y15y 5y7y10y 7y15y20y 5y10y20y 3y5y7y 2y5y10y 2y5y7y 1y3y5y 10y20y30y 2y3y5y 1y2y3y Avg P/L Min Max # Trades Hit rate Std Dev Risk adj returns 1.5 1 3 20 100% 0.5 3.00 0.5 0 1 21 100% 0.2 2.88 5.0 -7 8 20 95% 3.3 1.50 3.8 -2 7 8 88% 2.8 1.36 1.8 0 4 34 85% 1.4 1.28 6.2 -10 12 34 91% 5.8 1.06 2.1 -4 7 14 71% 3.8 0.55 1.5 -11 12 17 65% 6.7 0.22 0.7 -9 6 16 63% 5.0 0.15 0.2 -5 28 16 13% 8.0 0.03 -0.2 -4 9 5 20% 5.3 -0.04 -0.3 -4 6 19 37% 2.4 -0.13 -0.6 -3 1 2 50% 2.8 -0.20
have been well correlated with the carry and roll on the trade. In addition to carry, we note that butterflies anchored in the long and intermediate sector of the curve have been more mean reverting than front-end butterflies. To determine how mean reversion of various butterfly trades performed in 2011, we tested a simple trading rule that involved initiating a long position in the belly versus the wings when the belly appeared cheap by more than 1.5 standard deviations on a curve and level neutral basis. The reverse was done when the belly appeared rich. The trade was then held for a 1-month period. Exhibit 42 shows the results of such a trading rule and shows that intermediate and long-end butterflies had higher risk-adjusted returns by following such a strategy. By contrast, front-end curve
Exhibit 43: Supply concessions have been the largest and most consistent going into the bond auction, a trend we expect will continue in 2012
Exhibit 44: While supply concessions ahead of the bond auction have fallen since Operation Twist began, they are still sizable
Average P/L in select weighted curves* from being short the sector being auctioned in the days prior to its auction versus being long the sector in the days post auction calculated between January and October 2011; bp unless otherwise specified Centered Pre auction Post auction Wtd. around Trade auction in Avg Chg Hit rate Avg Chg Hit rate
Average change in 30-year - 0.85 *7-year curve in eight business days before the bond auction during QE2 (January to June 2011), post QE2 (July and August) and Operation Twist (October and November); bp 10
8 6 4 2 0 QE2 Jul-Aug Average in different periods Operation Twist
2y 2s3s 1.5 80% 1.0 60% 2y 2s10s 2.4 50% 1.7 60% 3y 2s3s 1.0 80% -0.6 50% 3s5s 1.0 40% 1.6 60% 3y 3s7s 1.4 40% 0.7 80% 3y 3s5s -0.8 50% -0.3 60% 5y 5y 5s10s 0.9 40% 0.7 60% 5y 5s30s 3.0 50% 2.5 70% 7y 3s7s 2.5 70% 4.6 70% 7y 7s10s 0.3 50% 0.5 60% 7s30s 2.7 60% -1.2 50% 7y 2s10s 0.4 70% 1.1 50% 10y 5s10s 1.1 70% 0.6 30% 10y 10y 7s10s 1.6 70% 0.0 40% 30y 2s30s 1.6 80% 1.1 60% 30y 5s30s 4.1 80% 1.1 40% 30y 7s30s 4.0 80% 3.6 50% * Weighted trades include 0.57*3s-2s, 0.31*10s-2s, 0.57*5s-3s, 0.58*7s-3s, 0.85*10s-5s, 0.64*30s-5s, 0.95*10s-7s, and 30s-0.85*7s. Curves are weighted to be level neutral.
2011 YTD average
Exhibit 45: Poor liquidity tends to amplify auction cyclicals

30-year 0.85 * 7-year Treasury rates averaged in the business days around the 30-year auction in periods of high market depth* and low market depth (Data over last 3-years has been used); %
2.16 2.11 2.06 2.01 1.96 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 Business days around 30-year auction High market depth Low market depth
1.85 1.80 1.75 1.70 1.65
trades exhibited weak mean reversion, with the simple trading rule producing poor risk-adjusted returns. In sum, our yield curve bias in 2012 will be to favor negative carry curve trades in the front end of the curve where convexity is attractive. In intermediates, we expect positive carry trades to outperform and will look to add to long carry positions as high carry points become cheap on the curve. Supply cyclicals Trading strategies around Treasury auction concessions had a mixed track record in 2011, with returns weakest in trades around 2-, 3-, and 5-year auctions. The largest and the most consistent supply concessions have been in the very long end of the curve. Exhibit 43 shows the performance of various trades in the days before and after auctions in different sectors. As shown in the exhibit, hedged for the level of rates, the 7s/30s curve has steepened going into the bond auction in 9 out of 11 months this year by an average of 4bp. Notably, this steepening has persisted despite Fed purchases in the very long end via Operation Twist. Exhibit 44 shows that even
* Market depth is calculated as the 10-day moving average of average size of the top three bids and offers, in $mn, for the on-the-run 10-year Treasury note, averaged between 8:30 a.m. and 10:30 a.m. daily. Periods with market depth of greater than $100mn are classified as high market depth periods.
though the magnitude of the supply concession has been smaller since Operation Twist began, it is still sizable. With liquidity poor, we expect these concessions to persist in 2012, and we favor curve steepeners going into bond auctions. Our outlook is motivated in large part by our expectation that market depth will likely remain poor in 1H12, as bank deleveraging continues. Historically, supply concessions have increased in periods of poor market liquidity, as shown in Exhibit 45. Thus, we
125
Exhibit 46: A strategy of buying issues with the highest yield errors versus issues with the lowest yield errors has been profitable during Operation Twist
Results from back-testing a simple trading rule*; table shows average change**, standard deviation of changes, minimum change, average change/ standard deviation and hit rate*** for par asset swap spread difference and yield difference for each bond pair
Par asset swap Yields Yrs to mat Avg chg Min chg Avg/Std.Dev Hit Rate Avg chg Min chg Avg/Std.Dev Hit Rate 6-7 1.3 -1.3 0.8 80% -1.0 -4.3 -0.6 40% 7-8 2.1 0.3 1.9 100% 0.9 -2.1 0.6 67% 8-9 -0.7 -2.2 -0.6 20% 0.8 -1.0 0.7 73% 9-10 -2.0 -4.9 -1.0 13% 0.9 0.4 2.6 100% 10-15 1.3 -4.1 0.5 80% -1.9 -5.2 -0.9 13% 15-20 -1.6 -3.2 -1.5 7% -0.4 -2.8 -0.3 53% 24-27 1.2 0.6 3.2 100% 0.8 0.1 1.6 100% 27-30 0.3 -0.6 0.6 80% 0.5 -0.3 0.9 80%
Exhibit 47: Mispricings are largest in the 8- to 10-year sector of the curve
RMSE of Treasuries in select parts of the Treasury curve; bp
4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0
6y-8y 8y-10y 20y-30y
Latest observation
* Trading rule involves buying the issue with the highest yield error in each maturity bucket versus issue with the lowest yield error in each bucket on a daily basis starting September 21, 2011. Holding period is one month. ** Positive number for average change implies that bond with highest yield error outperformed the bond with the lowest yield error *** Hit rate equals number of trades in which spread widened/ total number of trades in each maturity bucket
May 11
Aug 11
Nov 11
Exhibit 48: New-issue 10s have cheapened relative to old 10s in the weeks after they have been auctioned
Adjusted* and unadjusted on-the-run / triple olds yield spread in the days after a new 10-year is auctioned; %
recommend that investors actively position for a steeper curve ahead of bond auctions, particularly in periods in which risk aversion is elevated and balance sheet is constrained. Trading around the Fed The Feds purchase operations still continue to provide short-term trading opportunities in the Treasury market. Its preference to consistently purchase securities that are cheap to a fitted curve has resulted in further compression in the dispersion of valuations along the Treasury curve, as measured by the root mean square error of a fitted Treasury curve (RMSE). Even though this has been a well-identified trading theme for some time (through QE1 and QE2), the strategy of owning issues cheap on the curve in anticipation of Fed buying has remained profitable, even under Operation Twist. Exhibit 46 presents the results from following a simple daily strategy of buying issues that are cheap on the curve and selling similar maturity issues that are trading rich; the strategy has been most profitable in the 6- to 8-year sector, and 25to 30-year sector. As a result, the dispersion of valuations in these sectors has compressed to their lowest level in a year (Exhibit 47). On the other hand, as also shown in the exhibit, dispersion in the 8- to 10-year sector still has room to compress further relative to its recent history. Typically, the richest issue in this sector is the on-the-run 10s, and these have historically tended to cheapen relative to the older securities as they get reopened in the weeks
126
-0.030 -0.035 -0.040 -0.045 -0.050 -0.055 -0.060 -0.065 -0.070 0 10 20 30 40 50 # of business days after a new 10-year auction 60 New/ olds spread adj. for 7s10s curve New/ olds spread
0.160 0.155 0.150 0.145 0.140 0.135 0.130 0.125 0.120
* Adjusted for the 7s/10s yield curve. Equals on-the-run 10-year yield triple olds yield 0.25 * 7s/10s curve
Exhibit 49: The yield pickup of synthetic assets constructed by buying JGBs/ Bunds and swapping back to USD is substantial
Yield on synthetic asset* minus Treasury yield for various government bonds and sectors; bp Annualized riskSpread pick of synthetic adjusted return** Sector versus US Treasuries; bp
JGB Bunds JGB Bunds UST 2y 117 30 2.1 0.2 0.3 3y 124 23 1.6 0.2 0.2 5y 133 15 1.1 0.4 0.4 * 2-year synthetic yield equals JGB 2-year yield basis swap spread + (USD swap yield Yen swap yield) + 6M/3M Libor basis. ** 3-month return/ risk. 3-month return is one-fourth the yield plus the roll. Risk is (5-year standard deviation of 3-month changes in yield)*(modified duration).
after they are first auctioned (Exhibit 48). In general, we would favor overweighting off-the-run 10s versus newly issued on-the-runs to benefit from supply-related cheapening as well as the Feds preference for purchasing cheap off-the-run issues.
Creating front-end synthetic Treasuries The cross currency basis of USD with JPY and EUR was pushed further into negative territory this year, as the pullback by US money funds lending to European banks resulted in increased demand for USD funding. At its current level of -80bp, the 2-year JPY/USD cross-currency basis is near the lows of the last decade, while the EUR/USD basis is at its lowest since December 2008. This has made it very attractive for investors to create synthetic USD assets constructed by buying foreign government bonds that are swapped back to USD (see grey box).
Synthetic Treasuries and cross currency basis swaps A cross-currency basis swap is an agreement to exchange floating rates in a foreign currency against USD Libor. The transaction involves an exchange of foreign currency up front with floating payments exchanged over time. The floating payments equal USD Libor versus foreign Libor plus a spread where, by convention, the spread on the foreign Libor is referred to as the basis swap spread (Exhibit A). Currently, spreads for 2-year USD/JPY basis swaps equal -80bp meaning an investor can pay 3-month Yen Libor minus 80bp versus receiving 3-month USD Libor flat for the next eight quarters. The large negative basis spread currently makes it attractive for a USD floating-rate investor to create synthetic floating rate USD assets by buying floating rate assets in Yen and swapping them to USD floating-rate assets with a cross-currency basis swap. For example, a US investor can synthetically create a 2-year floating rate USD asset that earns USD Libor + 80 bp by 1) investing in a Yen floating-rate asset that pays Yen Libor flat and 2) entering in a cross-currency basis swap at a spread of -80. The yield pick-up embedded in the negative basis swap also creates opportunities for fixed-rate investors. An investor benchmarked to 2year Treasuries, for example, can create a synthetic 2-year Treasury note with a much higher yield by instead buying 2-year JGBs and using the basis swap and interest rate swap markets to convert the cash flows back into fixed-rate USD flows. To convert the JGB cash flows into fixed-rate USD flows, the investor 1) buys the JGB and asset swaps it into a semiannual Yen floater, 2) uses a 3s/6s basis swaps to convert JPY semiannual cash flows into quarterly cash flows, 3) enters a cross currency basis swap (paying Yen Libor + spread versus receiving USD Libor) to convert it to a USD floater, and 4) converts the floating rate USD cash flows into fixed semiannual cash flows using plain vanilla interest rate swaps. Of course, these trades can be collapsed into a single trade to effectively convert the 2-year JGB cash flows into a 2-year semiannual fixed-rate USD asset. The cash flows on this trade have the investor paying a Yen coupon equal to the yield on the JGB (offsetting JGB cash flows) and receiving a USD coupon (Exhibit B). The yield on the synthetic 2-year Treasury can be calculated as: Synthetic yield = JGB asset swap spread + 6M/3M Libor basis basis swap spread + USD fixed swap rate At current levels, this implies Synthetic 2-year yield = -25 + 12 - (-80) + 77 = 1.44% or 117bp higher than 2-year Treasury notes. Exhibit A: Basic cross currency swap
JPY 7,716 mm initial exchange
Exhibit B: 2-year fixed semiannual currency swap

JPY 7,716 mm initial exchange
Investor
Yen Libor + spread USD Libor
Dealer
Investor
JPY 0.12% USD 1.44%
Dealer
US $ 100 mm initial exchange
US $ 100 mm initial exchange
127
Exhibit 50: and at multi-year highs

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 May 10 Nov 10
Exhibit 51: The size of the STRIPS market fell in 2H11

P-STRIPS outstanding; $bn
Synthetic 2-year Treasury yield* minus 2-year US Treasury yield; %
210 205 200 195 190 185 180 175 170

May 11 Nov 11
Nov 09
May 10
Nov 10
May 11
Nov 11
* Synthetic 2-year Treasury yield equals JGB 2-year yield basis swap spread + (USD swap yield Yen swap yield) + 6M/3M Libor basis.
Exhibit 52: STRIPS in the very long-end of the curve continued to gain market share in 2011
P-STRIPS outstanding by sector as a percentage of total; %
JGBs in particular look very attractive when swapped back into US dollars. A trade constructed in this fashion in the 2-year sector currently results in a pick-up of 117bp versus similar maturity Treasuries (Exhibit 49). Not only is this incremental yield high relative to recent history (Exhibit 50), but the unlimited USD swap lines provided by the Fed limit the potential widening in the EUR/JPY basis. Currently, the 3-month JPY/USD FX basis is trading at -70bp. Converting this basis (which is quoted versus Libor) to an OIS basis (OIS basis = Libor basis + JPY 3M Libor/OIS spread USD 3M Libor/OIS spread) translates to an OIS basis of -100. Because this spread is equivalent to the penalty rate the Fed charges banks to use its currency swap line (OIS+100), we should not expect any further widening in the 3-month basis. This limits the downside risks in buying JGBs swapped to US dollars.
70 60 50 40 30 20 10 0 2002 2005
<5.5-yrs 5.5-to 10-yrs 10- to 17-yrs 17+ yrs
2008
2011
Exhibit 53: Low yields, a flat curve, and Operation Twist are likely to keep net stripping activity muted in 1H12
A model for quarterly changes in P-STRIPS outstanding; $bn
Outlook on STRIPS
The size of the P-STRIPS market is unchanged this year, with $197bn outstanding. However, the intra-year pattern exhibited more volatility: net stripping activity picked up in Q2 bringing the market to its largest size in a decade, but fell in 2H11. The quarter ending October 2011 saw the largest 3-month decline in the size of the market since January 2009, as the very long end of the curve flattened massively and Operation Twist started taking long-end supply out of the market (Exhibit 51). Investor demand for STRIPS remained largely in the very long end of the curve. P-STRIPS with more than 17 years to maturity now
Variable Intercept 10-year yields; % 10s/30s curve; % Change in S&P 500; points Quarterly Treasury duration purchases by the Fed; $bn of 10-year equivalents R2 = 50%; Std. error= 5
Coefficient T-statistics -55.3 -5.7 10.5 5.4 24.9 5.8 0.01 1.7 -0.0209 -1.5
Model fitted for quarterly data over a 5-year period. 10-year Treasury rates and the 10s/30s curve are quarterly averages.
comprise 66% of the market versus 60% at the beginning of the year (Exhibit 52).
128
Exhibit 54: Operation Twist Impact: Issues that were purchased in size by the Fed in October were also reconstituted the most during the month
Change in stripped amount for the top 10 issues purchased by the Fed in the 10to 30-year sector in October versus Fed purchases of the bond in October; $bn
Exhibit 55: Issues with high yield errors like Aug-40s and May-40s will likely lead reconstitution activity given the Feds preference for issues trading cheap on the curve
Amount stripped by issue ($bn) versus current yield error (bp)
1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Fed purchases in Oct; $bn
30 25 20 15 10 5 0 -2.5 -2.0 -1.5 N-41
A-41
M-41
F-41 N-39 N-40 F-40 F-36 M-38 F-38 M-37
A-40 M-40
A-39 M-39 F-39 F-37
-1.0
-0.5
0.0
0.5
Yield error; bp
Looking ahead, we expect net stripping activity to stay muted and for the market to continue to shrink in 1H12. This view is in part motivated by our outlook that the market environment will be characterized by high risk aversion as the crisis in Europe escalates, which will keep yields low and the very long end of the curve flat causing risky assets to underperform. Historically, a flat curve, low yields, and weak equities have all caused the STRIPS market to shrink (Exhibit 53). Fed purchases of Treasuries via Operation Twist will also have a significant impact and is likely to result in increased reconstitution activity for the purpose of delivering whole bonds into the Fed. As a rule of thumb, quarterly Treasury duration purchases of $100bn of 10year equivalents by the Fed have lowered P-STRIPS outstanding by around $2bn per quarter. In 1H12, we expect duration buying by the Fed to average $180bn of 10-year equivalents per quarter, suggesting that the market should shrink by $4bn per quarter due to this factor alone. With 32% of Fed purchases in maturities greater than 20years and the bulk of the STRIPS market concentrated in that sector of the curve, we expect reconstitutions to be particularly high. Indeed, there was already evidence of this in October (the first month of Operation Twist). Exhibit 54 shows that issues that were purchased in size by the Fed in October were also the ones that were reconstituted the most. In sum, we look for the STRIPS market to shrink further in 1H12. Whole bonds that are trading cheap relative to the fitted Treasury curve will likely lead the
Exhibit 56: Increased reconstitutions are likely to drive C-P spreads narrower in the coming months
Coupon minus Principal STRIPS yield spread in the 2022-23 sector* (%) versus the total amount of P-STRIPS outstanding ($bn) % $bn
0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 2010 2011 Average C-P spreads; % P-STRIPS outstanding; $bn
210 205 200 195 190 185 180 175 170
* Average C-P spread between issues maturing on 8/15/22, 11/15/22, 2/15/23 and 8/15/23.
way in reconstitution activity (Exhibit 55), since the Fed concentrates its purchases in such issues (almost 75% of its purchases have occurred in positive yield error bonds thus far in Operation Twist). This has two implications for the STRIPS market. First, shorter maturity Cs should richen versus similar maturity Ps, since reconstitution of longer maturity bonds will reduce the relative supply of shorter maturity Cs. Indeed, this has been one driver of the C-P spread recently (Exhibit 56). This richening of Cs will likely occur most in Cs that correspond to coupon dates of highly stripped whole bonds that have positive yield
129
errors in the very long end of the curve. Second, Ps corresponding to bonds that are likely to be reconstituted should richen relative to other surrounding Ps. In sum, we expect the size of the STRIPS market to decline further in 2012. Short maturity Cs should richen versus similar maturity Ps, as should Ps on long-dated bonds that are currently cheap to the par curve.
carry curve trades in the front end of the curve where convexity is attractive. In intermediates, we expect positive carry trades to outperform and will look to add to long carry positions as high carry points become cheap on the curve. The long end of the curve is biased slightly steeper; our mid-year target for the 10s/30s Treasury curve is 110bp The 10s/30s curve is slightly flat to fair value; we favor positioning for a steeper curve in advance of long-end supply to take advantage of the large supply concession in the long end. An increase in reconstitutions in the STRIPS market should cause shorter maturity Cs to outperform Ps We expect the size of the STRIPS market to decline further in 2012. Short maturity Cs should richen versus similar maturity Ps, as should Ps on long-dated bonds that are currently cheap to the par curve.
Trading themes
Stay long duration in early-2012 targeting 10-year yields to reach 1.70% in 1Q12 A worsening of the European sovereign debt crisis should accelerate the flight-to-quality bid into Treasuries early next year pushing yields lower; rich valuations, a deteriorating supply/demand imbalance, and poor technicals will limit the upside, however, and keep rates range bound later in 2012. Position for lower front-end yields in 2012 Front-end rates should trend lower in 2012 driven by declining front-end supply, further policy rate guidance by the Fed that flattens term premium, an increase in the supply of excess reserves following QE3, and the end of Operation Twist; look for GC repo to average near 5bp, and 2-year yields to fall to 17bp. Avoid crowds in 2012 A secular decline in liquidity will lead to heightened whipsaw risk in 2012 and cause Treasury yields to be volatile within a broad trading range; trading strategies based on our measure of investor positions that fade crowded consensus trades are likely to perform well in 2012 as they did this year. Create synthetic 2-year Treasury notes by buying JGBs and currency swapping them back to USD Synthetic Treasuries constructed by buying foreign bonds and currency swapping them back to US dollars are at the most attractive levels of the last decade. The yield on a synthetic 2-year note created by asset swapping a 2-year JGBs currently equals 1.44% or 117 bp higher than 2-year Treasury yields. We expect positive carry yield curve trades in the front end to perform poorly in 2012; favor convexity over carry in the front end of the curve and carry in the intermediate sector of the curve Positive carry yield curve trades in the front end are likely to perform poorly reflecting the extreme negative convexity in these trades. We recommend negative
130
US Fixed Income Strategy US Fixed Income Markets 2012 Outlook November 24, 2011 Srini RamaswamyAC (1-212) 834-4573 Praveen Korapaty (1-212) 834-3092 Alberto Iglesias (1-212) 834-5116 J.P. Morgan Securities LLC
US Interest Rate Derivatives

There is a new normal in the swaps market convexity-hedging flows and issuance-related swapping matter less, but European markets and Fed purchases matter more FRA-OIS spreads should remain under pressure into 1Q12, but concerted central bank policies should cause it to narrow thereafterlook for FRA-OIS to widen in the short term, but eventually narrow to 50bp by mid-year. Intermediate maturity swap spreads are similarly biased wider in the near term, but should end 1H12 close to current levels Thanks to sticky front-end Treasury yields, short-expiry swaptions on short tails are effectively like options on front-end spreads. We recommend utilizing 1Mx1Y payer / receiver swaptions to initiate asymmetric exposure to front-end spread widening / narrowing Fed purchases of longer-end Treasuries will offer opportunities in 1H12trade the impact of intramonth swings in purchase pace on swap spreads, and look to trade maturity-matched swap spread switches based on the Feds metric of value Initiate steepeners between intermediates and the long end hedged with Eurodollars, to position for yield curve normalization as well as carry Look to initiate positive carry, belly richening butterflies as attractive ways to gain empirically convex exposure to higher front-end yields Yield curves between the front end and intermediates will remain directional in 2012, but conditional curve trades are not cheapinitiate synthetic conditional curve trades, created by replacing swaptions at the front end with YCSOs Approach 2012 with a long gamma bias, but look for 3Yx10Y to decline to 6bp/day by 1Q12 end Bermudan receiver swaptions offer directional vega exposure, as well as exposure to cheap implied correlation and forward volatility. We introduce a novel approach to measure relative value in Bermudan swaptions, based on using approximating Canaries
Exhibit 1: Swap spreads are near the upper end of a relatively wide trading range in 2011
2011 YTD statistics Average High Low 23.7 53.2 12.6 28.2 50.1 20.4 25.0 44.3 16.2 20.0 34.1 10.5 11.5 22.7 3.9 -26.3 -19.8 -39.0
Statistics regarding maturity matched swap spreads in various benchmark sectors; bp
2Y 3Y 5Y 7Y 10Y 30Y
Current 53.2 50.1 44.3 34.1 19.3 -27.3
Start 18.1 23.9 15.9 9.6 5.5 -21.6
Range 40.6 29.7 28.1 23.5 18.8 19.2
* As of 11/17/2011
Exhibit 2: Collateralization rates have been on the rise across all OTC derivative markets
Percent of trade volume subject to credit support agreements
All OTC derivatives Fixed Income derivatives Credit derivatives FX derivatives Equity derivatives Commodities 2003 2004 2005 2006 2007 2008 2009 2010 2011 30 51 56 59 59 63 65 70 70 53 58 58 57 62 68 63 79 79 30 45 59 70 66 74 71 93 93 21 24 32 37 36 44 36 57 58 27 45 45 46 51 52 52 71 72 17 25 33 44 40 39 39 62 60
Source: ISDA Margin Survey
Swaps
Swap spreads across the curve traded in significantly wide ranges over the past year, and have recently widened to the upper end of those ranges (Exhibit 1). The move wider late in the year has been mainly due to the steadily deteriorating sovereign debt crisis in Europe, which has now spread well beyond its origins in Greece. The resulting moves in swap spreads have been reminiscent of 2008, although trading ranges this year remain small in comparison; 10-year spreads, for instance, traded in a 75bp range in 2008. The volatility in swap spreads is a reflection of the fact that spreads have been influenced by numerous factors in significant ways. Intermediate spreads narrowed going into the second quarter as markets priced in the end of QE2, but worsening conditions in Europe steadily have pressured spreads wider since then; more recently, the insufficient nature of policy measures announced after the recent EU summit has sparked renewed widening in spreads, bringing them closer to the wides of the year in the 2- to 10-year sector of the curve. In short, spreads have been largely driven by the Feds Treasury purchases and by the European banking system
131
crisisfactors that swap market participants would not normally view as endogenous. Indeed, endogenous factors that have historically exerted considerable influence on swap spreadsmortgage market hedging flows and the swapping of corporate issuance, as well as longer-term drivers such as budget deficit expectations and funds rate expectationshave played a less significant role this year. With deficit expectations remaining largely stable and with the Fed on hold for several years, longer-term drivers have not experienced enough volatility to influence swap spreads in a significant way. Also, the numerous well-understood hurdles to refinancing mortgages have made mortgages ever less negatively convex; this fact, together with falling portfolios at the two major GSEs, has largely helped to explain the decline of mortgage-hedging flows as a driver of swap spreads. Somewhat less permanently, low and sticky yield levels appear to have lessened the imperative to swap fixed rate debt issuance on the part of many high grade corporate issuers, including those who traditionally swap their issuance. Such behavioral shifts on the part of market participants are only one among numerous changes impacting the swaps market. Another key transformative development is of course the ongoing implementation of the DoddFrank reforms, as they pertain to the derivatives markets. Much uncertainty remains in this regard, and we discuss the current status of regulatory implementation; however, one key change that has already taken place in the swaps market is a move towards OIS discounting as a quasistandard for valuing off-market swaps. To be sure, this is not exactly a new development, and fixed income derivatives markets have led the broader trend towards higher levels of collateralization in recent years (Exhibit 2). Given this trend, the move towards OIS discounting (which is arguably closer to the funding rates for commonly used collateral) has been in the making for sometime now. Nonetheless, there is growing evidence that markets have gravitated towards value measures that are more consistent with OIS discounting. For instance, par asset swap spreads for premium/discount bonds appear to serve as better signals of value when computed using OIS discounting, as opposed to Libor discounting. In other words, swap market participants will need to adjust to a world where the reference curve (with respect to which asset valuations are measured) and the discount curve are different; we discuss this further in a later section.
Exhibit 3: No convexity to hedgethe magnitude of convexity in current coupon mortgages is near historical lows
Absolute value of FNMA 30-year current coupon TBA Libor OA-convexity;
3.0 2.5 2.0 1.5 1.0 Current 0.5 0.0 Feb 00
Oct 02
Jul 05
Apr 08
Jan 11
Exhibit 4: The declining footprint of mortgage hedgers in the swaps market is also due to mandated reductions in the GSEs retained portfolios
Total outstanding swaps notional amount held by Fannie Mae and Freddie Mac, versus their total retained portfolio size; $bn $bn
2000 1800 1600 1400 1200 1000 800 Dec 06 Swap notional
1700 1650 1600 1550 1500 1450 1400 May 08 Sep 09 Feb 11
Retained portfolio
Swap spread drivers: the new normal As we consider the likely outlook for intermediate maturity swap spreads over the course of next year, it is just as important to recognize what will not be driving swap spreads. These include several traditional and longstanding factors that influenced swap spreads in the past, such as (most notably) budget deficit expectations. To be sure, should the path of fiscal policy deviate from expectations in a sufficiently large way, budget deficit expectations will undoubtedly re-emerge as a key driver
132
A tactical model for benchmark 10-year maturity matched swap spreads

It is no exaggeration to say that the world of swap spreads is experiencing its own new normal. Traditional, long-term drivers have simply become less volatile and more range-boundFed funds expectations are virtually assured to remain stable for a long period of time, and budget deficit expectations have become quite stable. Traditional tactical factors have become less relevant too; declining portfolio sizes at the GSEs, and structural impediments to refinancing mortgages have made mortgage-hedging flows relatively insignificant in terms of driving swap spreads. In addition, anecdotal evidence points to reduced swapping of debt issuance, thanks to low outright yield levels. Long-term structural factors remain relevant, and budget deficit expectations could still prove significant in 1H12 if fiscal policy were to deviate materially from current expectations. That said, trading swap spreads in the 10-year sector now requires a different, more tactical model to assess fair value, which is outlined here. This model is more short term in nature out of necessity, given the recent nature of the developments discussed above. The factors incorporated in our model are (i) 6-month FRA-OIS spreads (to capture bank funding pressures and Libor expectations), (ii) the bank stock index (a broader measure of financial system health), (iii) weekly open market gross purchases (i.e., not net of sales at the front end) of Treasuries by the Fed (which matters for swap spreads even under a balance sheet neutral construct as with Operation Twist), and (iv) weekly swapped corporate issuance. The first three factors are relatively straightforward; the lastswapped issuancehas been more interesting, because low-yield levels have likely deterred the amount of fixed rate issuance that has been swapped this year. Understanding the aggregate swapping activity of high grade debt issuers is difficult given sketchy data, but necessary nonetheless since moves in yields could alter swapping behavior by material amounts. Therefore, here we attempt to estimate the fraction of high grade issuance that gets swapped as a function of rates. Our approach is as follows. For a representative sample of non-financial companies over a 3-year period, we examined data on the aggregate amount of fixed-rate issuance (as a percentage of overall fixed rate issuance) that was swapped to floating, using information from financial statements. As seen in Exhibit A1, the fraction of overall non-financial corporate debt issuance that gets swapped to floating has indeed been sensitive to yield levels, falling sharply as 10-year swap yields fall below 3%. Also interesting is that the data suggests that swapping activity may rise again in very low yield environments. While the data in these ultra-low yield regimes is admittedly limited, it stands to reason that swapping activity might begin to rise again as yields fall below a second threshold, perhaps reflecting expectations of the Fed being on hold for a very long time. Unfortunately, such data cannot be compiled for financial issuers; it is nonetheless important to account for potential changes in their swapping behavior too, given that these institutions are both large issuers and also tend to swap considerable fractions of their issuance. Moreover, anecdotal evidence suggests that these issuers too have curtailed their swapping activity as yields have fallen. To estimate their swapping behavior as a function of yields, we take the leap of faith that the fraction of financial issuance that is swapped may be modeled as a linear function of the corporate issuance swap-fraction. That is, we assume that financial issuance fractions are a similar non-linear function of yields, but could differ in magnitude since financials typically tend to swap a greater portion of their fixed-rate debt to floating. We estimate this function to be consistent with broad anecdotal evidence regarding the percentage of financial issuance that gets swapped. To be specific, our linear transformation causes financial issuance swapping percentages to fall to 30% if corporate swapping fractions go to zero, and financial issuance swapping fractions go to 100% when corporate swapping fractions go to 100%. Exhibit A2 presents a time series of our estimate for total swapped monthly high grade issuance, calculated by multiplying actual monthly financial and nonfinancial issuance by our estimate for swapping fractions (based on yield levels at the time of issuance). Armed with an approach to estimate issuers debt swapping percentages, we can now present our model for benchmark 10-year maturity matched swap spreads. This new, high-frequency model for swap spreads is detailed in Exhibit A3. As a rule of thumb, each 10bp of FRA-OIS widening implies a 2.5bp widening in 10-year spreads; $10bn in weekly Treasury gross purchases is worth 1bp of widening in spreads; $10bn in swapped high grade issuance (estimated per our framework) would narrow swap spreads by 2.5bp; and a 10 point rise in the BIX bank stock index would cause swap spreads to narrow 1.3bp.
Exhibit A1: Issuance swapping fraction vs. yields
25%
50
Exhibit A2: Estimated swapped issuance
Exhibit A3: A model for 10-year swap spreads
20% 15% 10% 5% 0% 2.0 2.5 3.0 Low yield periods
High yield periods
40 30
Overall fit
20 10 0 Nov 08
3.5
4.0
May 09
Dec 09
Jul 10
Jan 11
Aug 11
Factor Coefficient T-stat Current Intercept 20.4 6.3 FRA-OIS 0.26 10.2 72.3 Wkly Treasury purchases by Fed 0.10 4.6 10 Swapped HG issuance -0.25 -4.1 1.7 BIX -0.13 -6.5 119.0 R^2 73% Projected 10Y swap spread 23.9 Actual 19.3
of swap spreads. In this regard, it is still possible (but perhaps not likely) that fiscal policy adjustments will be enacted sometime next year; despite the initial failure of
the Congressional super committee, mandatory cuts do not take effect until 2013, and thus significant time remains for Congressional action on fiscal policy.
133
However, barring such a surprise, the fact remains that 1year ahead budget deficit expectations have remained (and are likely to remain) very stable, moving in a $200bn range. As a result, assuming there are no unforeseen developments on the fiscal policy front, deficit expectations will likely not particularly impact 10year swap spreads in 1H12. Similarly, given a historically low level of negative convexity in mortgages (Exhibit 3), mortgage hedging flows have not been, and will likely not be, a significant driver of spreads in 2012. The footprint of mortgage hedgers in the swaps market is also declining for a much more obvious reason mandated caps on the retained portfolios of Fannie Mae and Freddie Mac point to a decline in the swap hedging activity of these hedgers (Exhibit 4), who have traditionally been sizeable players in the swaps market. In order to reflect this altered market environment that has now become the new normal, we have augmented our more traditional longer-term benchmark 10-year swap spread model with a considerably revised tactical model. Not only is this model more short term in nature (out of necessity given the recent nature of the developments discussed above), but it also addresses newly emergent tactical forces impacting swap spreads Fed purchases of Treasuries at the longer end, behavioral shifts with respect to swapping of debt issuance, and of course the European crisis. A detailed description of this tactical model is included in the grey box on benchmark 10-year swap spreads. Swap spread trading themes in 2012 Such a model has several implications for swap spreads looking ahead into 1H12. First and foremost, our outlook on the various drivers points to widening in intermediate swap spreads heading into year-end and in 1Q12, and an eventual narrowing (to slightly below current levels) by mid-year 2012; this is laid out in our swap spread forecast in a later section. Second, swap spreadseven in the 10-year sectorwill likely remain quite vulnerable to exogenous risk from the European crisis. Exhibit 5 presents an estimate of the partial impact on 10-year swap spreads, if each of the drivers moves by an amount equal to one standard deviation of rolling 3-month changes over the past year. Not only are FRA-OIS spreads the single most important driver of intermediate spreadsas noted in Exhibit 5, but they are also significantly inversely correlated to the BIX (whose partial beta is negative), implying that the two
Exhibit 5: Ranking swap spread driversFRA-OIS spreads and bank equity prices are likely to be the dominant drivers of swap spreads
Partial impact of a 1 standard deviation move* in each of the drivers in our model for benchmark 10-year maturity matched swap spreads; bp
3 2 1 0 -1 -2 -3 FRA-OIS Fed purchases Corp. issuance BIX * Estimated as the standard deviation of rolling 3-month changes over the past year. Impact estimated as the 1 SD move in each driver multiplied by its beta in our model.
Exhibit 6: The weekly pace of Treasury purchases by the Fed can exhibit significant tactical variability, even as it remains stable on average during Operation Twist
Weekly gross purchase of Treasuries by the Fed; $bn
15
10
most significant drivers of swap spreads are likely to move in correlated fashion, amplifying the effect of European developments on swap spreads. Hedging swap spread positions with 35% risk (rather than the 25% partial beta) in FRA-OIS spread positions would help mitigate the effective risk exposure to the European crisis. Third, although the Feds Treasury purchases are likely to be stable on average (with Operation Twist scheduled to continue through the end of June 2012, and with QE3 unlikely to include Treasuries until after Operation Twist is concluded), the actual schedule of Treasury purchases
134
Exhibit 7: which should make for tactical trading opportunities, since spreads have tended to narrow in periods when the Feds purchase pace declines sharply, and widen when the opposite is true
Cumulative change in 10-year maturity matched swap spread, net of a 35% riskweighted 6-month forward FRA-OIS hedge, averaged in periods when the weekly pace of Treasury purchases by the Fed fell sharply* and rose sharply**; bp
Exhibit 8: Our estimate of swapped high grade issuance also exhibits considerable intra-year seasonal patterns, with the FebMay period seeing the biggest rise in pace
Weekly swapped high grade issuance*, averaged by month over the past five years; $bn
3 2 1 0 -1 -2 -3 -4 -6 -4 -2 0 2 4 # business days around selected date * Dates used are 1/5/11, 7/15/11 and 11/2/11. ** Dates used are 11/18/10, 1/20/11 and 5/11/11. -10 -8 Decline in pace Rise in pace
9 8 7 6 5 4 3 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec * J.P. Morgan estimate, based on a model for the fraction of financial and nonfinancial debt issuance that is swapped
Exhibit 9: causing swap spreads to be biased narrower in that period on average

Cumulative change in 10-year maturity matched swap spread adjusted for 6-month forward FRA-OIS in the 2-month period from the first business day of March, averaged over the past four years; bp
by the Fed will likely create attractive tactical trading opportunities in swap spreads. It is interesting to note that there can be considerable variation in the pace of gross Treasury purchases on a week-to-week basis; for instance, the weekly pace of gross purchases declined from a mid-October high of $14bn to under $5bn in just two weeks (Exhibit 6). Such tactical swings in the weekly pace of Fed purchases tend to impact swap spreads (adjusted for FRA-OIS spreads). To illustrate this tactical effect, we identified the three dates that saw the biggest 2-week upward change in rolling weekly purchases, as well as the three dates that saw the biggest downward change. As seen in Exhibit 7, spreads (adjusted for FRA-OIS) narrowed in the periods that saw sharp declines in the pace of buying, while periods that saw a sharp pickup in the pace of purchases were characterized by a widening in swap spreads. Given that the Fed will be publishing schedules at the end of each month for its purchases in the following month, shifts in pace will be known ahead of time, and this tactical effect can therefore be traded. Fourth, barring considerable disruptions in primary issuance markets due to the European crisis, typical intrayear seasonal patterns in issuance will likely offer attractive opportunities as well. Over the past five years, high grade swapped issuance has tended to be far below average in the months of February and July, while reaching highs in the months of May and September
-5
-10
-15 0 5 10 15 20 25 30 35 Business days after the beginning of March 40
(Exhibit 8). The best way to position for this intra-year seasonal pattern is to initiate swap spread narrowers in early March. Benchmark 10-year swap spreads, adjusted for FRA-OIS, have narrowed significantly in the March/April period in each of the past four years, and narrowed by a modest amount in 2007. Given the significant narrowing in spreads in this period on average (Exhibit 9), we would look to initiate spread narrowers in early March.
135
Modeling the duration exposure in variable annuities and its impact on 30-year swap spreads
Since the start of QE2 in 4Q10, long-end swap spreads have been mainly driven by three factorsthe slope of the 2s/30s Treasury curve, the Feds gross purchases of Treasuries, and the shifting duration of the variable annuity universe. Estimating the last of these three factors is itself a challenge. Pricing variable annuities (henceforth referred to as VAs) is an exceedingly complex undertaking, requiring the joint modeling of long-term interest rates as well as equities; in addition, actuarial risks stemming from life insurance related guarantees, and other features make the product path dependent, adding to the complexity. It is nonetheless important to capture the nonlinear dependencies of VA duration exposure with respect to equities and long-term swap yields, since VAs have been a popular product for insurance companies for many years with over $1.5tn estimated to be outstanding, and the interest rate in these products can be highly nonlinear. A common product is to pay policy holders a variable return based on the performance of the S&P 500 (or another benchmark) but provide downside protection by guaranteeing a minimum income stream for a specified time period. In rising equity markets, the product creates little risk to insurance companies, since equity returns are merely passed through to policyholders. However, in falling equity markets, the minimum guaranteed income stream becomes more binding. Effectively, as the moneyness of the embedded equity put increases, insurance companies increasingly become short a fixed income annuity, requiring them to add long duration hedges as a result. Recognizing that our objective is not to price variable annuities accurately, but merely to capture the trend in their duration exposures as well as their nonlinear relationship with equities/yields, we devise a simpler approach. Our approximation approach is based on three principles. First, we start with the assumption that the duration of the VA universe may be approximated by a weighted combination of the durations of a set of much simpler VA lite instruments, which we refer to as VA kernels. This is not unlike series approximation techniques commonly used in mathematics to solve difficult problems. One example of a VA kernel is a simple product where a policyholder pays $100 on (say) January 1, 2007, intended to be invested in the equity market for a 10-year (fixed) horizon, with a guaranteed withdrawal amount of $5 per year for the subsequent 20 years. Several VA kernels may be created by varying the start date, the length of time of the intended equity investment, and the minimum guarantee amount. Effectively, we price the complex VA universe as a linear combination of simpler VA kernels, each of which is priced in a manner that captures the nonlinearities with rates and equities. Second, we price the present value of this instrument by ignoring actuarial risks and using an option pricing framework. We use implied distributions from the swaptions market as well as long-term S&P vols and correlation estimates to calculate the price. We also use numerical tweaks to calculate the partial exposure with respect to long-term swap rates (i.e., duration). Third, we use a calibration approach to solve for the appropriate weights on the various VA kernels. Our calibration relies on the anecdotally known fact that VA risk exposures were significant influences on long-end swap spreads in certain periods of time, such as 4Q08; we may thus solve for non-negative coefficients that maximally explain the portion of long-end swap spread behavior not explained by other factors in those select periods of time. In order to mitigate circularity (since we plan to use VA duration estimates to model long-end swap spreads), no data after 2008 has been used in calibration, and out-of-sample performance has been tested and found to be reasonable. The aggregate duration of the VA universe is shown in Exhibit B1; as can be seen this remains near historical highs and could worsen if yields and equities were to decline in 1Q12 on the back of continuing deterioration in Europe. Armed with this estimate for VA duration, we may now model 30-year swap spreads as a function of the three factors outlined above. Exhibit B2 presents the statistics from regressing 30-year benchmark maturity matched swap spreads versus the 2s/30s curve, cumulative gross purchases of Treasuries by the Fed (i.e., not net of front-end sales, and since the start of QE1), and the duration of the VA universe in 20-year equivalents, over the past 15 months (i.e., since QE2 expectations became a significant factor, resulting in altered dynamics at the long end of the Treasury curve). As a rule of thumb, each 10bp flattening in the 2s/30s Treasury curve would widen 30-year spreads by 2.8bp, a fall in VA duration by $10bn 20-year equivalents would cause a 5bp widening in spreads, and the cumulative effect of the Feds long-end purchases under Operation Twist ($400bn in all) should be about 3.6bp (not including indirect effects due to the impact of such purchases on the curve). As seen in Exhibit B3, this model has been successful in fitting the observed behavior of 30-year swap spreads. Exhibit B1: Estimated duration of VA universe Exhibit B2: 30-year swap spread model
275 250 225 200 175 150 125 100 Jan 07 May 08 Sep 09 Feb 11
Exhibit B3: Actual spreads versus model fair value

-15 -20 -25 30-year maturity matched swap spread; bp Model
Factor Intercept 2s/30s Treasury curve, % VA aggregate duration; $bn 20s Cum gross Tsy purchases by Fed; $bn R^2 Projected 30Y swap spread Actual
Coefficient T-stat 160.7 18.4 -27.7 -20.7 -0.5 -24.1 0.009 17.0 74% -25.2 -27.1
Current 2.897 240.4 1200.5
-30 -35 -40 -45 Nov 10 May 11 Nov 11
136
Exhibit 10: The partial sensitivity of long-end swap spreads to VA duration has indeed been declining, but reports of its demise (as a factor impacting spreads) are as yet exaggerated
Exhibit 11: Appearances can be deceptivethe seemingly wrong-way correlation between VA duration and long end swap spreads
30-year maturity matched swap spread versus estimated VA duration* $bn 20-year equivalents bp
Rolling partial beta of 30-year maturity matched swap spreads versus VA duration*; bp per $bn 20-year equivalents
260 255
-18 -20 -22 -24 -26 -28 -30 -32 VA duration 30-year swap spread -34 -36 -38 Sep 11 Oct 11 Nov 11
-0.45 -0.50 -0.55 -0.60 -0.65 -0.70 -0.75 Feb 11 May 11 Aug 11 Nov 11 VA beta
250 245 240 235 230 225 220
* J.P. Morgan estimate
* Based on rolling 15-month regressions of 30-year benchmark maturity matched swap spreads versus VA duration, the 2s/30s Treasury curve and cumulative gross Treasury purchases by the Fed. Regressions are over 15 months because our current model for 30-year swap spreads is designed to cover the period from August 2010, when QE2 became a significant factor affecting long end spreads.
Long end swap spreads: does VA hedging still matter? Swap spreads at the long end of the curve are broadly at the mercy of two large technical forcesFed purchases of Treasuries (particularly at the long end of the curve), and the shifting duration hedging needs of life insurance companies variable annuity portfolios (VA duration, for short). Although the significance of the latter has been diminishing, as insurance companies broaden their use of Treasury-based derivatives and/or cash Treasuries, we believe it is too soon to dismiss VA duration needs as a driver of long-end swap spreads. Estimating the duration of the VA universe is a complex undertaking, and we have outlined an approximate scheme for doing this in previous research (see Interest Rate Risk in Variable Annuities, J.P. Morgan Research Note, September 28, 2011). A short summary of our approach to estimating VA duration, as well as its use in our revised model for 30-year benchmark maturity matched swap spreads, is presented in the grey box on long-end swap spreads. Several points are interesting to note with respect to our long-end spread model. First, the partial sensitivity of long-end spreads to VA duration has indeed been declining in magnitude (Exhibit 10). That said, reports of its demise as a factor impacting long-end spreads are
Exhibit 12: masks a more subtle messagethe recent strong equityrate correlation has caused the 2s/30s curve to become negatively correlated with VA duration, producing offsetting impacts on long-end spreads
Rolling 3-month correlation between weekly changes in the S&P and 30-year swap yields, versus the effective beta* of 30-year swap spreads versus VA duration; bp per $bn 20-year equivalents
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 May 11
S&P/30Y yield correlation Effective VA beta
0.05 0.00 -0.05 -0.10 -0.15 -0.20 -0.25 -0.30 -0.35 -0.40 -0.45 -0.50 Nov 11
Jul 11
Sep 11
* Calculated as partial beta of 30-year spreads with respect to VA duration (as detailed in the footnote to exhibit 10), plus the partial beta of long end spreads with respect to the 2s/30s curve (from the J.P. Morgan 30-year spread model) times the recent beta of the 2s/30s Treasury curve with respect to VA duration. Recent beta calculated over rolling 3-month periods.
exaggerated, at least for now, and it is equally important not to overstate the extent of this decline. To all appearances, it would seem that VA hedging has become insignificant, or even a contrarian indicator of long-end swap spreads. Indeed, in recent months, long-end spreads have widened as VA duration has risen (Exhibit 11). However,
137
appearances can be deceiving, and this masks a more subtle message, which is that in recent months VA duration and the 2s/30s curve have become much more negatively correlated than before, causing their impacts to offset each other; moreover, this negative correlation is itself a consequence of the elevated positive correlation between long-end yields and the S&P. In other words, increased equity-rate correlation has caused VA duration to rise sharply as long-end yields fall and the 2s/30s curve flattens. As a result, the partial impact of increased VA duration needs has been mitigated by the counterdirectional exposure of long-end spreads to the curve (Exhibit 12). Such correlations could persist for some time, producing limited net impact on long-end swap spreads. However, this should be seen as a happy coincidence, and is unlikely to persist in the medium term. Should equities and rates decouple in 1H12, as indeed we expect them to, such a canceling effect will be less pronounced, and the impact of VA duration on long-end spreads will likely be apparent once again. Therefore, we continue to rely on a fair-value framework for long end spreads that includes VA duration as well as the 2s/30s curve. Front-end spreads At the front end of the curve, swap spreads have become pure plays on the European crisis, exhibiting strong correlation to FRA-OIS spreads but little else. This will likely remain the case in 1H12, and opportunities in this sector will likely revolve around trading the mispricing in front-end spreads relative to FRA-OIS on a hedged basis. Thus, a view on FRA-OIS is essential to determine the likely evolution on front-end swap spreads in 2012. To this end, we present a simple model that attempts to explain the behavior of FRA-OIS differentials in recent months. USD FRA-OIS spreads have been driven by two factors recentlythe EUR/USD OIS FX basis, and semiperipheral European sovereign CDS spreads. (Although the commonly quoted EUR/USD FX basis is with respect to a Euribor/Libor basis swap, it is more useful to adjust this quoted spread to be with respect to an EONIA/OIS basis swap. We do this because (i) EONIA and OIS rates are more reflective of the marginal cost of funds for banks in the Eurozone and the US, and (ii) the existence of the Feds dollar swap lines at a 100bp penalty bounds this EUR/USD OIS FX basis at -100bp. For a more
Exhibit 13: Our model for FRA-OIS spreads suggests that FRA-OIS could widen into year-end, but narrow going into mid-year
Statistics from regressing 3-month forward constant maturity FRA-OIS differential (bp) versus the 3-month EUR/USD FX OIS basis (bp) and semi-peripheral sovereign CDS spreads* (bp), and projections for FRA-OIS at end of 4Q11 and 1H12 based on assumptions for the drivers.
1H12 Factor Coefficient T-stat Current 1Q12 Intercept 4.6 3.8 EUR/USD FX OIS basis -0.43 -11.9 -73 -100 -40 Semiperipheral spreads 0.06 7.7 406 450 475 R^2 95% FRA-OIS fair value 60 74 50 Actual 63 * Average of Italy, France and Spain 5Y CDS spreads. Regression based on 6 months of history, and are as of COB 11/18/2011. Shaded values represent projections.
Exhibit 14: Swap spread forecast
J.P. Morgan projections for maturity matched swap spreads in various benchmark sectors; bp
2Y swap spread 5Y swap spread 10Y swap spread 30Y swap spread
Current 49 40 17 -29
1Q12 53 45 27 -23
1H12 38 34 18 -23
* Current values as of COB 11/18/2011. Projections assume 3-month forward FRAOIS spreads evolve according to the trajectory shown in the previous exhibit.
detailed discussion of cross-currency basis swaps, see Decoding the FX basis market for signs of $ funding stress, Pavan Wadhwa et al, August 23, 2011. A more negative basis implies a higher USD borrowing rate at which a European borrower would be indifferent to borrowing in USD versus borrowing in euros and swapping to USD, and may thus be thought of as a fundamental funding market factor. We also use semiperipheral spreads as a second factor to capture the broad correlations between credit spreads and metrics of European stress. Exhibit 13 presents statistics regarding such a model, and also uses the model parameters to project 3-month forward FRA-OIS spreads by year end and by mid-year 2012. To make these projections, we are assuming that in the near term, conditions deteriorate further into year end and in 1Q12, and the EUR/USD OIS FX basis reaches 100bp, which should act as a near-term floor thanks to unlimited central bank funding. We also make projections for European stress metrics based on estimates from our European rates strategists, and look for a near-term widening in our measure of semi-
138
peripheral spreads. Finally, we assume that policy actions to be enacted sometime in 1H12 will include concerted actions by central banks (alluded to by Rosengren last week) such as dropping the penalty on the Feds dollar swap lines to 50bp from 100bp. Under such a scenario, FRA-OIS spreads may be expected to widen to nearly 75bp in the near term, but narrow back to 50bp by midyear 2012 despite a steady widening in semi-peripheral spreads. Given these projections for FRA-OIS spreads, which is a key determinant of spreads in the 2- to 10-year sector, and based on our projections for other inputs, our estimates for maturity matched swap spreads in various sectors as of the end of 1Q12 and 1H12 are shown in Exhibit 14. Broadly speaking, we see the potential for swap spreads to widen going into 1Q12; however, since front-end spreads are generally wide to fair value currently, this upside is greatest in the 10-year sector. Heading into 2Q12, we would look for swap spreads to narrow, a view that is premised upon favorable policy actions in Europe as discussed earlier. In the 30-year sector, we expect spreads to widen initially and then remain largely range-bound. Front-end spread options? The correlation between front-end swap spreads and FRA-OIS spreads is a reflection of a broader phenomenon at the front end of the yield curve. Subsequent to the Fed announcing its commitment to low rates until at least mid-2013, OIS rates and Treasury yields have fallen and the OIS curve has become both flat and sticky; as a result, the volatility in front-end swap spreads is now almost entirely due to the volatility of the swap rate itself. Put differently, bearish positions at the front end are effectively the same as front-end swap spread wideners. This is strikingly evident in Exhibit 15, which shows that 2-, 3- and 5-year swap spreads have all become highly correlated to front-end swap rates, with little differences between the three sectors. This is quite usefuloptions on front-end yields do exist, and are effectively options on front-end swap spreads. Short expiry payer swaptions and/or payer swaption spreads on front-end swap yields (such as 3Mx1Y payer swaptions, for instance) can be used to create asymmetric exposure to wider front-end swap spreads. As we have noted often in recent weeks, given their limited-risk nature, such trades can offer better risk-reward than outright swap spread wideners or narrowers. Indeed, in the past month, we have recommended payer swaption
Exhibit 15: Swap spreads across much of the front end have become highly correlated to front-end swap yields with little distinction across sectors
Maturity-matched swap spreads in 2-, 3- and 5-year sectors versus the 3Mx1Y forward swap yield; past three months; bp
45 40 35 30 25 20 0.4 0.45 0.5 0.55 0.6 3Mx1Y forward swap yield; % 0.65 0.7 2Y 5Y 3Y
Exhibit 16: Sectors where the Fed is actively buying Treasuries are likely to see a convergence in yields between issues of similar duration
Dispersion* of bond yields in the 3- to 6-year sector (where the Fed is not buying) and the 20- to 30-year sector (where the Fed is buying); bp
1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 Sep 11 Oct 11 Nov 11 3- to 6-year 20- to 30-year
* Calculated as the root mean squared yield error of bonds in that sector relative to an overall par Treasury fitted curve
spreads as proxies for front-end swap spread wideners, as well as receiver swaption spreads in place of spread narrowers (see US Fixed Income Markets Weekly, Interest Rate Derivatives, dated September 23, 2011 and November 18, 2011, respectively). This state of affairs is likely to persist into next year. Going into 2012, a key part of our strategy with respect to front-end swap spreads will be to trade them using payer and receiver swaption spreads as proxies.
139
O-spreads and spread curve relative value With Operation Twist already underway and slated to continue until the end of 1H12, relative value spread switch trades will likely prove to be a profitable trading strategy as we head into 2012. The logic underlying this is quite simple; experience from the Feds Treasury purchases under QE1 as well as QE2 suggest that the Feds issue selections are based on yield differentials of various bonds with respect to a fitted curvei.e., yield errors (see Treasuries). Thus, it is reasonable to expect convergence between rich and cheap issues, measured through metrics that mirror the Feds, particularly in sectors where the Fed is actively purchasing Treasuries. Such convergence is indeed already evident since the commencing of Operation Twist; yield errors are converging at the longer end of the curve, producing little dispersion across bonds in those sectors, while a similar fall in dispersion is not evident in the shorter end (Exhibit 16). Of course, yield errors are not directly tradable, but switch trades on a maturity matched swap spread basis are a close proxy. Such trades should prove attractive in periods and sectors characterized by active open market Treasury purchases by the Fed. But what relative value metrics are suitable for sectors of the curve or periods of time where the Fed is not actively purchasing Treasuries? Historically, asset swap spreads (some times called a proceeds asset swap spread) have served as a useful metric of value. The logic behind this measure is this: suppose that an investor puts up $100 to purchase a Treasury bond regardless of its actual full price, finances (or lends) the difference between the bonds full price and par via a swap where the investor pays the bonds
Exhibit 17: An illustration of the proceeds asset swap spread and Ospread
A schematic illustration of the proceeds asset swap spread for a given bond
Dirty price Investor
Bond Bond coupons Swap counterparty 3M Libor + spread
Dirty price minus par
For bond: 8% Nov 2021 Spread: -12.9bp under Libor discounting -17.4bp under OIS discounting
coupon cash flows versus receiving Libor plus a spread on the floating leg. This spread is called the proceeds asset swap spread (or simply, asset swap spread), and is a clean metric for comparing bonds with different coupons and prices (see Exhibit 17 for an illustration). Traditionally, these asset swap spreads were solved for by requiring that cash flows from the associated swap, discounted using swap curve zero rates, should produce an NPV that offsets the amount being financed (or lent). That is, the swap curve served as both the reference curve as well as the discount curve. In principle, however, the discount curve should depend on the funding curve for the type of assets that would be used to collateralize the mark-to-market variations in the swap. Since US dollar interest rate swaps are commonly
Exhibit 18: Might is rightthe Feds preferred metric displaces the markets value metric in sectors and times where the Fed is actively buying Treasuries, but O-spreads are likely to be the best value metric otherwise
Statistics regarding the back-testing of trading strategies* using three metrics of value: yield error, asset swap spreads (based on swap curve discounting) and o-spreads (or asset swap spreads to the swap curve, calculated using OIS discounting)
2- to 5-year sector 7- to 15-year sector Value metric: Yield error Value metric: ASW Value metric: OSW Value metric: Yield error Value metric: ASW Value metric: OSW During Jul - Oper. During Jul - Oper. During Jul - Oper. During Jul - Oper. During Jul - Oper. During Jul - Oper. Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist 230 146 64 20 230 146 64 20 230 146 64 20 230 146 64 20 230 146 64 20 230 146 64 20 202 135 53 14 196 119 60 17 208 125 63 20 139 103 18 18 137 86 43 8 139 87 42 10 28 11 11 6 34 27 4 3 22 21 1 0 91 43 46 2 93 60 21 12 91 59 22 10 88% 92% 83% 70% 85% 82% 94% 85% 90% 86% 98% 100% 60% 71% 28% 90% 60% 59% 67% 40% 60% 60% 66% 50% 4.2 4.2 4.7 2 4 3.7 4.9 3.1 4.7 3.9 6.6 3.1 2.5 2.9 1.5 1.5 3.7 3.2 5.1 1.2 4 3.8 5.2 0.9 -2.4 -0.7 -4.7 -1.3 -1.2 -1.4 -0.7 -0.5 -1.1 -1.1 -0.8 N/A -1.6 -1.6 -1.6 -0.4 -3.1 -3.2 -3.6 -1.9 -3.5 -3.4 -4.2 -2.6 3.4 3.8 3.1 1 3.2 2.7 4.5 2.5 4.1 3.2 6.5 3.1 0.9 1.5 -0.7 1.3 0.9 0.6 2.2 -0.7 1 0.9 1.9 -0.8
# trades # wins # losers Hit ratio Avg Gain Avg Loss Overall avg
* Trading strategy is as follows: every day, the cheapest and richest bonds in each sector are identified, based on each of the three value metrics. The rich bond is then sold and the cheap bond is bought, on a maturity matched swap spread switch basis. Trades are held for a month and unwound thereafter. We back-tested these strategies from the start of QE2 through mid-November.
140
collateralized by Treasuries, it would be more correct to use OIS-curve based zero rates to discount future cash flows (since Treasuries fund at GC rates, which are rather close to OIS rates, and since a well developed OIS curve exists). With regulatory reform forcing almost all swaps to be collateralized going forward, a more correct way to define a proceeds asset swap spread would be to still define a spread versus the Libor curve (as in the previous case), but using the OIS curve for discounting cash flows. We introduce new terminology to reference this new measure of value, and will refer to it as proceeds asset swap o-spread (or simply, o-spread) heretofore in our research. O-spreads will be understood as a reference to spreads on the floating leg of a swap, where the spread has been solved for using OIS zero rates for discounting purposes. Which of these metrics works well in practice? To answer this question, we back-tested trading strategies based on three different value metrics, in two sectors of the curve, over the past year. Every day, we identified the richest and cheapest bonds in each of the two sectors, based on each of three value metrics (yield error, asset swap spread and o-spread). We then initiated a long in the cheap bond and a short in the rich bond, but on a maturity-matched swap spread basis (for the purpose of lessening P/L noise due to curve movements). All trades were held for a month and unwound thereafter. The two curve sectors we chose are the 7- to 15-year sector (a beneficiary of Fed purchases during QE2 as well as Operation Twist), and the 2- to 5-year sector (which benefited from Fed purchases during QE2, but not during Operation Twist). Statistics regarding our results are presented in Exhibit 18. Three points are interesting to note from our results. First, yield errors are the best metric for sectors/periods where Fed purchases are a factor. Put differently, Fed purchases distort normal value metrics, and essentially impose the Feds value metric in its place. This is evident from the fact that hit ratios were highest in both curve sectors during QE2 when using yield errors as the value metric. More recently under Operation Twist, using yield errors once again results in the highest hit ratios in the 7to 15-year sector. Second, when Fed purchases cease to be a factor, the resulting distortion effect on value metrics is quick to dissipate. This is seen in the fact that during the brief 3month hiatus in Fed purchases in the 7- to 15-year sectors, yield errors were a poor indicator for
Exhibit 19: The yield curve has flattened significantly this year in all sectors
Year-to-date statistics regarding the slope of the curve in various sectors; bp
1s/2s 2s/3s 3s/5s 5s/7s 7s/10s 10s/30s
Current 4.6 9.9 42.7 42.6 40.1 55.1
Start 35.8 48.3 90.8 65.5 57.0 76.7
2011 YTD statistics Average High 25.5 56.4 36.1 59.8 76.8 98.0 59.8 68.2 54.7 63.2 77.0 93.5
Low 0.4 9.9 42.7 40.8 36.9 51.3
YTD chg -31.2 -38.4 -48.1 -22.9 -17.0 -21.7
* As of 11/17/2011 COB
Exhibit 20: Investors reaching for carry will only find it further out on the curve
3-month carry and slide in various sectors of the yield curve; bp
14 12 10 8 6 4 2 0 ED2 1Y ED6 * As of 11/17/2011 COB 2Y ED10 3Y 5Y 7Y 10Y 30Y
convergence trades, underperforming asset swap spreads as well as o-spreads as a value metric. This is also evident in the fact that since the ending of Fed purchases in the 2- to 5-year sector of the curve, yield error has underperformed the other two value metrics. Third, there is some evidence to suggest that o-spreads rather than traditional asset swap spreadsare a better value metric when Fed purchases are not a factor. This is seen by looking at the performance of the three different value metrics in the 2- to 5-year sector, which has been free of Fed purchase distortions since the end of QE2. Looking ahead to 2012, we anticipate that the general trend in these results will continue to holdconvergence trades at the long end of the curve will continue to be attractive when chosen using yield errors as a measure of value; for other sectors of the curve, convergence trades indicated by o-spreads as a value metric will likely prove profitable. With Fed purchases likely to be a mainstay of
141
Exhibit 21: At the front end of the swap curve, carry is now a mirage rising spot Libor rates at the front end has decreased the usefulness of carry as a measure of the attractiveness of Eurodollar trades
1-week moving average of carry-beta* versus spot 3-month Libor; %, inverted axis
Exhibit 22: Yield curves are now highly directional with yield levels thanks to ultra lowand stickyfront-end yields
Rolling 3-month beta between weekly changes in various yield curves and weekly changes in the corresponding longer end yield;
3 3-month Libor (inverted)
0.28 0.30 0.32 0.34 0.36 Carry beta 0.38 0.40
0.8 0.6 0.4 0.2 0.0 -0.2 May 11 2s/5s
5s/10s
-1 0.42 09 Aug 19 Aug 29 Aug 08 Sep 18 Sep 28 Sep 08 Oct * Beta between the 1-month P/L on a cross section of Eurodollar sector curve and butterflies versus the 1-month ago slide on those trades
Jul 11
Aug 11
Oct 11
markets in 2012, such relative value switch trades will be a key part of our market strategy in the year ahead. Swap yield curve The yield curve has been on the move in 2011, flattening significantly in all sectors of the curve (Exhibit 19). The flattening mostly occurred in 3Q, leading into and after the August FOMC meeting, as a number of factors conspired to flatten the curve. At the very front end of the curve, the biggest driver has been the Feds conditional commitment, announced at the August FOMC meeting, to hold the funds rate near current levels until mid-2013. Although falling short of a firm commitment, the Feds indication that the current economic outlook would likely be consistent with a low funds rate for at least two years had a significant impactfront-end term premium collapsed, OIS rates fell, the OIS curve flattened sharply, and yield levels declined sharply across much of the front end of the curve, out to the 5-year sector. The very front end of the swaps curve has been additionally biased flatter by the widening in Libor/OIS spreads. Further out the curve, the key driver has been the Feds Operation Twist. Much of this backdrop is likely to stay unchanged as we head into next year. Operation Twist will continue to dominate rates markets through the end of 1H12, and the Feds mid-2013 conditional commitment is likely to remain valid for quite some time, which should keep the very front end of the curve very flat for the foreseeable
future. In addition, the overhang from Europe will likely persist, weighing on funding spreads and thus the front end of the swap curve. The implications for yield curve trading strategies in 2012 are three-fold. First, investors will need to reach further along the curve for carry, as the flatness of the very front end has erased carry in the front and red Eurodollar sectors (Exhibit 20). Second, the very front end of the swap curve1-year swaps and front Eurodollarswill likely be the least attractive sector to earn carry, since longs in these sectors are vulnerable to a worsening of the European crisis. Carry in this sector is more optical than real, as spot Libor rates continue to rise, making forwards more likely to be realized than spot. Ironically, wider FRA-OIS spreads can lead to better carry/slide optically, making front-end longs look more attractive, even as rising spot Libor rates means that carry in this sector is merely a mirage. Empirical evidence supports this view. Exhibit 21 shows that the usefulness of carry as an indicator (measured as the beta between the 1-month P/L on a cross section of Eurodollar sector curve and butterfly trades and the exante carry/slide on those trades) has declined as spot Libor has risen; in other words, slide on the Eurodollar curve ceases to predict future spot yields when funding pressures are at work in pressuring Libor higher. Thus, going into 1H12, we will not look to earn carry at the front end of the Libor swap curve (unless of course if policy actions were to cause a more sanguine outlook on
142
the European crisis at some point). Indeed, as we noted in an earlier section, rather than owning the front end of the Eurodollar curve for the sake of carry, being short the sector via short-expiry payer swaptions offers the most asymmetric way of positioning for deterioration in Europe. Third, the yield curve is likely to exhibit a very high degree of directional exposure to yield levels, an artifact of ultra-low yields in the less-than-5-year maturity sector. This is seen in the elevated positive betas between various yield curves and yield levels (Exhibit 22). Hedge curve steepeners with Eurodollars Based on these observations, three types of trading themes are likely to prove attractive in 2012. First, forward curve steepeners (or spot curve steepeners with the front end being anchored in sectors that offer the best carry) hedged with shorts in the Eurodollar sector are likely to be attractive ways to earn carry while also protecting against exogenous risk from Europe. To identify the best trades, we identified the potential reward and risk in numerous such combinations. We estimated the risk weight on the Eurodollar hedge as the 1-year average of the rolling 3-month beta between weekly changes in each curve and the selected Eurodollar yield. Based on this hedge ratio, we then estimate the potential gain as the 3-month carry/slide, plus 50% of the potential upside, which in turn is defined as the average level of the weighted yield spread since the start of 2010 minus its current level. In order to properly assess risk, we focus on a downside-vol measure (in addition to a more typical standard deviation based volatility estimate). Specifically, we estimate the downside risk by assuming that semi-peripheral spreads (an average of Spain, France and Italian sovereign CDS spreads) widen by 100bp, and estimate the likely move in the weighted spread, and we additionally assume a 1-standard error downward move from this relationship. Thus, our downside risk measure reflects an attempt to capture the effect of deteriorating conditions in Europe. Finally, we calculate an efficiency ratio, which is the ratio of potential gain to the downside risk. Exhibit 23 presents a visual snapshot of results, and charts the potential gain versus the magnitude of the downside risk, while Exhibit 24 presents more detailed statistics regarding numerous such trades, ranked according to their efficiency ratios. As can be seen, 10s/30s and 5s/30s steepeners, hedged with shorts in red
Exhibit 23: Steepeners such as 10s/30s and 5s/30s, hedged with shorts in red Eurodollars, are attractive ways to earn carry and position for yield curve normalization, while hedging European exogenous risks
Potential gain* versus the magnitude of downside risk** for various curve steepeners hedged with short Eurodollars; bp
35 30 25 10s/30s 20 15 5s/10s
5s/30s
2s/10s and 3s/10s 10 1s/5s, 2s/5s and 3s/7s 5 10 20 30 40 50

Downside risk; bp * Upside defined as average of the weighted yield spread corresponding to each trade since 1/1/2010, minus its current value as of 11/14/2011. Potential gain defined as carry/slide plus one half of the upside. ** Downside risk estimated as the expected decline in the weighted yield spread for a 100bp rise in semi-peripheral spreads, plus a 1-standard error downward move in addition. Beta with respect to semi-peripheral spreads estimated based on a regression of monthly changes since 1/1/2010.
Exhibit 24: Detailed statistics regarding the potential reward and downside risk in yield curve steepeners hedged with shorts in red Eurodollars
Beta of various yield curves against Eurodollar yields*, the current weighted yield spread corresponding to each trade**, 3-month carry and slide, potential upside***, 3month standard deviation of weekly changes, downside risk**** and efficiency ratio***** for various yield curve steepener trades hedged with red Eurodollar shorts.
Curve ED Beta Spread Carry Upside (%) (bp) (bp) Vol (bp) Downside Efficiency vol (bp) ratio
3M fwd 10s/30s 6th -0.24 0.73 4.4 26.5 11.0 20.8 0.85 1Y fwd 10s/30s 6th -0.25 0.61 4.1 22.2 9.2 18.4 0.83 6M fwd 10s/30s 6th -0.24 0.69 4.0 25.0 10.3 20.0 0.82 2Y fwd 10s/30s 6th -0.24 0.42 4.2 17.3 7.5 15.9 0.81 2Y fwd 5s/30s 6th -0.47 1.16 9.0 32.3 16.7 32.1 0.78 3M fwd 5s/30s 6th -0.37 1.71 7.7 51.6 25.5 43.2 0.78 1Y fwd 5s/30s 6th -0.46 1.53 8.0 42.1 21.5 38.0 0.77 3Y fwd 5s/30s 6th -0.39 0.75 7.5 25.8 12.9 26.9 0.76 3Y fwd 5s/10s 6th -0.20 0.53 4.0 12.1 7.9 14.4 0.70 2Y fwd 5s/10s 6th -0.23 0.74 4.8 15.0 10.4 17.6 0.69 3Y fwd 3s/10s 6th -0.32 0.97 7.9 18.2 13.8 25.5 0.67 3Y fwd 2s/10s 6th -0.40 1.30 9.6 22.9 17.8 32.5 0.65 2Y fwd 3s/10s 6th -0.36 1.34 7.4 24.4 18.4 30.6 0.64 1Y fwd 5s/10s 6th -0.21 0.92 3.9 19.9 13.3 21.6 0.64 2Y fwd 5s/10s 5th -0.26 0.76 5.0 11.2 10.0 16.8 0.63 2Y fwd 2s/10s 6th -0.42 1.71 7.2 32.2 24.0 37.6 0.62 * 1-year average of 3-month beta of weekly changes in the curve versus weekly change in ED yield. ** Current value as of 11/14/2011. *** See footnote in previous exhibit. **** See footnote in previous exhibit. ***** Efficiency ratio defined as the ratio of 3-month carry plus 50% of upside to downside volatility.
143
Exhibit 25: Yield curve butterflies begin to exhibit non-linear behavior as front end yields become sticky near the zero bound
A curve neutral (25:75 weighted) 2s/10s/30s swap yield butterfly spread, versus 2year swap yield; %
0.2 0.1 0.0 -0.1 -0.2 -0.3 0.4 0.5 0.6 0.7 0.8 0.9 2-year swap yield; % 1 1.1
front-end rally or a sell-off. Moreover, since the underlying reason for this nonlinearitythe stickiness of yields when near the zero boundaryis also affecting swaption implied volatility skews, we may use information from swaption skews to infer which butterflies are likely to exhibit the most nonlinear behavior (see Implied Weights in Yield Curve Relative Value Trades, J.P. Morgan Research note, May 25, 2011, for a detailed discussion of our framework). Exhibit 26 outlines several butterflies that are likely to exhibit the most asymmetric behavior with respect to rates under different rate scenarios, thanks to such nonlinearities. Some of these trades have positive (or very slightly negative) carry, which also makes them attractive ways of positioning for a large, unanticipated, move higher in front-end yields. Use payer swaptions to position for an eventual normalization in rates Third, with yields extremely low at the front end, and with steep implied volatility skews, payer swaption spreads (1:1 weighted) now offer attractive reward-topremium ratios (Exhibit 27), and are attractive as limited-risk ways of positioning for an eventual normalization in rates towards higher levels, should that occur in 2Q12 as our rates strategists expect. Although we expect yield levels to fall in the early part of 2012, we do expect such a move to higher rates in the middle of next year (see Treasuries). Given the highly uncertain nature of policy developments in Europe, and the resulting risks in outright short-duration positions, we view 1:1 weighted limited-risk payer swaption spreads as the preferred way to position for the normalization in yield levels that we expect as we head towards the middle of next year. Moreover, with yields expected to drift initially lower in 1Q12, we would expect to find even better entry levels for such trades in the early part of next year, and would look to initiate payer swaption
Eurodollars are among the best such trades; variations involving 3- and 2-year forward 2s/10s steepeners (again hedged with red Eurodollars) are also attractive, but slightly less so. As we head towards 2012, we would look to a framework such as this to identify the best ways to position for a normalizing in yield curves. Take advantage of non-linearities in the yield curve due to sticky front-end yields Our second yield curve trading theme takes advantage of yield curve nonlinearity in low yield regimes. One consequence of yields being very close to the zero bound at the front end of the curve is that yield curve butterflies will likely exhibit nonlinear behavior with respect to yield levels. This is strikingly evident by looking atfor instancea curve neutral 2s/10s/30s butterfly (25:75 weights on the wings) versus 2-year yields (Exhibit 25). This suggests that curve-neutral belly-richening butterflies locally hedged for exposure to front-end yields should experience empirically convex upside, in either a
Exhibit 26: The nonlinearity implied by the skew is pronounced for various benchmark butterflies
Left weight given a shift in left rate of -25bp 0bp +25bp +75bp Right weight given a shift in left rate of -25bp 0bp +25bp +75bp
Projected evolution of left- and right-leg weights for several butterflies under various left leg swap yield shifts* and projected upside** in each scenario and carry (bp).
Projected Projected Projected upside from upside from upside from -25bp scenario +25bp scenario +75bp scenario 3M Carry
Butterfly
3s/7s/15s 1.32 1.03 0.85 0.69 0.34 0.36 0.34 0.30 3.3 2.1 10.4 1.2 2s/7s/15s 0.86 0.60 0.46 0.35 0.62 0.64 0.62 0.54 2.7 1.3 4.8 -3.7 3s/10s/30s 0.78 0.60 0.48 0.37 0.63 0.66 0.63 0.55 1.5 0.8 3.1 -0.4 5s/7s/15s 1.07 1.02 0.99 0.96 0.15 0.15 0.15 0.13 0.6 0.3 1.7 2.6 * Our methodology is as follows. For each left leg rate shift, we calculate the right leg rate shift based on the current ratio of ATM vols. We then combine the left and right rate shifts with the current implied weights to calculate a belly rate shift for that scenario. Third, we use the current implied vol skews to estimate the change in then-ATM implied vols for those rate shifts in each leg. Last, armed with estimates for then-ATM implieds in all three legs, we calculate new implied left and right weights. ** The projected upside is estimated as sum over left and right leg values of half the change in weight from the un-shifted scenario times the corresponding shift in rate.
144
Exhibit 27: Payer swap spreads (1:1 weighted) are attractive as limitedrisk ways of positioning for an upward drift in rates by mid-year
Exhibit 28: Yield curves have become highly directional with rates, and options markets are pricing this in to some extent but not yet fully
Implied* and empirical** betas between the 3-month forward 2s/10s curve and 3Mx10Y forward swap yield; 0.9
Empirical beta Implied beta
Statistics regarding 1:1 weighted payer swaption spreads (buy the ATMF strike, sell the A+50 strike) with a 6/30/2012 expiry on various underlying tails
Yields, % Tail 1Y 18M 2Y 3Y 5Y 7Y 10Y 15Y 30Y
Spot Forward bp 0.752 0.846 9.3 0.780 0.859 7.9 0.799 0.897 9.8 0.897 1.080 18.3 1.324 1.570 24.6 1.750 1.963 21.3 2.151 2.317 16.6 2.503 2.615 11.2 2.701 2.764 6.3
Slide Premium bp yield 12.1 12.1 12.2 13.3 16.1 17.6 18.6 19.2 20.1
Max gain 37.9 37.9 37.8 36.7 33.9 32.4 31.4 30.8 29.9
Ratio 3.1 3.1 3.1 2.8 2.1 1.8 1.7 1.6 1.5
0.8 0.7 0.6 0.5 0.4 Aug 11
* As of COB 11/17/2011
Sep 11
Oct 11
Oct 11
Nov 11
spreads on the back of declines in yield levels. Conditional trades with a twist The fourth yield curve trading theme that we will attempt to take advantage of in 1H12 involves exploiting the high degree of directionality between yield curves and yield levels. With yield curves anchored at the front end expected to be highly directional with yield levels, conditional curve trades (i.e., conditional flatteners in a rally and conditional steepeners in a sell-off) will likely be an ongoing trading theme in 1H12. Options markets are pricing in this directional nature of yield curves to some extent, but not fully. Although swaption implied volatility on short tails (such as 2s) is well below implied volatility on longer tails (such as 10s), the implied beta between the curve and long end yields is well below empirical betas (Exhibit 28). This
*Implied beta calculated as 1-(3Mx2Y / 3Mx10Y) implied bp vol ratio. ** Empirical beta calculated as the 3-month beta between weekly changes in the 3M fwd 2s/10s curve and 3Mx10Y forward swap yield.
allows for the construction of premium-neutral weighted conditional curve trades at the current time; however, with yield curves expected to remain highly directional for quite some time, we would not expect the options markets to persistently allow for attractive entry levels on such trades. How else can one exploit yield curve directionality once the swaptions markets have priced in the relative vol differentials fairly? To answer this question, we explore a variant of the usual conditional curve trades. Our variant is based on two observations: first, the 10s/30s curve is highly correlated (in magnitude, but with a negative beta) to front-end yields, and second, implied volatility on the 10s/30s curve in the YCSO market has tended to trade rich. This suggests that YCSOs on the
Exhibit 29: Conditional curve trades with a twistsince the 10s/30s curve is well correlated to front-end yields, and since YCSOs on 10s/30s are rich, 2-year tail swaptions can be replaced with options on the 10s/30s curve to create synthetic conditional curve trades at better entry levels
Implied volatilities (bp/day), notionals ($mn) and premium($) for conditional curve trades and corresponding values for trades in which the left leg option is replaced by a YCSO curve cap.
Trade Expiry Left Right 3m 2Y 5Y 3m 2Y 7Y 3m 2Y 10Y 3m 2Y 30Y 6m 2Y 5Y 6m 2Y 7Y 6m 2Y 10Y 6m 2Y 30Y
Desired Trade Implied vol Type Left Right Rec 3.46 5.26 Rec 3.46 6.61 Rec 3.46 7.86 Rec 3.46 8.96 Rec 3.36 5.28 Rec 3.36 6.41 Rec 3.36 7.46 Rec 3.36 8.26
Notionals Left Right -100 40.8 -100 29.8 -100 21.7 -100 9.6 -100 40.7 -100 29.7 -100 21.6 -100 9.6
Net prem Replacement for left leg Net Beta wrt on modified Premium Expiry YCSO Imp. Vol Left leg Notional Type Premium trade $111,655 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 $ (373,100) $195,317 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 $ (289,438) $272,424 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 $ (212,331) $339,559 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 $ (145,196) $167,224 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 $ (569,328) $265,495 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 $ (471,057) $355,936 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 $ (380,616) $423,282 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 $ (313,270)
145
Exhibit 30: Synthetic curve tradeswhere the front end leg has been replaced with a beta-weighted position in the 10s/30s curveare likely to track the desired original trade
6-month forward 2s/10s swap curve, versus the yield spread corresponding to replacing the 2-year leg with a beta weighted 6-month forward 10s/30s curve position
Exhibit 31: A tale of two different vol markets in 2011being short volatility was profitable until July, and long volatility positions have prevailed afterwards
1-month Gamma and Vega P/L* from long straddles in 3Mx10Y swaptions ; bp of notional
5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 Sep 11
Yield spread for synthetic trade; % 6M fwd 2s/10s; %
2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4
60
Gamma PL Vega PL
40 20 0 -20 -40
Nov**
-60
Jan Jun Mar Feb May Apr Aug Sep Oct Jul
Oct 11
Nov 11
*Calculated as 6Mx10Y minus (6Mx10s/30s curve divided by -0.34)
10s/30s curve may be used in place of swaptions on 2year tails; specifically, receiver swaptions on 2-year tails may be replicated via similar expiry caps on the 10s/30s curve, and payer swaptions on 2-year tails may be replaced with floors on the 10s/30s curve. Exhibit 29 presents details for a sample set of conditional curve trades, with the 2-year leg being replaced by options on the 10s/30s curve; for instance, the short receiver swaption position in a conditional bullish flattener would be replaced with short position in 1-look caps on the 10s/30s curve with the same expiry. The short payer swaption leg in a conditional bearish steepener would similarly be replaced with a short floor position on the 10s/30s curve. As can be seen in the table, such modified synthetic conditional curve trades are interesting because they can mimic an unweighted conditional curve trade and also take in premium currently, which cannot be done right now in the swaptions market. A visual confirmation that such a trade would indeed track the yield curve (which, after all, is the ultimate intent of a conditional curve trade) is seen in Exhibit 30. The yield spread corresponding to a 6-month forward 2s/10s curve trade, where the 2-year leg has been replaced with a beta weighted position in the 10s/30s curve, is shown in this exhibit; as can be seen, this spread does indeed closely track the 2s/10s curve, which was the original desired trade in this example.
* Returns calculated using the J.P. Morgan Volatility Index, which assumes daily delta rebalancing and zero transaction costs. Options are re-struck at the start of each month. ** November-to-date P/L.
Exhibit 32: Expect liquidity to remain poor: risk appetite continues to decline as the European crisis spreads into semi-peripheral countries
Reported average quarterly VAR for the 9 largest investment banks* versus 3-month moving average of semi peripheral spreads**; $mn
140 130 120 110 100 VAR Semiperipheral spreads
50 100 150 200 250
90 300 Dec 09 Apr 10 Jul 10 Oct 10 Feb 11 May 11 Aug 11 * Average VAR reported by JPM, GS, MS, BAC, C, UBS, CS, Soc Gen and DB. 3Q11 VAR is an estimate, projected from the results reported to date (excluding C). ** Average of 5-year CDS spreads on France, Italy and Spain.
146
Exhibit 33: Intra-year seasonals are likely to be unhelpful to market liquidity, as evidenced by the tendency of market depth to reach its lows in December
Percentage deviation of market depth in the 10-year sector from average over the past five years
Exhibit 34: Larger market moves become likely in periods of low market depth, as a price-taker will need go deeper down the order stack to trade a given amount
3.0
Regression of market width* spread against market depth** in the 10-year sector; bp
40% 20% 0% -20% -40% -60% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec * Market depth is calculated as the average size of the top three bids and offers, in $mn, for the on-the-run 10-year Treasury note, averaged between 8:30 a.m. and 10:30 a.m. daily.
2.5 2.0 1.5 1.0 0.5 150 200 250 Market depth; $mn * Market width calculated as the spread between mid-market and the level at which a fixed size ( $150mn) of the on-the-run 10-year Treasury note can be traded, assuming a static order book. Bid-side and offer-side widths are averaged here. ** see footnote in previous exhibit 50 100
Options
Short volatility strategies have thus far managed to remain profitable in 2011 year-to-date, but it has been a clear case of two different market environments. From the start of the year through the end of July, selling gamma proved considerably profitable overall. In contrast, since August, the opposite has been the case, and long gamma strategies have steadily gained ground (Exhibit 31). While the outperformance of long gamma strategies in August is a typical seasonal phenomenon related to declines in market depth, heightened macro uncertainty stemming from Europe has thus far forestalled a recovery in market depth, helping preserve conditions that are favorable to long gamma positions. Looking ahead towards the remainder of this year and 1Q12, our best guess is that risk appetite will remain low, and liquidity conditions will remain poor. Declining risk appetite is perhaps most evident in the daily VARs reported by major investment banks in their quarterly financials. As seen in Exhibit 32, VARs remain on the downtrend, reflecting growing caution as the European crisis has spread into semi-peripheral countries. In the near term as we head into year end, seasonals are unfavorable as well and could make for worsening risk
Exhibit 35: A model for seasonally-adjusted market depth
Model* for market depth in the 10-year sector, and forecast for 2012 Chg in driver, from current Factor Coefficient T-stat 1Q12 2Q12 4Q12 Intercept 303.1 97.0 Semi-peripheral spreads -0.47 -27.8 70 125 70 Cross-asset correlation -76.1 -19.2 -0.1 -0.3 -0.3 77% R-squared 26.4 Std error Impact on market depth -26 -36 -10
Projected market depth 70 60 86 * Model based on regression of seasonally-adjusted market depth against semiperipheral spreads and cross asset correlation, which we proxy using the rolling 3month correlation between weekly changes in the S&P and 30-year swap yields. Semi-peripheral spreads are an average of 5-year CDS spreads on France, Spain and Italy, and are used to capture broader risk aversion. Regression estimated over two years of history.
appetite and market depth. As seen in Exhibit 33, market depth has exhibited a tendency to decline to annual lows in December. This should prove unhelpful to market liquidity and could support elevated realized volatility. One way to see this is to estimate the width of the market at which a given fixed size can be transacted by a pricetaker. For instance, an investor needing to (say) buy $150mn 10s might simply lift the highest offer in a period when market depth is high, but might need to lift (say) the top three or four offers in periods of lower market depth, assuming a static order book. Examining the implied width of the market for a given fixed size (we
147
use $150mn) is insightful. As Exhibit 34 shows, this implied market width grows nonlinearly with depth, highlighting the fact that the market is much more prone to violent moves in periods of low market depthas even small-sized trades produce sizeable moves. Looking past year end, the key question then is whether market depth will begin to improve with a renewal of risk appetite next year. While this has historically been the case seasonally, there are good reasons not to expect such a pick-up after year-end 2011. In order to better understand the drivers of market depth, which has become a key determinant of realized volatility, we analyze the drivers of market depth on a seasonallyadjusted basis (given significant intra-year variation due to seasonals). As seen in Exhibit 35, market depth (adjusted for intra-year seasonals) has been well explained by two factors over the past two years. The first factor is a measure of cross asset correlations, proxied here by the rolling equity-rate correlation. All else equal, rising cross-asset correlations effectively increase the risk in an investors portfolio, in turn leading to cuts in notional amounts of risk-taking in individual markets. As short-term cross-asset correlations have risen in markets, measured VARs in portfolios have likely risen as well, prompting risk reduction; market depth would likely suffer in such periods. A second factor that has also biased market depth lower has been the steady contagion from the European crisis, which we measure by using an average of semiperipheral spreads. Finally, although not formally in our model, we note that some of this decline in market depth is also structural, reflecting stricter capital and leverage requirements on banks. Looking ahead, cross-asset correlations are likely to remain elevated or even increase further in the near term going into year end, but decouple thereafter in 2012. Semi-peripheral spreads, however, are likely biased wider, in the view of our European rates strategists. Incorporating their views into the framework of our model for market depth suggests that the typical newyear rebound in market depth may be unlikely; we project that market depth is likely to average similar or lower levels for much of 1H12, as also seen in Exhibit 35. What does this mean for delivered volatility? To answer this, we turn to a model for realized volatility we have
Exhibit 36: Look for realized volatility to decline only modestly after year end, and remain elevated in 1H12
Statistics regarding a model* for 1-month ahead realized volatility on 3Mx10Y forward swap yields, and forecast for 2012
Factor Intercept 3M fwd 2s/10s curve (%) Market depth ($mn) 3Mx10Y yield (%) H.F. leverage index R-squared Std error Impact on del. Vol (bp/day) Projected realized vol 8.4 Coefficient T-stat -0.6 -0.6 4.0 -0.017 -1.5 5.85 72% 0.97 -0.8 7.6 -0.8 7.6 -1.7 6.7 5.0 -11.1 -2.9 15.6 Current 1.49 98 2.29 1.4 Chg in driver, from current 1Q12 2Q12 4Q12 -0.3 -26 -0.35 -0.1 0.4 -36 0.45 -0.4 0.5 -10 0.65 -0.5
* Model based on 1-year regression of 1-month 3Mx10Y realized volatility against ex-ante levels of the 2s/10s swap yield curve, market depth in the 10year sector**, 3M forward 10-year swap yields and the hedge fund leverage index***. ** Market depth is calculated as the average size of the top three bids and offers, in $mn, for the on-the-run 10-year Treasury notes, respectively, averaged between 8:30 a.m. and 10:30 a.m. daily. *** Leverage defined as the sum of the absolute value of each of the six beta coefficients in the following multiple regression: daily excess returns on the IQ Hedge Global Macro beta index regressed against daily excess returns on 1) J.P. Morgan global bond index ex-US, 2) J.P. Morgan US 7-10Y bond index, 3) J.P. Morgan US 7-10-year minus 1-3-year bond index, 4) MSCI G7 index, 5) USD (J.P. Morgan USD cash index minus J.P. Morgan global cash index), and 6) GS Commodity index. Higher values imply more leverage.
Exhibit 37: Mortgage market convexity needs are near historic lows
Convexity exposure in Fannie Mae and Freddie Macs portfolios*; $bn 10s per 10bp
-2 -3 -4 -5 -6 -7 -8 -9 -10 2009 2010 2011
* Estimated from their reported interest rate sensitivities.
used in recent months. This model explains realized volatility in the 10-year sector using yield levels, the curve, market depth and an index of hedge fund leverage as independent variables. It is presented in Exhibit 36 together with the details of our projections for 2012. As can be seen, we generally project elevated realized volatility over much of 2012 (i.e., levels comparable to current implieds), and would thus look for an
148
Exhibit 38: Longer-dated swaption implied volatility has been well explained recently by three factorsthe level of volatility overall, mortgage market convexity, and inflation expectations
3Yx10Y implied volatility versus in-sample model fair value *
Exhibit 39: Look for 3Yx10Y swaption volatility to decline to 6bp/day in 1Q12
Statistics regarding a model* for 3Yx10Y implied volatility and forecast for 2012
Factor Coefficient T-stat Intercept 1.4 9.7 GSE portfolio cvx ($bn 10s/10bp) -0.17 -17.1 10Y inflation swap rate; % 1.0 21.3 3Mx10Y implied vol; bp/day 0.30 37.9 80% R-squared 0.335 Std error Actual implied vol; bp/day Projected implied vol, bp/day Current -2.6 2.26 7.65 1Q12 -2.1 1.81 7.6 Projections 2Q12 4Q12 -3.3 1.96 7.6 -3.6 1.96 6.7
10 9 8 7 6
3Yx10Y implied vol Model
6.9 6.4
5.9
6.2
6.0
* See Exhibit 38
Exhibit 40: Implied correlation has plunged in anticipation of Operation Twist
5 Nov 08 Jun 09 Dec 09 Jul 10 Jan 11 Aug 11 * Model based on 3-year regression of 3Yx10Y implied vol against GSE portfolio convexity (see Exhibit 37), 10-year inflation swap yields (%) and 3Mx10Y implied vol (bp/day).
Implied correlation between 6M forward swap yields from YCSO market. Average value from 8/18/2010 to 8/18/2011, value as of 8/18/2011 and drop during the period 8/18/2011-11/18/2011; %
100%
Avg
8/18/2011
Chg 8/18/2011-11/18/2011
6Mx10Y
6Mx30Y
1Yx10Y
1Yx30Y
2Yx10Y
2Yx30Y
6Mx5Y
1Yx5Y
2Yx5Y
environment that is largely favorable to long gamma positions. In the early part of the year, the main driver of this is likely to be heightened risk aversion and low market depth that combine to create illiquid conditions that support high realized volatility. Later in the year, however, the driver is likely to be higher rateswe expect US yields to decouple from the European crisis later in 2012, and rising yields will likely emerge as the main driver sustaining elevated realized volatility. Long-dated volatility As we contemplate the outlook for longer-dated swaption implied volatility in the year ahead, it is useful to revisit the key drivers of implieds in this sector in recent years. In addition to the overall level of volatility, which would of course drive implieds all across the vol surface, there have been two important drivers. The first is mortgage market convexity, which has of course steadily declined in magnitude, resulting in declining demand for intermediate expiry swaption volatility. We use the imputed convexity of the two GSEs portfolios net of hedges (which they report monthly) as our preferred variable to capture mortgage market convexity needs in the aggregate (Exhibit 37). A second factor that has been relevant in recent years is long-term inflation expectations. As we note in the Cross Sector Overview, there are not many attractive alternatives to position for a sharp rise in long-term
50%
0%
-50%
-100%
Exhibit 41: and that, coupled with an inverted volatility curve, has led to cheap forward volatility
Beta weighted differential between forward and spot volatility, current level and 2-year Z-score
Richness/cheapness of forward vol
2-year beta of forward vs spot vol Forward minus spot vol spread* (bp/day) 2-year average (bp/day) 2-year Z-score
Forward
Spot
6Mx5Yx5Y 6Mx5Yx10Y 1Yx5Yx5Y 1Yx5Yx5Y 1Yx5Yx30Y 6Mx1Yx10Y 1Yx1Yx10Y 1Yx1Yx10Y 6Mx6Mx10Y 1Yx6Mx10Y 1Yx6Mx10Y
5Yx5Y 5Yx10Y 5Yx5Y 5Yx5Y 5Yx30Y 1Yx10Y 1Yx10Y 1Yx10Y 6Mx10Y 6Mx10Y 6Mx10Y
1.02 1.00 0.93 0.93 1.03 1.12 1.10 1.10 0.96 0.90 0.90
-0.6 -0.4 -0.2 -0.2 -0.9 -1.4 -1.4 -1.4 -0.7 -0.2 -0.2
-0.3 -0.2 0.2 0.2 -0.5 -0.9 -0.7 -0.7 0.4 0.7 0.7
-3.4 -3.1 -2.9 -2.9 -2.7 -2.2 -2.1 -2.1 -1.9 -1.9 -1.9
* Beta weighted.
149
Exhibit 42: A Bermudan swaption may be thought of as spanning numerous European swaptions, corresponding to each of the various possible exercise dates
Graphic representation of a 1Yx5Y Bermudan swaption with semi-annual exercise points, and its spanned European swaptions. t=0 t=1Y t=6Y 1Yx5Y Bermudan swaption
Spanned European swaptions t=1Y (first exercise) 5Y swap t=1.5Y 2Y expiry 3.5Y expiry t=3.5Y 4Y swap
Exhibit 43: also, it can be thought of as spanning numerous Canary swaptions

Graphic representation of a 1Yx5Y Bermudan with annual exercise points, and its 10 spanned Canary swaptions 1Yx5Y Bermudan swaption t=0 t=1Y t=6Y
t=1Y Spanned Canary swaptions t=2Y 1Y(2Y)x5Y Canary t=3Y 1Y(3Y)x5Y Canary t=4Y 1Y(5Y)x5Y Canary t=2Y t=5Y 2Y(3Y)x4Y Canary t=4Y t=3Y 3Y(4Y)x3Y Canary t=4Y
2.5Y swap
t=5Y
5.5Y expiry (last exercise) 6M
inflation expectations, other than outright bearish trades or payer swaptions. Indeed, recent years have seen considerable demand for trades designed to protect against a rise in long-term inflation, via payer swaptions/ payer swaption spreads (or CMS caps or cap spreads), although this demand has ebbed and flowed with market sentiment in this regard. These two factors, together with 3Mx10Y swaption implied volatility (as a proxy for the overall level of volatility) do indeed successfully explain most of the variation in implied volatility levels in the intermediate expiry sector, as seen in Exhibit 38. We also project the likely evolution of the drivers in Exhibit 39. In particular, we project the convexity of the GSEs portfolios based on the most recent known data point, the recent relationship of this variable with yield levels, and our projections for yield levels in 2012; we project 10year inflation swap yields based on our projections for TIPS breakevens (see TIPS); and we draw upon our forecast for short expiry swaption implied volatility detailed earlier. Putting it all together, our projections point to declines in 3Yx10Y implied volatility to 6bp/day by the end of 1Q12, with range-bound levels thereafter.
Exhibit 44: The BC basis is consistently narrower than the BE basis under different yield curve scenarios
1Yx5Y ATM receiver Bermudan/Canary basis and Bermudan/European basis for the CTD divided by the Bermudan vega* for various parallel shifts of the yield curve**; %
120% BC basis/Berm. vega 100% 80% 60% 40% 20% 0 25 Shift in yields; % * Bermudan vega in bp of notional per bp shift in daily vol ** COB 11/10/2011 -150 -75 -25 75 150 BE basis/Berm. vega
Bermudan Swaptions As investors look ahead to devising their volatility strategies in 2012, several underlying principles will likely be important. First, given extremely low rates across much of the curve, implied volatility across much of the volatility surface has become highly directional with rate levels.
150
Exhibit 45: The Canary CTD mimics the Bermudan risk characteristics better than individual European counterparts, both for delta
0.3 0.2 0.1 0.0 -0.1 -0.2
Exhibit 46: and vega
Deviation of the vega profile of selected European and Canary swaptions from the Bermudan vega profile plotted against parallel shifts of the yield curve; %
Deviation of the delta profile of selected European and Canary swaptions from the Bermudan delta profile plotted against parallel shifts of the yield curve; %
3% 2% 1% 0% -1% -2% -3%
1Y(2.5Y)x5Y Canary 1Yx5Y European 2.5Yx3.5Y European
1Y(2.5Y)x5Y Canary 1Yx5Y European 2.5Yx3.5Y European
-0.3 -150 -100 -50 0 50 100 150 The delta profile is obtained as a % of the original delta (that is, for no change in the yield curve). Each curve is obtained by subtracting the European or Canary delta profile from the Bermudan delta profile. The 1Y(2.5Y)x5Y Canary and the 1Yx5Y European are the CTDs.
-150 -100 -50 0 50 100 150 The vega profile is obtained as a % of the original vega (that is, for no change in the yield curve). Each curve is obtained by subtracting the European or Canary vega profile from the Bermudan vega profile. The 1Y(2.5Y)x5Y Canary and the 1Yx5Y European are the CTDs.
Exhibit 47: Sources of value in the 5Yx30Y Bermudan receiver swaption

11/01/2011
All else equal, this argues for structuring portfolios to have directional vega exposurei.e., positions that become long vega in a sell-off are attractive. Second, one consequence of the Feds recent initiation of Operation Twist has been the plunge in correlation (Exhibit 40). While this is most noticeably true for correlations between front-end and back-end rates, it is also true more broadly. Last, with liquidity likely to remain poor, realized volatility is likely to remain high as well and preserve an inverted implied volatility curve across expiriesi.e., short expiries are likely to remain elevated relative to longer expiries. One corollary of the last two pointscheap implied correlation and an inverted volatility curveis that forward volatility appears cheap (Exhibit 41). Thus, vega strategies for 2012 should seek to leverage the current cheapness of implied correlations and forward volatility, while creating positions that gain vega in a selloff. Interestingly, Bermudan receiver swaptions offer a way to do all of those in one package. Since Bermudan swaptions can be exercised at several points in time, the value of near-term exercise grows with the moneyness of the option. Thus, Bermudan receivers, for instance, should increasingly track short-expiry receiver swaptions in a rally, but effectively behave more like longer-dated vega instruments in a sell-off (see Introduction to Bermudan Swaptions and a Framework for Analysis,
Properties* of Bermudan swaption, Canary CTD and volatility curve as of COB
Premiums from a Bermudan model 5x30 Berm receiver 1891.5 bp 5Y(14Y6M)x30Y Canary 1741.1 bp Vega 33.1 bp of notl per 0.1 bp/day Alternate fair values for Canary based on: Implied correlation 1848.6 bp Cheapness of Canary (& Berm) In price terms In vol units 107.5 bp of notional 0.32 bp/day
plus
Mispricing in correlation markets Canary price using implied correlation 1848.6 bp Canary price using realized correlation 1851.7 bp Correlation driven cheapness of Berm In price units In vol units 3.1 bp of notional 0.009 bp/day
plus
Mispricing of 5x30/14.5x20.5 vol curve 0.001 bp/day
equals
Total mispricing 0.335 bp/day
151
* Premia quoted in bp of notional and vega in bp of notional by 1/10 bp shifts in daily volatility.
J.P. Morgan Research Note, July 1, 2003 for more details). In addition, Bermudan swaptions are also naturally long exposure to correlation and forward volatility. However, the complexity of Bermudan swaptions (due to several possible exercise dates) means that they have exposure to a number of different correlations and forward volatilities. It is therefore unclear exactly which implied correlations and which forward volatilities a given Bermudan structure is chiefly exposed to, if such simplification is even possible. More broadly, while Bermudan receiver swaptions may offer a slate of desirable characteristics, it is unclear if they offer cheap means to add such exposures since developing a relative value view on Bermudan swaptions is challenging. In this section, we build on our previous research on Bermudan swaptions to develop exactly such a relative value metric for each Bermudan structure. Our approach is as follows. First, we identify a simpler instrument that is spanned by a Bermudan swaption that most closely mirrors the Bermudan itself. In our earlier work, we noted that each Bermudan swaption may be thought of as spanning a basket of European swaptionsone for each possible exercise date. Moreover, the premium of the Bermudan swaption must be higher than each of the European swaptions spanned by it, and the European swaption with the highest premium in that basket may be thought of as the best European approximation to the more complex Bermudan, at least locally (Exhibit 42). (We label that particular European swaption as the Bermudans CTD, motivated by the Treasury futures analogy and notwithstanding the fact that it is the most expensive European swaption in the basket, not the cheapest). While that observation is useful in developing some intuition regarding the behavior of Bermudan swaptions, the relative instability of the Bermudan swaptions European CTD has in practice meant that it does not easily lead to a value metric. In this piece, we examine a natural generalization of that idea that does lead to a value metric. Much like the Bermudan swaption may be thought of as spanning a basket of one-exercise date European swaptions (one for each possible date on which the Bermudan can be exercised), we may also think of the Bermudan swaption as spanning a basket of twoexercise date Canary swaptions, with one Canary for
Exhibit 48: The relative value metric shows several sectors as currently mispriced
Components* of our relative value metric for Bermudan receiver swaptions and total in various sectors expressed in price terms (bp of notional) and Bermudan volatility terms (bp/day) as of COB 11/01/2011
Sector 5Yx30Y 6Mx30Y 1Yx30Y 3Yx30Y 2Yx30Y 6Mx2Y 6Mx10Y 3Yx10Y 5Yx10Y 3Yx7Y 1Yx10Y 2Yx10Y 5Yx5Y 6Mx3Y 6Mx5Y Bermudan / Canary basis cheapness* Price Vol 107.5 93.5 91.0 93.2 86.1 2.6 7.7 13.8 12.0 4.5 7.2 8.2 6.4 1.6 1.6 0.32 0.33 0.31 0.30 0.28 0.24 0.09 0.12 0.09 0.06 0.08 0.08 0.10 0.09 0.05 Correlation cheapness** Price Vol 3.1 -1.1 -0.9 0.7 -0.4 -0.2 7.3 5.1 8.3 6.0 3.5 4.5 -0.3 1.9 3.8 0.01 0.00 0.00 0.00 0.00 -0.02 0.09 0.04 0.06 0.08 0.04 0.04 -0.01 0.11 0.11 Vol curve cheapness*** Price Vol 19.3 13.3 13.7 15.9 14.6 0.3 1.7 0.3 1.3 1.0 0.9 0.4 1.6 -1.6 -3.3 0.06 0.05 0.05 0.05 0.05 0.03 0.02 0.00 0.01 0.01 0.01 0.00 0.02 -0.09 -0.10 Total Price 129.8 105.7 103.9 109.7 100.4 2.8 16.6 19.2 21.5 11.5 11.6 13.1 7.7 1.9 2.0 Vol 0.39 0.37 0.36 0.35 0.33 0.25 0.20 0.17 0.17 0.14 0.13 0.13 0.12 0.11 0.06
* Canary CTD premium using implied correlation minus Premium of the Canary CTD implied by Bermudan price. ** Canary CTD premium using realized correlation minus premium using implied correlation. *** Canary CTD premium using the second forward rate volatility shifted by vol curve mispricing (from a 2-year regression of vol curve versus second vol level) minus premium using current implied volatilities and realized correlation.
Exhibit 49: Total mispricing of the 1Yx10Y Bermudan receiver swaption over the past four years
Total mispricing in the 1Yx10Y Bermudan receiver swaption; bp/day
0.2
0.0
-0.2
-0.4
-0.6 Nov 07 May 08 Dec 08 Jun 09 Jan 10 Aug 10 Feb 11 Sep 11 * Negative values indicate cheapness of Bermudan swaptions
each possible pair of dates on which the Bermudan swaption can be exercised (Exhibit 43). We recall here that a Canary option is one that can be exercised on exactly two dates; for instance, a 1Y(2Y)x5Y has possible exercises one and two years from now into swaps that end in six years. Moreover, just as was the case earlier, the price of the Bermudan swaption must be at least as large as the highest premium canary that is spanned by the Bermudanwe refer to that particular
152
Canary as the Canary CTD, in keeping with our earlier terminology. But why is it useful to think of a Bermudan swaption as spanning a set of Canaries? Because the Canary CTD is much more stable than the European CTD, across a range of rate, curve and volatility moves, making it a useful simplification (Exhibits 44-46). Moreover, canaries also are exposed to exactly one correlation and forward volatility, which will be useful to us. For instance, the 5Yx30Y Bermudan currently has the 5Y(14.5Y)x30Y as its Canary CTD, with resulting exposure to 5Y forward 9.5Yx20.5Y correlation and forward volatility. We may now develop a relative value metric. In doing so, we note that there are three possible sources of value. First, the price of the canary CTD in the Bermudan swaptions market (or from a suitably calibrated Bermudan swaption pricing model) may differ from the price of a canary that can independently be calculated from vanilla implied volatilities and implied correlations from the YCSO market. Second, using an empirical framework, we may conclude that implied correlations in the YCSO market are themselves rich or cheap, based on comparisons to realized correlations over time. Third, again using an empirical framework, we may take a view on the volatility curve corresponding to a given canary. Last, we may total up the mispricing from all three sources, and divide by the vega of the Bermudan swaption to compute the richness or cheapness of the Bermudan swaption, expressed in basis points per day. This is illustrated in detail below for the case of the 5Yx30Y Bermudan receiver swaption in Exhibit 47. The sources of value are also illustrated for the currently most mispriced Bermudan receiver swaptions in Exhibit 48, while Exhibit 49 shows a time series of the total mispricing in the past four years for the particular case of the 1Yx10Y Bermudan receiver. Large negative values of our total mispricing should be taken as a signal to buy the Bermudan. As a way to analyze the usefulness of our metric we evaluated the performance of buying several Bermudan receiver swaptions at previous points in time in 2011 when our metric suggested cheapness. Three strategies were tested: (i) hedging only the delta, (ii) hedging the delta as well as the vega by selling the European CTD swaption, and (iii) hedging the delta as well as vega by selling a portfolio of replicating European swaptions. Our results in this regard are preliminary, and much more analysis is needed. However, our initial results are
Exhibit 50: More empirical analysis is needed, but our results currently suggest that our metric does indeed identify value in Bermudan swaptions, and also that the best way to monetize this metric is to buy the Bermudan swaption against the European CTD on a delta hedged basis
Mispricng at inception Bermudan vol terms (bp/day) -0.44 -0.44 -0.43 -0.38 Delta hedged returns over 3-months Vega hedged to European CTD bp of notl. bp/day bp of notl. bp/day 88.5 0.4 25.0 0.1 -105.4 -0.4 311.3 1.3 -125.8 -0.5 61.9 0.3 -217.4 -0.9 185.8 0.8
3-month return of long Bermudan positions delta hedged*, vega hedged with respect to the European CTD, and vega hedged to a replicating portfolio**;
Vega hedged to replicating portfolio notl. bp/day 8.5 0.0 28.5 0.1 -25.9 -0.1 -152.6 -0.6
Trade start Bermudan date 1/3/2011 5Yx30Y 3/1/2011 4/1/2011 5/2/2011
* Options are delta hedged daily and rolled every month. Returns assume zero transaction costs. ** Replicating portfolio refers to a weighted combination of European swaptions that are spanned by the Bermudan swaption. The weights are solved for using a scheme to replicate the Bermudan swaptions characteristics under a range of scenarios. See Introduction to Bermudan Swaptions and a Framework for Analysis, J.P. Morgan Research Note, 7/1/2003 for details.
promising, and suggest that such a relative value metric is in fact useful in signaling cheapness of Bermudan swaptions (Exhibit 50); our (limited) results currently appear to suggest that buying the Bermudan versus the European CTD swaption (and dynamically delta-hedging the switch) is the best way to monetize the cheapness of Berms.
Regulatory update
Under the Dodd-Frank Act, several government agencies such as the CFTC, SEC, FDIC, and the Federal Reserve were tasked with rule-making responsibilities. We review developments this year on this front, as well as some potential changes in markets that could result from the broader regulatory reform effort. Proposed and finalized rules The CFTC, the key rulemaking body as it pertains to derivatives markets, was tasked under the Dodd-Frank Act to regulate Swap Dealers (SD) on capital and margin requirements, as well as on more stringent recordkeeping and reporting. The act also required the CFTC to specify rules related to the migration of trading standardized derivatives to regulated exchanges or Swap Execution Facilities (SEF). The process of arriving at final rules has proved fairly time intensive. In July, the CFTC issued an order clarifying the effective date of the provisions in the swap regulatory regime established by Title VII of the Dodd-
153
Frank Act. The order provides temporary relief from provisions that that would have become effective July 16, 2011 up to the effective date of final rules or December 31, 2011. Specifically, the CFTC provided relief from certain provisions in the Commodities and Exchange Act (CEA) that do not require rule-making but reference one or more terms regarding swap entities and instruments (such as the terms swap, swap dealer, major swap participant, or eligible swap participant) that the act requires be further defined. The CFTC has, jointly with the SEC, issued notices of rulemaking to further define these terms. In its latest amendment to its July order, the Commission is proposing to extend the latest date of expiry of relief from December to July 16, 2012 as it has not yet finalized its definition of these terms. As such, several aspects of the rule-making are still in various stages of either being open to comments or under discussion by the commission. Rules in this stage include, but are not restricted to, key items affecting OTC derivative markets as relating to margin requirements for uncleared swaps, capital requirements for Swap Dealers (SD) and Major Swap Participants (MSP), and collateral type and custody rules. Other items yet to be finalized include rules on Derivatives Clearing Organizations (DCOs) with regards to financial resources and the resource waterfall in case of a default by a clearing member, as well as end-user exceptions from the mandatory clearing requirement. Several rules were in fact finalized, such as those related to swap data repositories, on particular products such as agricultural swaps, and certain enforcement-related items, but as noted earlier, key rules remain outstanding. The Dodd-Frank Act also required the Federal Reserve to implement several provisions, sometimes in conjunction with other agencies. As part of this inter-agency effort (between the Fed, FDIC, FCA, FHFA and OCC), a proposal seeking comment was issued in April that would require swap entities regulated by the five agencies to collect minimum amounts of initial margin and variation margin from counterparties to non-cleared swaps and non-cleared, security-based swaps (with a commercial end-user exception as long as the entities margin exposure is below an appropriate credit exposure limit). The amount of margin required under the proposed rule would vary based on relative risk of the counterparty and of the swap. The proposal would apply to new, non-cleared swaps entered into after the proposed rules effective date. Existing capital standards that apply to swap entities as part of their prudential regulation
Exhibit 51: Cleared interest rate swap volumes have risen significantly over the last few years
Percent of notional outstanding cleared via LCH
60%
50%
40%
30%
20% 2007 Source: ISDA, BIS, LCH 2008 2009 2010
Exhibit 52: Cleared volumes of other interest rate derivatives have risen over the year
Percent of notional outstanding centrally cleared in 2011
72% 70% 68% 66% 64% 62% 60% 58% Dec 10

Source: Trioptima
IRS OIS IR Basis Swap (rhs)
31% 28% 25% 22% 19% 16% 13% 10%
Feb 11
Apr 11
Jun 11
Aug 11
Oct 11
already address non-cleared swaps, and therefore are not separately addressed in the rule. The Fed also requested public comment on the Volcker Rule requirements, developed jointly with the FDIC, OCC, SEC and CFTC. Briefly, the Volcker Rule prohibits insured depository institutions, BHCs, and their subsidiaries or affiliates from engaging in short-term proprietary trading of any security, derivative, and certain other financial instruments for the entitys own account, subject to certain exemptions. The act also prohibits exempt transactions that may result in a conflict of interest with customers or in a material exposure to high-risk assets or trading strategies as defined by the rule. The proposal clarifies the scope, as well as provides
154
certain exemptions to prohibitions in the act. Transactions in certain instruments, such as obligations of the US government or a US government agency, the GSEs, and state and local governments are exempt from the statutes provisions. As expected, it also exempts market-making, underwriting, and risk-mitigation hedging. It also provides commentary to assist entities in distinguishing market making-related activities from proprietary trading. Finally, we note that next year brings the observation period for the Liquidity Coverage Ratio (LCR) proposed under Basel III. Relevant regulators will start collecting data to assess LCRs, with mid-2013 being the deadline before which any changes to the current LCR proposal may be announced. Our short-term analysts note that in its existing form, LCR is likely to be fairly onerous on banks. In its current form, this requirement would result in increased demand for scarce high-quality collateral. Market developments and outlook Despite extensions in the rule-making process, markets have moved ahead towards a central clearing model. This trend towards a clearing model predates the new regulations (Exhibit 51). As of October 2011, around half of the $511tn interest rate derivatives outstanding were cleared. A majority of the cleared derivatives were interest rate swaps (87%), which are about 60% of notional of interest rate derivatives outstanding. Over the last year, roughly two-thirds of interest rate swaps outstanding (based on USD-equivalent notionals) have been centrally cleared, and the percentages of other products currently cleared (OIS and basis swaps) have grown substantially (Exhibit 52). Another trend that predates post-crisis regulations is the rise in the level of OTC derivative collateralization. As seen in Exhibit 2, margin coverage in fixed income derivatives is high, and covered over half the trade volume well before the crisis. One outcome of the move to central clearing and more stringent collateral requirements has been the move to OIS discounting as discussed earlier. The move would result in a one-time accounting gain (loss) depending on whether the counterparty was long (short) an in-themoney swap position. More interesting is the question of whether there would be a significant switch from Liborbased hedges to OIS-based ones. Given the much higher
levels of liquidity in Libor swaps, we think this is unlikely in the near term. Another potential consequence of regulatory changes is a change in issuer behavior. Corporate issuers that swap their fixed-rate debt issuance may incur higher costs irrespective of whether they may avail of the end user exemption from mandatory clearing because their dealer counterparties will incur higher capital costs on uncleared swaps. The proposals could also impact callable issuance. Many issuers of callables routinely sell the embedded option on issuance to monetize favorable pricing and achieve a lower funding cost. While swaptions are not yet a cleared product, they will nevertheless require their own collateral. Even though this should not be an issue for some larger issuers like FNMA and FHLMC that carry a significant amount of eligible collateral, other issuers like FHLB that do not (and non-US issuers such as KFW that do not typically fully collateralize their derivative positions) would have to evaluate the cost savings from callable issuance against additional collateral costs. On the margin, this could result in lower callable issuance and lower vol supply. We note, however, that the rules are not finalized with regards to which of these entities will be eventually subject to mandatory clearing and margin requirements, even as entities seek exemptions from various provisions of the act.
Trading themes
Look for FRA-OIS spreads to stay wide or widen further into 1Q12, but begin to narrow thereafter In the near term, worsening conditions in Europe are likely to bias the EUR/USD cross currency basis narrower, bias the semi-peripheral European sovereign spreads wider, and preserve upward pressure on FRAOIS spreads. However, looking past this likely nearterm intensification, policy actions (a potential reduction in the penalty on the Feds USD swap lines to 50bp from 100bp) are likely to eventually dominate, pushing FRA-OIS spreads narrower to 50bp by the end of 1H12. Swaptions on short tails are effectively options on front-end swap spreadsuse receiver/payer swaptions in place of front-end spread narrowers/wideners With the Feds conditional commitment to low rates until mid-2013 helping to keep front-end Treasury yields low and sticky, front-end swap spreads have
155
become highly correlated to front-end swap yields. Thus, options on front-end yields can effectively serve as options on front-end swap spreads. Buy payer swaptions or 1:1 weighted payer swaption spreads in place of front-end swap spread wideners, and buy receiver swaptions or receiver swaption spreads as proxies for front-end spread narrowers. Position for wider 10-year maturity matched swap spreads in the near term In the near term, FRA-OIS spreads will likely continue to widen, and banking stock valuations will likely remain under pressure, both of which should pressure 10-year swap spreads wider. We expect 10year swap spreads to hover near 27bp towards the end of this year and in early 1Q12. but look to initiate narrowers towards the end of 1Q12 High grade corporate issuance tends to exhibit strong intra-year seasonal patterns, and issuance-related swapping is likely to pick up over the March-May period. Intermediate maturity swap spreads have consequently tended to narrow in this period. In addition, as we look into next year, the growing likelihood of concerted action by central banks (including perhaps lower penalties on USD swap lines with the Fed) should also help swap spreads to turn the corner. Look for narrowing to 18bp by mid-year. Trade intra-month swings in the pace of Fed purchases of longer-end Treasuries While the Fed tends to purchase similar amounts each month in total, the actual schedule of purchases (which is known ahead of the month) can produce significant intra-month variations in pace. Intermediate maturity swap spreads have tended to narrow in periods characterized by a sharp slowing in pace, while widening when the opposite is true. Take advantage of relative value opportunities in sectors subject to Fed purchases A simple strategy of buying the cheapest issue (measured using yield error as a metric) and selling the richest issue on a maturity matched swap spread switch basis has been attractive in periods and sectors where the Fed is actively purchasing Treasuries. With Operation Twist slated to continue until the end of 1H12, we would look to make such switches a key part of our relative value trading strategy in 1H12.
Position for a steeper yield curve between intermediates and the long end, hedged with red Eurodollar shorts Steepeners anchored in the 5-year sector and beyond offer attractive carry and slide in addition to being considerably mispriced relative to front-end yields. Also, thanks to the flattening of the front end of the curve, shorts in red Eurodollars have small carry costs, and also help mitigate exogenous risks. Take advantage of yield curve nonlinearities induced by the stickiness of front-end yields near the zero bound As front-end yields become highly sticky near the zero bound, their relative volatility versus points further out the curve falls, causing locally level and curve neutral belly-richening butterflies to exhibit empirical convexity versus front end yields. Butterflies that are likely to exhibit the greatest asymmetry can be identified by examining implied volatilities and skew information. In addition, some of these butterflies have positive carry and slide, making them particularly attractive ways to position for an unanticipated front end selloff. The yield curve should remain highly directional with rates in 1H12; combine YCSOs with swaptions to create synthetic conditional curve trades that offer better entry levels than regular conditional curve trades Sticky front-end yields are likely to cause the yield curve between the front end and intermediate yields to remain highly directional with yield levels for the foreseeable future. However, swaptions markets have priced such directionality in (albeit to varying degrees), and entry levels are often unattractive. By exploiting the correlation of the 10s/30s swap curve with front-end yields, and given the richness of YCSOs on the 10s/30s curve, replacing swaptions at the front end of the curve with suitably weighted YCSOs results in a synthetic conditional curve trade that can be constructed at much better entry levels. Stay long gamma going into year end and in 1H12 A persistent crisis in Europe and poor risk appetite should cause market depth to stay depressed for much of 4Q11 and 1H12, creating conditions favorable for long gamma positions.
156
Look for intermediate-expiry swaption implied volatility to drop into 1Q12, but stay range-bound thereafter Intermediate expiry swaption implied volatility currently looks rich; look for 3Yx10Y swaption implied volatility to fall to 6bp/day by the end of 1Q12, and remain range-bound thereafter. Look to add exposure to Bermudan receiver swaptions Cheap implied correlations and an inverted implied volatility curve have made forward volatility cheap. In addition, the strong relationship between implied volatility and rates that emerges in low yield regimes makes it desirable to hold positions that become longer vega in a sell-off. One attractive way to initiate directional vega exposure, while also positioning for a rise in implied correlations and forward volatility, is to replace European receiver swaptions with Bermudan receiver swaptions.
157
Japan Rates Research Global Fixed Income Markets 2012 Outlook November 24, 2011 Takafumi YamawakiAC (81-3) 6736-1748 Yuya Yamashita, CFA (81-3) 6736-1493 J.P. Morgan Securities Japan Co., Ltd
Japan
2011 was a noteworthy year for the JGB market in three respects: 1) The Great East Japan Earthquake of 11 March 2011 2) JGB-UST decoupling, and 3) A weakening of the ultra-long sector. However, JGB stayed within a narrow range of 0.95-1.35% in 2011 Our main scenario points to a continuation of range-bound trading while our risk scenarios suggest that JGB yields are more likely to fall than rise. The downside in yields could be tested if the global economy loses momentum in 1H2012 while a gradual cyclical recovery in 2H12 should trigger only moderate rises in yen interest rates. We are therefore anticipating ranges of 10Y JGB yields as 0.8-1.1% for 1H12 and 0.91.3% for 2H12 This report considers why Japanese sovereign debt can be considered relatively safe by international standards while also identifying some cause for concern in the longer term We would consider any rise in JGB yields that might be triggered by the FY12 issuance plan to be little more than extremely short-lived noise We expect short-end swap rates to remain sticky under the BoJs interest rate policy. We also have in mind a risk scenario where LIBOR rises on a possible worsening of the crisis in Europe. We have a flattening bias on JPY swap curve as a whole We have a widening bias on short-end swap spreads. We think the cheapness of JGB vs. swap rates in the super-long sectors will be sustained into 2012 Large negative cross-currency basis spreads make it attractive to create synthetic USD bonds by buying JGBs and swapping them to USD. With higher FX hedge cost, Japanese investors may accelerate the repatriation of USD funds back to JGBs We have a tightening bias on the TIBOR/LIBOR fixing spread on positive convexity. We expect more tightening if LIBOR increases than widening if LIBOR falls We think the short-to-medium-term 3s/6s basis spreads are biased wider
Exhibit 1: JGB stayed in a narrow range of 0.95-1.35% in 2011

10YJGB yield history; %
1.50 1.40 1.30 1.20 1.10 1.00 0.90 0.80 Jan-10 Jul-10 Jan-11 Jul-11 0.95% high correlation w ith UST 1.35% QE2 Earthquake Decoupling from UST
The JGB market in 2011

Three key points What sort of year has 2011 been for the JGB market? US Treasury yields rose sharply in January and February as equity prices were driven higher by the Fed's late-2010 QE2 announcement and the Obama administration's decision to extend the so-called Bush tax cuts. The 10Y JGB yield hit its year-to-date high of 1.35% as yen bonds were caught up in this global selloff, but the domestic market then turned unusually volatile in the wake of the 11 March 2011 Great East Japan Earthquake, with losses on stock positions apparently triggering a certain amount of profit taking in JGBs towards the 31 March end of the Japanese fiscal year (Exhibit 1) Yen interest rates followed UST yields lower from midApril through July, but then failed to keep pace with further declines from August onwards. The year as a whole has thus been characterised by a rather gradual downtrend in JGB yields since May. 2011 was a noteworthy year for the JGB market in three respects. 1: The Great East Japan Earthquake of 11 March 2011 The Great East Japan Earthquake of 11 March 2011 appears set to be one of the defining events of Japan's already long history. While the initial earthquake ended up causing relatively little direct damage, the subsequent tsunami destroyed a number of towns and villages and triggered a major crisis at the Fukushima Daiichi nuclear power plant. Japan is
158
by no means unaccustomed to earthquakes and tsunami, but the virtually unprecedented size of this tsunami nevertheless shocked the nation, with significant negative ramifications for consumer spending, manufacturing supply chains, and electricity supply capacity. The collapse and rebound of industrial production provides a particularly clear picture of the extent to which economic activity was impacted by the disaster (Exhibit 2). The following couple of months saw JGB market participants focus on: 1) the potential for post-quake reconstruction to drive a V-shaped economic recovery; and 2) the prospect of an increase in JGB issuance to pay for post-quake rebuilding. The subsequent economic recovery fell somewhat short of initial expectations, however, and with the JGB market asked to absorb only a limited amount of additional issuance, a significant rise in yen interest rates was averted. 2: The Euro zone crisis and JGB-UST decoupling UST-JGB correlations declined markedly as the European sovereign debt crisis dominated headlines from August onwards. UST and German Bund yields fell sharply as the US economy too began to lose momentum, but JGB yields dropped only slightly during this period. Exhibit 3 illustrates the extent to which yen interest rates were left behind by the fall in UST yields. 3: A weakening of the ultra-long sector The performance of ultra-long JGBs was a key difference between 2010 and 2011. The 30s/40s curve steepened sharply this year (Exhibit 4), as demand for these longest-dated bonds dropped dramatically while demand for other parts of the super-long sector remained strong. The 40s have a very thin investor base compared with that of the 20s, however, and we will be monitoring this recent slackening of the supply/demand in the 40s to see if it continues to put steepening pressure to the JGB yield curve in 2012.
Exhibit 2: The collapse and rebound of industrial production data provide a particularly clear picture of the extent to which economic activity was impacted by the disaster
Industrial production seasonal adjusted; CY 2005 = 100
100 95 90 Earthquake 85 80 75 Jul-10 Oct-10 Jan-11 Apr-11 Jul-11
Exhibit 3: UST-JGB correlations declined markedly from August onwards.

UST 10Y vs. JGB 10Y since October 2010; %
1.50 1.40 1.30 1.20 1.10 1.00 0.90 0.80 0.70 1.50
JGB 10y r
Aug11-Nov 11
Oct10 - July 11 UST 10y r 2.00 2.50 3.00 3.50
Exhibit 4: The 30s/40s curve steepened sharply this year

JGB 30s/40s curve history; bp
30 25 20 15 10 5 0 Oct-10 Jan-11 Apr-11 Jul-11 Oct-11
159
Outlook for the JGB market in 2012

Six factors support a relatively narrow trading range While we do not expect volatility to remain as low as it has been in 2H11, we nevertheless expect to see a continuation of relatively range-bound trading in 2012. Looking at the key factors summarised in Exhibit 5, we see little prospect of significant yield movements in either direction. Below, we discuss our main scenario and risk scenarios for each of these factors. 1: Trade balance Main scenario: Large current account surplus despite near-zero trade balance (neutral bias on yields: 75%) Exhibit 6 shows Japans trade balance, which remained in deficit this year as exports fell in the wake of the earthquake, and import value was driven up by high commodity prices. The trade deficit exceeded 400bn (seasonally adjusted) in both April and May, but had narrowed to 21.8bn as of September. Japan may find it difficult to book a consistent trade balance surplus in 2012, given the prospect of exports being held back by a global economic slowdown, but we expect it to remain roughly in equilibrium (i.e. near zero) over the year as a whole. This would see Japan's massive income surplus translate into a similarly large current account surplus, implying that the market should have little problem absorbing even relatively high levels of JGB issuance.
Exhibit 6: Japan's trade balance remained in deficit this year and the current account surplus has been gradually shrinking recently
Current account surplus in Japan; tn
35 30 25 20 15 10 5 0 -5 -10 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 Income Current Transfer Trade Balance Serv ice
Risk scenario: Severe overseas economic slowdown or rise in commodity prices (upward bias on yields: 20%) Japans trade balance would likely plunge deep into deficit under a severe global recession, thereby fueling speculation that a current account surplus might be difficult to sustain in the longer term and potentially triggering a selloff of JGBs by foreign investors.
Exhibit 5: Six factors support for the JGB market moving in a relatively narrow range, but with a downward bias in yields
Six factors for the JGB market, their bias towards yield move and our subjective probabilities assigned to each scenario
Probability 1 Trade Balance Yield Down 5% Yield Neutral 75% Yield Up 20%
Comments If trade balance stays in the negative for a long time, some investors may have concerns about JGB sustainability. If the BoJ indroduce further easing such as increasing Rinban, extending the maturity of the Asset Purchase Program, and lowering the IOER If Japanese economy goes into a recession, tax increase might be extended. If US rates stay around the current levels for a long time, lifers will shift their investment money from USTs to JGBs. If banks face difficulties in non-performing loans, they will buy JGBs more aggressively. If UST stay below 2%, JGB yields will follow USTs.
BoJ Additional Easing
35%
60%
5%
3 4 5 6
Domestic Politics Lifer's Investment Activity Bank Activity UST Rates
10% 40% 30% 30%
65% 50% 60% 50%
25% 10% 10% 20%
160
2: Additional BOJ easing Main scenario: Additional easing limited to further expansion of Asset Purchase Program (APP) (neutral bias on yields: 60%) The Bank of Japan eased monetary policy on three occasions following the 11 March disaster.
14 March 2011: Increase in the amount of APP: + 5tn 4 August 2011: Increase in the amount of APP: + 10tn 27 October 2011: Increase in the amount of APP + 5tn
3: Domestic political situation Main scenario: Decision on consumption tax hike helps to maintain a minimum level of fiscal discipline (neutral bias on yields: 65%) Fiscal discipline became a major policy priority after Yoshihiko Noda took over from Naoto Kan as Prime Minister, with the need for sharply higher spending in the wake of the 11 March disaster seemingly counterbalanced by a strong commitment to responsible fiscal management. The Noda administration also appears determined to hike the consumption tax rate. In fact, the administration has set its guidline to hike the consumption tax rate from 5% to 10% in two stages from October 2013. Fiscal policy thus appears unlikely to be an Achilles heel for the JGB market provided that the Noda government can remain in office. Risk scenario: Political squabbling sees fiscal discipline thrown out the window (upward bias on yields: 25%) A severe global economic downturn could potentially see the fiscally hawkish Noda administration overthrown, in which case the new government could be expected to loosen the fiscal reins and relegate a consumption tax hike. 4: Investment by life insurers Main scenario: Shift out of foreign bonds into JGBs (neutral bias on yields: 50%) Japan has also seen an increase in USD funding costs (Exhibit 7), with 3M roll cost rising to around 0.8% annualised. This means that less than 1.5% of carry is now earned by buying 10Y USTs, as a result of which domestic investors have started to shift back into JGBs. The focus is life insurance companies as their activities are the key for the demand and supply balance of the long end of the curve. Their buying has been already modestly increasing over the past few years (Exhibit 8). Some investors may also have replaced a portion of their FX-hedged foreign bond holdings with unhedged positions, but full substitution appears unlikely given the risk it would entail. We therefore expect to see a gradual shift into (super-long) JGBs that should help to limit any rise in yields towards the far end of the curve. Risk scenario 1: Renewed investment in foreign bonds following a resolution of the Euro zone crisis (upward bias on yields: 10%) Confidence in USD funding may be boosted if a significant improvement in the Euro zone situation drives 10Y UST and German bund yields substantially higher,
161
However, each of these actions was limited to an expansion of the central banks APP, with the BoJ refraining from exercising any of its other options (2-4 below): 1) Further expansion of APP 2) Lengthening of duration of APP JGB purchases (from 2 years to 5 years) 3) Lowering of the interest rate paid on excess reserve balances 4) Increase in Rinban outright JGB purchases With overseas central banks seemingly reluctant to lower IOER (interest on excess reserves) rates or increase their outright bond purchases, we see little prospect of the BoJ opting for any of the options (2-4) unless the domestic economy were to slow dramatically. Central banks are increasingly prone to criticism for both direct underwriting of government debt and excessive purchases via the secondary market. With the BoJ understandably keen to avoid any negative surprises on the monetary policy front, we do not expect to see any announcements with sufficient impact to knock JGB yields out of relatively narrow trading ranges. Risk scenario: Yen driven higher by aggressive Fed easing (downward bias on yields: 35%) There is at least some chance that the BoJ will bow to pressure from markets and politicians by announcing slightly more aggressive measures, with (2) perhaps the most likely option (other than a further APP expansion) by virtue of its relative ease of implementation. If USD/JPY were to drop below 70 following a QE3 announcement by the Fed, then we would expect the BoJ to respond by lengthening the duration of its APP JGB purchases, thereby generating downward pressure on the curve as a whole and the 5Y yield in particular.
in which case European bonds and USTs could start to look more attractive to domestic lifers from a relative value perspective (versus JGBs). Risk scenario 2: Increased investment in JGBs as the Euro zone situation continues to deteriorate (downward bias on yields: 40%) A further escalation of the Euro zone crisis would probably make lifers increasingly risk averse, with a selloff in domestic equities and the corresponding need to reduce risk exposure likely to accelerate purchases of super-long JGBs. 5: Investment by banks Main scenario: Little change in earnings, little need for aggressive trading (neutral bias on yields: 60%) Recent financial results indicate that bad-debt disposal costs remain well under control, suggesting that there is little need for banks to chase relatively risky returns in the JGB or the broader securities markets. We therefore expect to see only moderate trading activity, with banks likely to take profits when yields fall and buy on (price) dips when yields rise. This should help limit overall market volatility. Risk scenario: Economic downturn increases reliance on trading profits (downward bias on yields: 30%) It is possible that a further escalation of the Euro zone crisis could have negative ramifications for Japan and the rest of Asia, in which case domestic banks may be forced to step up their bond trading with a view to covering higher bad loan costs. This could result in a slightly more volatile market, with JGB yields likely to decline at least temporarily. 6: Correlation with US Treasuries Main scenario: JGB yields unlikely to rise sharply if the 10Y UST yield remains somewhere around 2.5% (neutral bias on yields: 50%) As should be evident from Exhibit 3, the JGB market was much more sensitive to movements in UST yields prior to August 2011. A return to pre-August correlations would imply that a 10Y UST yield of around 3.0% should be commensurate with a 10Y JGB yield of around 1.20%. (Our UST research team is forecasting that the 10Y UST will stay below 3.0% throughout 2012.) Our main scenario for 2012 Our main scenario for each of the above six factors points to the continuation of range-bound trading, while
162
Exhibit 7: The increase in USD funding cost has reduced the attractiveness of USTs. We therefore expect to see a gradual shift into (super-long) JGBs that should help to limit any rise in yields towards the far end of the curve
3M FX hedge cost in USD/JPY: % annualised
1.00 0.80 0.60 0.40 0.20 0.00 Oct-10 Jan-11 Apr-11 Jul-11 Oct-11
* 3M FX hedge cost is calculated as annualized 3M forward discount vs. spot FX level. Source: Bloomberg
Exhibit 8: Lifers activities are the key for the demand and supply balance of the long end of the curve. Their buying has been modestly increasing over the past few years
3M moving average of monthly purchase amount in super long end from lifers; bn
1,400 1,200 1,000 800 600 400 200 0 Apr-08 Apr-09 Apr-10 Apr-11 3M moving average
our risk scenarios suggest (on balance) that JGB yields are more likely to fall than rise. The downside could be tested if the global economy loses momentum in 1H12, while a gradual cyclical recovery in 2H12 should trigger only moderate rises in yen interest rates. We are therefore anticipating ranges of 10Y JGB yields as 0.80%1.10% for 1H12 and 0.90%1.30% for 2H12.
Japanese sovereign risk

Still relatively safe, but some longer-term concerns remain A continued escalation of the European debt crisis has focused substantial attention on Japanese sovereign risk, but the stability of domestic interest rates over the past few years suggests that a sovereign shock is still quite a remote prospect. This report considers why Japanese sovereign debt can be considered relatively safe by international standards, while also identifying some cause for concern in the longer term. Why the JGB market is still relatively safe? 1: Massive domestic savings: A healthy balance sheet for the nation as a whole Exhibit 9 shows (inward and outward) financial assets and liabilities for the various sectors of the Japanese economy, with equities counted separately owing to their significant price volatility. These figures should help give an overall picture of Japanese financial flows and leverage. The household sector is seen to hold net financial assets totaling about 1,048tn, while the social security systems pension reserves amount to about 175tn in net assets. The (central + local) government sector has net liabilities totaling 840tn, while non-financial private corporations have around 39tn in net liabilities.
Foreigners net liabilities total 348tn, implying an equivalent amount of Japanese claims against foreign entities. Overall, we believe that Japan's balance sheet is healthy, with the existence of massive domestic asset holdings a key difference from the current situation in Europe. 2: High levels of issuance digested with little difficulty The past behaviour of these various sectors should help explain why the market as a whole has had little difficulty in absorbing continued increases in JGB issuance. Foreigners net liabilities (= Japans net assets) tend to decline when the yen is strengthening and increase when the domestic currency is weakening. In the absence of particularly dramatic exchange rate fluctuations, however, it seems likely that Japan will continue to gradually accumulate net assets by virtue of its consistent current account surplus. The existence of a current account surplus means that the government sectors fiscal deficit is being more than offset by increases in the net assets of the private sector (households + corporate). The corporate sectors net financial liabilities have declined sharply in recent years owing to a diminished need for borrowing, while the household sector has continued to build up its net assets. Some analysts have suggested that Japans adverse demographics an
Exhibit 9: Given massive domestic asset holdings, we believe that Japan's balance sheet is healthy
Financial assets and liabilities by sectors; tn
Financial Institution Inward Asset 2,430 Outward Asset 244 Inward Liability 2,642 Outward Liability 30 Net Asset 10 Equity Liability 105 Non Fin Corporate Inward Asset 607 Outward Asset 106 Inward Liability 480 Outward Liability 3 Net Asset -39 Equity Liability 407 Household Inward Asset 1,392 Outward Asset 9 Inward Liability 443 Outward Liability 0 Net Asset 1,048 Equity Liability
Equity Asset 113
Equity Asset 138
Equity Asset 90
Government Inward Asset 128 Outward Asset 88 Inward Liability 1,109 Outward Liability 2 Net Asset -840 Equity Liability 18
Social Security Inward Asset 147 Outward Asset 31 Inward Liability 24 Outward Liability 0 Net Asset 175 Equity Liability
Foreigners Inward Asset 212 Outward Asset 34 Inward Liability 217 Outward Liability 472 Net Asset -348 Equity Liability
Equity Asset 73
Equity Asset 21
Equity Asset 95
163
aging and shrinking population will eventually push the household sectors net assets into a downtrend, but this has thus far been prevented by significant transfers of income from the government sector to households (Exhibits 10&11). The resulting decline in corporate demand for funds and increase in deposits have left banks awash with surplus funds to invest in JGBs, thereby helping to keep yen interest rates low and relatively stable. We expect that this mechanism will continue to function for as long as Japans current account balance remains in surplus. 3: Current account surplus has declined slightly but is unlikely to disappear It is of course necessary to recognise that Japans current account surplus has been gradually shrinking recently as shown in Exhibit 6. The trade balance deteriorated sharply in the wake of the Great East Japan Earthquake of 11 March, owing to supply chain disruptions and a surge in oil imports. We expect to see only a moderate surplus going forward given the likelihood that Japan will continue to import fossil fuels while exporters are hit by a global economic downturn. Japan continues to book a large income surplus, however, and we expect this to remain somewhere in the order of 10tn per annum unless there is a significant decline in net claims against foreigners. This should also provide substantial support for the JGB market. 4: Difficult to envisage a default scenario While there has indeed been some discussion of the possibility that Japan might default on its debt, we consider this an extremely unlikely scenario for the following reasons. Shifting out of JGBs during a flight to quality will entail foreign currency risk, making this a very difficult option to contemplate for most market participants. (Shifting from Greek bonds into German bunds involves no such risk). We expect support from domestic investors not willing to take currency risk. Japan has its own central bank, which should be able to purchase JGBs in the event of a crisis. This makes it theoretically possible for Japan to avoid a default indefinitely, i.e. the BoJ is theoretically able to underwrite JGBs in the primary market with necessary law changes.
164
Exhibit 10: Despite adverse demographics, the household sector has continued to build up its net assets
History of net asset in household sector; tn
1,100 1,050 1,000 950 900 850 800 750 700 1997 1999 2001 2003 2005 2007 2009
Exhibit 11: on the back of significant transfers of income from the government sector to households
History of net liability by sectors; tn
900 800 700 600 500 400 300 200 100 0 1997 1999 2001 2003 2005 2007
Gov ernment
Foreigners
Corporate
2009
Even if the yen were to weaken as a consequence of JGB purchases by the BoJ, Japans ample foreign reserves mean that this would actually help to reduce the states net debt. (In the case of Argentina, depreciation of the home currency triggered a sharp increase in USD-denominated liabilities.) A depreciation of the yen would also be likely to boost exports and thereby bring about a significant increase in the current account surplus, with positive ramifications for the JGB market. We therefore see very little risk of Japan facing a Greecelike sovereign debt crisis.
5: Some gradual progress towards more stable fiscal footing There were concerns that the three supplementary budgets made necessary by the Great East Japan Earthquake of 11 March would require a sharp increase in JGB issuance, but the government has been able to deplete frontloaded JGB issuance to the point where the market will only be asked to absorb an additional 2tn (or thereabouts), which suggests that any impact on the overall supply/demand balance should be minimal. Moreover, the government appears to be making slowbut-steady progress in its efforts to return Japan to a more sustainable fiscal footing. The governments medium-term fiscal framework covering the period through FY2014 (endorsed by the Cabinet in August) seeks to limit annual general account spending (excluding debt-servicing costs) to 71tn and new JGB issuance to 44tn. Cabinet Office projections indicate that meeting these targets would still leave Japan unable to achieve its longer-term goal of wiping out the primary fiscal deficit by FY2020, but it is nevertheless impressive that Japan has clearly announced its intention to maintain a certain level of fiscal discipline despite the massive shock of the 11 March disaster. The government has also indicated that the post-quake rebuilding process is to be managed outside the general budget, with the necessary funds to be repaid via temporary tax hikes including higher personal income and residential taxes and non-tax revenues. The Noda administration has set its guidline to hike the consumption tax rate from 5% to 10% in two stages from October 2013. Cause for concern in the longer term 6: Massive imbalance between the government and private sectors Japan faces a severe imbalance where the government sectors massive net liabilities (currently exceeding 800tn) are covered in their entirety by the domestic private sector (households + non-financial private corporations). It is also somewhat troubling that the banking sector holds such a large proportion of the private sectors net assets. This lack of diversification has the potential to exacerbate any upward movement in yen interest rates.
Exhibit 12: The lack of diversification in the banking sector has been the main driver of the sharp rises in JGB yields over the past decade
Long history of JGB 10Y yield; %
2.00 1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 00 01 02 03 04 05 06 07 08 09 10 11 Bank Trading Banks reduced their rates ex posure in a rate hike cy cle Bank Trading
As should be evident from Exhibit 12, there have been two particularly sharp rises in JGB yields over the past decade: JuneAugust 2003, when a surge in VaR for the banking sector triggered an across-the-board reduction of interest rate risk, and 20052006, when banks sought to reduce their exposure to medium-term JGBs ahead of a BoJ rate hike. Insufficient diversification on the part of investors was a contributing factor in both cases. 7: Super-long issuance outstanding continues to rise A sharp increase in outstanding issuance of super-long JGBs is also worthy of some attention. Total outstanding issuance beyond the 11Y sector has risen by around 40tn over the past two years and is projected to exceed 140tn by end-FY2011 (Exhibit 13). This extremely rapid pace of supply reflects large and regular offerings via 20Y, 30Y, and 40Y tenders as well as Auctions for Enhanced Liquidity. Japanese life insurers are already preparing for the launch of a domestic version of the Solvency II framework (currently being implemented in Europe) around 34 years from now, and their associated ALM measures appear likely to ensure relatively strong demand for super-long JGBs over the next 23years. Beyond that point, however, it may be necessary to find a new investor base for super-long offerings. The current debate over a possible lengthening of the average
165
duration of JGB holdings for public pension portfolios could be a step in the right direction.
Exhibit 13: Total outstanding issuance beyond the 11Y sector is projected to exceed 140tn by end-FY2011, increasing the need to find a new investor base for super-long offerings.
Long-term outlook for the JGB market

As indicated in Exhibit 14, we expect the 10Y JGB yield to remain below 1.5% over the coming 34 years before shifting into a 1.5%2.0% range for the next few years, with fiscal risk unlikely to become a major market factor until at least a decade from now.
Total market outstanding of 10Y+ JGB excluding 15Y CMT, nominal bonds only; tn
200
Forecast
150
100
50
0 03 04 05 06 07 08 09 10 11 12 13 14
Exhibit 14: We expect the 10Y JGB yield to remain below 1.5% over the coming 34 years before shifting into a 1.5%2.0% range for the next few years, with fiscal risk unlikely to become a major market factor over the next 10 years
Long-term forecast of 10Y JGB yield; %
3.00 2.50 2.00 1.50 1.00 0.50 0.00 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Medium Term High Volatility 10Y=1.5-2.0% Long Term Gradual Yield Rise 10Y>2.0% Short Term Rangebound 10Y=0.8-1.5%
166
Forecasts for FY2012 calendar base market issuance

The Ministry of Finance tends to announce its JGB issuance plan for the following fiscal year around Christmas. As indicated in Exhibit 15, we expect the FY2012 issuance plan (which is also likely to be announced before the end of 2011) to show an increase of 300-400bn per month in calendar base market issuance (supply at scheduled auctions from April through March) relative to the offering amounts that will be seen from December 2011 onwards. We would expect this to break down into a 100bn/month increase in issuance of 1Y, 2Y and 5T JGBs and a further 100bn/month increase in the 20Y sector (Exhibit 16). Increases of this magnitude are unlikely to have a significant impact on overall market sentiment, and we would therefore consider any rise in JGB yields that might be triggered by the announcement to be little more than extremely short-lived noise. Our key assumptions are as follows. We currently expect the FY2012 issuance plan to show: No more than 44tn in new financial resource bonds (thereby meeting the target set out in the governments medium-term fiscal framework)
We expect this to break down as follows: Around 150tn in calendar base market issuance (offerings to regular auction participants) 4.2tn in Non-price Competitive Auction II issuance 2tn in adjustments by depleting front-loaded issuance Around 3.5tn in issues to individual (retail) investors, which would be on a par with the third FY2011 supplementary budget Around 16tn in long-term JGBs held by the BoJ are due for redemption during FY2012
Exhibit 16: and we would expect this to break down into a 100bn/month increase in issuance of either 2Y or 5TY JGBs and a further 100bn/month increase in the 20Y sector
Calendar base JGB issuance in FY2012; JPYtn
Source: J.P. Morgan
Calendar Base 6mTB 1Y TB 2Y JGB 5Y JGB 10Y JGB 20Y JGB 30Y JGB 40Y JGB Liquidity
FY11 Dec 0.9 2.5 2.7 2.5 2.2 1.1 0.7 0.4 0.6
FY2012 Amt 0.9 2.5 2.7 2.6 2.2 1.2 0.7 0.4 0.6 1 12 12 12 12 12 8 4 12 Total 0.9 30.0 32.4 31.2 26.4 14.4 5.6 1.6 7.2 149.7
Exhibit 15: FY2012 issuance plan (which is also likely to be announced before the end of 2011) to show an increase of 300-400bn per month in calendar base market issuance
JGB issuance plan (JPM estimate) in FY2012; JPYtn FY2011 3rd Suple Legal Grounds 181.7 44.3 New Resource JGBs 109.3 Refunding JGBs 16.5 FILP JGBs 11.6 Reconstruction JGBs 0.0 JGBs for govt pension
Lowest 173.9 43.9 113.5 12.0 2.0 2.5 FY2012 Forecast 176.1 44.0 114.0 13.0 2.5 2.5 Highest 179.2 44.1 114.5 14.5 3.5 2.6
Comment Government Target Around 114 trJPY Redemption of FILP bonds is smaller than FY2011 3.5trn at a maximum Temp issuance for Govt pension
Issuance Methods Calendar Base Non-Price Comp Depleting Front Loaded For Households Bank of Japan
FY2011 3rd Suple 181.7 145.7 5.9 14.8 3.5 11.8
Lowest 148.0 4.2 1.0 2.5 15.0

FY2012 Forecast 176.1 149.7 4.2 2.5 3.5 16.2
Highest Comment

152.0 5.0 4.0 4.0 18.0
Calendar base will be around 150trn 3.75% of total amount of 2Y + JGBs Same size as 3rd supplementary budget of FY11 Same size as the redemption in BoJ portfolio in FY12.
167
Interest Rate Derivatives Swap Curve

Over the past year, the swap curve bull-flattened up to 10Y sectors while the curve in the super-long sectors was shifted down in a parallel-to-modest bull-steepening fashion (Exhibit 17). The direction and magnitude were generally in line with the movement of the JGB yield curve. Even after markets recovered from the postearthquake turmoil, the BoJ continued to provide abundant liquidity in response to the deepening debt crisis in Europe and global economic slowdown. In addition, the Fed introduced an explicit time frame of interest rate policy until mid-2013 in August and then the so-called Operation Twist in September. The BoJ is expected to hold the policy rate at a low level even longer than the Fed. The bull-flattening of the JPY swap curve up to 10Y sectors this year is consistent with such global low-for-long environment. 6M LIBOR effectively lays a floor to the swap rates up to 5Y. Therefore, in a bullish environment, the level of the shorter-term swap rate tends to be well supported while longer end rates have more room for downshift given a slow fall in 6M LIBOR. Therefore, in the course of a downshift of the curve, short-end flatteners begin to show convexity from a certain point. And butterflies, after the belly richens as rates decline, begin to cheapen from a certain point. In fact, as the curve flattens further out, the 2s/5s/10s fly did not richen in a rally this year. On the other hand, the 5s/10s/20s spread has been on a richening trend for about a year (Exhibit 18) and may start to show convexity going forward. The long end of the curve remained flat. The inversion in the long end of the forward swap curve, which had first taken place after Lehman shock in 2008, was not corrected either this year, although the extent of inversion declined. The 10Yx10Y / 20Yx10Ycurve is still in negative territory at -36bp (Exhibit 19). The supply and demand balance in 30Y sector remained one-sided in 2011 with strong receiving of the long end of the curve. Firstly, declining FX rates and risky assets put a flattening pressure on the long end of the curve. While the issuance sum of structured notes remained small, hedging activities for existing positions continued. Secondly, asset swap investors stayed away from the 30Y sector. Domestic investors buy ASW up to 20Y. Overseas investors, who had once invested in 30Y JGBs as a LIBOR+ floating asset by using ASW, did not come back to the market. Lastly, hedge funds sporadically paid
168
Exhibit 17: Swap curve bull-flattened up to 10Y sectors while the curve in the super-long sectors was shifted down in a parallel to modest bull-steepening fashion over the past year
Change over the past year; % bp
2.00 1.50 1.00 0.50 0.00 0 5
Change (RHS) Term structure (LHS)
5 0 -5 -10 -15 -20 -25 -30
10
15
20
25
30
Exhibit 18: As the curve flattens further out, the 2s/5s/10s fly did not richen in a rally this year whereas the 5s/10s/20s has richened consistently
Swap 2s/5s/10s and 5s/10s/20s 50;50 butterfly since 2011; bp
0 -2 -4 -6 -8 -10 -12 -14 -16 Jan-11 2s/5s/10s Butterfly (RHS) Apr-11 Jul-11 Oct-11 -45 -50 5s/10s/20s Butterfly (LHS) -20 -25 -30 -35 -40
Exhibit 19: The inversion in the long end of the forward swap curve remained, though the extent of inversion declined. The 10Yx10Y/20Yx10Y curve is at -36bp
10Yx10Y / 20Yx10Y curve since 2011; bp
0 -10 -20 -30 -40 -50 -60 Jan-11 Apr-11 Jul-11 Oct-11
super-long swap rate and bought long-dated payer swaptions as a so-called Japan fiscal premium trade with the anticipation of an increased Japanese fiscal deficit leading to higher long-term yields. However, they generally preferred benchmark 10Y tail to express the view compared with 20Yx10Y for a liquidity reason because the trade was essentially more macro-driven. Going forward, for short-end rates, we expect them to remain sticky under the BoJs interest rate policy as a main scenario. At the same time, we also have in mind a risk scenario that LIBOR rises on a possible worsening of the crisis in Europe, which we think is getting to be more of a reality. Short-term swap rates will likely rise if LIBOR rises. We have a flat-to-upward bias on shortterm swap rates. On the other side of the curve, we expect the flatness of the long-end swap curve to continue into 2012. The crisis in Europe still has a long way to go before it approaches a resolution. The crisis puts a downward pressure on long-term swap rates via two routes. The supply and demand balance is unlikely to see a large change. Risk appetite of foreign (European) investors is unlikely to recover enough for them to come back to 30Y ASW in the near term. Also, a worsening of the crisis would lead to a selloff in risky assets and a JPY appreciation, which would flatten the long end of the curve through dealers hedging activities. In sum, we have a flattening bias on JPY swap curve as a whole. With short-end rates held at a low level, we focus on convexity trades such as weighted flatteners and weighted butterflies. In addition, if the markets price in a longer time frame of a low interest-rate policy, the effect will likely spill over from shorter ends to longer ends, as investors seek carry and move along the curve. In that scenario, we expect a cheapening of the belly on shorter dated butterflies such as 2s/5s/10s and then possibly 5s/10s/20s spreads as a medium-term trend. For example, the 5s/10s/20s 50:50 fly has a convex profile vs. 5s/20s curve, therefore a curve neutral 5s/10s/20s fly is expected to show a convex P&L (Exhibit 20). In the long end, long-dated steepeners are attractive in carry considertations..
Exhibit 20: The 50:50 5s/10s/20s butterfly has historically showed positive convexity to the 5s/20s swap curve
Swap 5s/10s/20s butterfly vs. swap 5s/20s spread since 2010; bp
5 0 -5 -10 -15 -20 -25 -30 100 110 120 130 140 150 5s/20s sw ap curv e; bp y = 0.02 x 2 - 3.52 x + 188.77 R 2 = 0.46
Exhibit 21: Swap spreads have broadly tightened across the curve with the exception of below 2Y over the past year
Change of swap spreads over the past year; bp
30 20 10 0 -10 -20 -30 1 2 3 4 5 6 7 8 910 12 15 Maturity 20 25 30 30 20 10 0
Change (RHS) Term structure (LHS)
-10 -20 -30
Exhibit 22: A PCA* indicates that swap spreads tended to move in the same direction, but the move tended to be larger in the longer end. There was also another factor that impacted the 7Y sector and super-long sectors in the opposite direction
January 2011 November 2011; first two PCA factors, first factor explains 73%, second factor explains 15% unit
0.60 0.40 0.20 0.00 -0.20 -0.40 -0.60 2Y 5Y 7Y 10Y 20Y 30Y 1st factor 2nd factor
Swap spreads
Swap spreads tightened over the past year (Exhibit 21). For one, we think this was related to a rise in fiscal premium for JGBs. As mentioned earlier, the increase in JGB issuance associated with the aftermath of the 11 March earthquake has continued to be in focus. On the other hand, from a swap perspective, the swap curve
* The PCA is conducted on the daily changes in swap spreads across the curve. The data period is from January November 2011.
169
faced a bull-flattening pressure, led by super-long sectors on the back of a selloff of risky assets and JPY appreciation. To get a better understanding of this years trend, we conducted a PCA. Exhibit 22 shows the result of the PCA analysis of the daily changes in swap spreads over this year. The first and second PCA factors explain 73% and 15% of the total variance, respectively. The result indicates that swap spreads generally moved in the same direction but the move was larger in the longer end. This is in line with the above explanation that a rise in fiscal premium caused this years swap spread changes. The second factor, which is independent of the first one and smaller in its impact, had an effect on the 7Y sector and super-long sectors in the opposite direction. We can think of this factor as somehow related to flight-to-quality movement with richening of 7Y on JGB futures buying and cheapening of 30 on swap curve receiving. In general, JPY swap spreads are largely affected by 1) repo/LIBOR spread as a determinant of baseline values, and 2) JGB yield levels. Since December 2010, the repo/LIBOR spread has stayed within the 21-23bp range (Exhibit 23), therefore had only a limited impact on swap spreads. In fact, in a simple multiple regression analysis over this year, the repo/LIBOR spread does not have a meaningful explanatory power and JGB yield levels, in turn, have higher t-values. Partly mentioned in the Swap Curve section, we have a flat-to-upward bias on LIBOR. From a monetary policy perspective, we do not expect the BoJ to hike a rate ahead of the Fed. Neither do we expect the BoJ to step in to an IOER cut. However, a cut in IOER is still possible. In this scenario, it is likely that with LIBOR falling faster than repo rates, the repo/LIBOR spread will tighten. A more likely risk scenario is that markets face stress and LIBOR rises. In that scenario, LIBOR rises while repo rates are kept low and the repo/LIBOR spread widens. All in all, we have a widening bias on short-end swap spreads. For the super-long sectors, we think the cheapness of JGB vs. swap rates will remain into 2012, given our view that the flatness in the long end of the JPY swap curve will continue. From a relative value and shorter-term perspective, swap spread curve trades are worth considering. One way to see such relative values is to run simple multiple regression. For example, Exhibit 24 shows the residuals of swap spreads from the regression analysis where swap spreads are regressed on underlying JGB yields and
170
Exhibit 23: Since December 2010, repo/LIBOR spreads have stayed within the 21-23bp range, therefore had only a limited impact on swap spreads.
Spread between the S/N repo rate and 6M JPY LIBOR since 2010; bp
38 36 34 32 30 28 26 24 22 20 Jan-10 Jul-10 Jan-11 Jul-11
Source: Bloomberg, J.P. Morgan
Exhibit 24: Our JGB swap spreads model on underlying JGB yields and 1Y LIBOR/OIS basis* indicates that front end spreads are too wide while long end spreads are too narrow. However, we would not fade this relative value.
Residuals of swap spreads as of 17 November 2011; bp
8 6 4 2 0 -2 -4 -6 2 3 4 5 6 7 8 910 12 15 Maturity
* The swap spreads are regressed on the underlying JGB yields and 1Y LIBOR/OIS basis. The data period is from January November 2011. The regression equation for 10Y swap spread is, for example, 10Y swap spreads = 8.9 19.0 x 10Y JGB Yield - 0.48 x 1Y LIBOR / OIS
JGB cheap
JGB rich 20 25 30
1Y LIBOR/OIS basis (which we think better reflects day-to-day moves than the repo/LIBOR spread) as explanatory variables. It indicates that front end spreads are too wide while long end spreads are too narrow. We would use this model to assess the relative richness/cheapness on the swap curve, however given our view of swap curve flattening we recommend investors not to chase this relative value.
Cross-Currency Basis Swap

USD/JPY cross-currency basis spreads gradually narrowed in 1H11. However, they widened back (moved ve) in 2H11. In addition, we had two sharp widening in August and November (Exhibit 25). A medium-to-long-term driver is dollar funding pressure. Concerns about European banks have heightened. Some banks that found it difficult to raise dollar funding via standard funding markets probably headed for crosscurrency basis markets. USD/EUR cross-currency basis spreads widened, and USD/JPY cross-currency spreads followed. On a flow side, coupon swap (forward USD buying) flows from Japanese investors, which had once counteracted a widening of the spreads in the past, were substantially reduced as the USD/JPY FX rate continued to break record lows this year. Meanwhile, periodic samurai issuance put pressure in a negative direction to the spreads. Short-term and sharp widening in August and November is explained by the FX intervention by the MoF and BoJ, which created a significant number of short USD positions in markets. Our FX team estimates the amount of intervention to be around 4.5tn in August and 7.4tn in October. Some market participants covered those positions in the cross-currency basis market as well as FX markets. August, in particular, saw some investors spread tightening positions as a carry trade. The FX intervention triggered stop-loss activities on such positions, which accelerated further widening. In fact, the magnitude of widening was milder over October November. Probably a large part of speculative positions had been already cleared after the intervention in August. It is possible to estimate the break-even level of shortterm FX basis spreads. If a bank funds JPY cash at overnight rate and swaps it into USD by FX OIS basis swap, it effectively funds USD cash at USD OIS + X (Exhibit 26, left). Moreover, short-term FX LIBOR basis should move interactively with 3M LIBOR/OIS basis in each currency and FX OIS basis (Exhibit 26, right). Even if the overnight call market is not available, banks can resort to the BoJs liquidity providing operation which lends JPY cash at 0.1% p.a. for 3M or 6M terms in exchange for certain eligible collaterals. The cost involved in the above method should theoretically not exceed the cost via other dollar funding means. Currently, the BoJs dollar lending facility is in place. It lends unlimited USD at approximately USD overnight rate + 100bps in exchange for certain eligible
Exhibit 25: USD/JPY FX basis spreads gradually narrowed in 1H11. However, they widened back (moved ve) in 2H11
Cross currency basis spreads since 2011; bp
0 -10 -20 -30 -40 -50 -60 -70 -80 -90 -100 (bp) Apr-11 Jul-11 Oct-11 3M FX OIS basis 1Y FX LIBOR basis
Jan-11
Exhibit 26: (1) If a bank funds JPY cash at overnight rate and swaps it into USD by FX OIS basis swap, it effectively funds USD cash at USD OIS + X. (2) Short-term FX LIBOR basis should move interactively with 3M LIBOR/OIS basis in each currency and FX OIS basis. The cost involved in the above method should theoretically not exceed the cost via other dollar funding means
How to calculate the break-even level of short-term FX basis spreads
(1) Funding JPY from overnight market Fixing via JPY OIS Swapping JPY cash into USD by FX OIS basis JPY O/N JPY OIS - ( JPY OIS X) + USD OIS (2) Funding JPY from overnight market Converting it to JPY LIBOR via JPY LIBOR/OIS basis Swapping JPY cash into USD by FX LIBOR basis Convert it to USD OIS via USD LIBOR/OIS basis Net USD OIS + X Net JPY O/N - O/N + 3M JPY LIBOR- Y - (3M JPY LIBOR - Z) + 3M USD LIBOR - 3M USD LIBOR + USD OIS + W USD OIS - Y + Z + w
Exhibit 27: but 1Y FX OIS spread at the right is already below 100bp. The implied FX OIS USDJPY spreads exceed those seen in the actual FX OIS USD/JPY basis market.
Breakeven USD funding cost; bp
JPY 3M LIBOR/OIS basis (Y) 16.9 17.0 16.8 FX LIBOR USD/JPY basis (Z) 66.0 75.5 82.3 USD 3M LIBOR/FF basis (W) 49.3 58.9 66.2 JPY OIS + USD OIS + (Implied (-Y+Z+W) FX OIS) 109 -109 120 -119 124 -123 FX OIS USDJPY basis -110 -120 -125
1Y 2Y 3Y
Source: Bloomberg, J.P.Morgan
collaterals with the additional haircut of about 12.5% for equal to or less than the 1M term and 20% for more than 1M and equal to or less than the 3M term. Taking the cost of the haircut into account, the effective cost of USD funding via the BoJs dollar facility is higher than USD FF + 100bps. Therefore, FX OIS USD/JPY basis spreads
171
can go under -100. But the divergence should not be large, assuming that the two OIS rates are equal to the overnight rate in each currency and constant. In addition, if the spread charged on the dollar lending facility by the central banks is cut from the current 100bps, FX OIS USD/JPY basis spreads are expected to move in a positive direction accordingly, Exhibit 27 shows the current level of LIBOR/OIS basis spreads in each currency, FX LIBOR USD/JPY basis spreads and the FX OIS spreads implied from those numbers. 1Y FX OIS spread at the right is already below 100bp. Further, the implied FX OIS USDJPY basis spreads exceed those seen in the actual FX OIS USDJPY basis market. We have to be reminded that it starts from the assumption that a bank funds JPY at the overnight rate. In addition, it depends on many other variables such as how long the BoJs lending facilities (in JPY and USD) would be available, how much the collateral and additional haircut actually costs, and bid-offer. Lastly, there could be the possibility that markets overshoot at extreme stress, irrespective of these fundamentals. There are some implications from large negative USD/JPY cross-currency basis spreads. First, large negative basis spreads make it attractive for a USD investor to create synthetic USD bonds by buying JGBs and swapping them into USD with a cross-currency basis swap. The level of yield pick-up appears to at a historically attractive level (Exhibit 28). This applies even when compared with the difference of CDSs of the two countries (Exhibit 29). Although we need to bear in mind that CDSs of JGBs and USTs are mainly traded in USD and EUR, respectively, the comparison still gives us some guidance on the level of the pickups. Secondly, the return of FX-hedged foreign bonds has become less attractive (Exhibit 30). Some investors, when investing in foreign bonds, hedge their FX risk. They often use rolling 3M FX forward. As a result of sharp widening in the 3M term FX OIS basis spread, the cost of 3M FX hedge has become higher. If short-term basis spreads are stabilised at a large negative level, Japanese investors may accelerate the repatriation of USD funds back to JGBs. The focus is on life insurance companies because these investors would increase investment into super-long sectors with extra JPY cash. The return of FX-hedged 10-year USTs is now even lower than in August, when we saw 550bn net selling of medium- and long-term foreign bonds from life insurance companies, according to MoF data.
172
Exhibit 28: The level of yield pick-up appears to be at a historically attractive level on 1Y
Pick-up in yield of USD-asset-swapped 1Y JGB over 1Y USTs since 2010; bp
120 100 80 60 40 20 0 Jan-10 Jul-10 Jan-11 Jul-11
Exhibit 29: The level of yield pick-up appears to be attractive even when compared with the difference of CDSs of the two countries
Pick-up in yield of USD-asset-swapped JGBs over USTs; bp
JGB 1Y 2Y 3Y JGB ASW -25.5 -23.9 -20.6 JPY 3s/6s Basis 13.5 14.0 14.1 FX LIBOR USD/JPY basis -60.0 -70.0 -78.0 3M USD LIBOR + 48 61 73
Pickup vs UST 107 110 119
JGB/UST CDS spread 8 23 31
* For Exhibits 27&29, all the numbers are based on mid levels. For relevant basis spreads in each currency, cross-currency basis spreads, ASWs and CDSs, we used the nearest benchmark points. We calculated these numbers, with the assumption that USD/JPY conversion factors are the ratio of modified durations for simplification.
Exhibit 30: The return of FX-hedged foreign bonds has become less attractive with higher FX hedge cost, which could lead to further demand to the super-long sectors from life insurance companies.
20Y JGB yields and 10Y UST yields with FX hedge since 2010; %
3.50 3.00 2.50 2.00 1.50 1.00 0.50 Jan-11 Apr-11 20Y JGB
10Y UST hedged by 3M roll
Jul-11
Oct-11
FX hedge cost is calculated as 3-month forward point divided by spot FX rate Source: Bloomberg, J.P. Morgan
Volatility
The implied volatility surface as a whole was shifted downward with a steepening bias over the past year (Exhibit 31). Implied volatilities soared last December, as QE2 pushed up 10Y UST yields to above 3.50%. JGBs were sharply sold off and 10Y JGB yields moved higher from 0.85% in October to 1.3% in December. The selloff also triggered unwinding of short volatility positions, and volatilities spiked. Implied volatilities rose sharply soon after the 11 March earthquake. The BoJ provided abundant liquidity to ease the post-earthquake turmoil and maintained liquidity at a high level. Interest rates gradually declined thereafter. The 10Y swap rate and 10Yx10Y forward swap rate moved lower to 0.96% and 2.43%, respectively, from 1.21% and 2.67% at the beginning of this year. Steadily declining yields also lowered realised volatilities and then implied volatilities (Exhibit 32). From an investor flow perspective, first, the fall in the surface was led by constant selling flows in gamma sectors such as target buying and covered call from domestic investors. Second, issuance volume of structured notes remained small, but domestic banks kept selling callable structured deposit to mass retail. Banks generally sell options to hedge long vega positions. Their selling tends to take place in the middle of the surface as structured deposits often have a call option after 13 years and a maturity of 510 years. Such volatility selling flows are another factor for cheapening of the vega sector. On the other hand, there were sporadic buying flows in longer tail vega sectors from macro/leveraged investors. They bet on a long-term fiscal deficit deterioration scenario of the Japanese government. These flows pushed up and supported longer ends of the surface. Into 2012, we think that the current shape of the surface is likely to remain. We are neutral on gamma. Under a low-for-long regime, short-term interest rates will likely remain low and move within a narrow range and volatility selling flows are expected to continue. On the other hand, we cannot discount the possibility of a rise in gamma sector volatilities under a stressed scenario, given the situation in Europe. In long-tail vega sectors, we expect volatilities to remain relatively high. With absolutely low interest rates and volatilities in the JPY market, we do not expect selling demand to outweigh buying demand in these sectors. Rather, since an issuance increase of JGBs continues to be a major topic of debate, concerns about the long-term fiscal deficit of Japan are likely to be priced in and out from time to time.
Exhibit 31: The implied volatility surface as a whole was shifted downward with a steepening bias over the past year.
Implied volatility changes over the past year; bp/day
1Y tail 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 -1.4 -1.6 -1.8 3M
2Y tail
5Y tail
10Y tail
20Y tail
1Y
2Y
5Y
10Y
Exhibit 32: Implied volatilities soared last December as QE2 pushed up 10Y UST yields to above 3.50% They also rose sharply soon after the earthquake in March before declining later in the year.
1Yx10Y normal implied volatility since 2010; bp/day
5.0 4.5 4.0 3.5 3.0 2.5 2.0 Jan-10 Jul-10 Jan-11 Jul-11 UST Yield Rise after QE2 Earthquake
Exhibit 33: TIBOR remained sticky to fall relative to CP rate

TIBOR 3M and a-1 rated CP rate; %
0.60 0.50 0.40 0.30 0.20 0.10 0.00 Nov -09 May -10 Nov -10 May -11 Nov -11 CP rate a-1 3M D Tibor
Source: Bloomberg, JASDEC, J.P. Morgan

173
TIBOR/LIBOR Spread
There are three categories of TIBOR-LIBOR spreads: 1) spot fixing TIBOR-LIBOR spread; 2) IMM 3M TIBORLIBOR spreads whose liquidity runs up to the 8th contract; and 3) Term TIBOR/LIBOR spreads which are the difference between swap rates usually against TIBOR 6M and LIBOR 6M, but for short maturities also against TIBOR 1M and LIBOR 1M. TIBOR fixing remained sticky to fall while CP rates were stable at a low level (Exhibit 33). We think the reason for the stickiness has remained unchanged over the past few years, i.e. market practice is that TIBOR is used as a reference rate to corporate lending. As the credit spread charged to domestic companies continued to be tight (Exhibit 34), TIBOR tends to be maintained at a high level relative to the policy rate. As a result, the 3M TIBOR/LIBOR fixing spread remains positive. There were two periods in the past when TIBOR/LIBOR fixing spread was negative (Exhibit 35). One was just after the end of the BoJs quantitative easing in 2006 when LIBOR priced in the future rate hike earlier, and the other was after the subprime shock in summer 2007 as LIBOR jumped. We think that the former case is not impossible, but very unlikely. Since the Fed introduced an explicit time frame of interest-rate policy until mid2013 in August, the BoJ is expected to continue the current interest-rate policy even longer than the Fed. The latter case is more likely given the current situation in Europe. But even if it happens, the magnitude may be milder as central banks are committed to avoiding a liquidity crisis. With regards to term TIBOR-LIBOR, we think that negative spreads are unlikely unless the spot fixing TIBOR-LIBOR spread becomes negative. Short-term spreads are generally driven by the spot fixing spread. For longer tenors, the spreads are largely driven by domestic banks hedging activities. They use TIBOR as their internal pricing and TIBOR swaps fit the purpose. However, as LIBOR swaps are more liquid, they often use LIBOR swap and hedge the TIBOR/LIBOR risk from time to time. Their choice of hedge timing is the source of term TIBOR-LIBOR fluctuation. However, term TIBOR-LIBOR spreads generally moved in a very tight range with limited flows and a sticky spot-fixing spread this year. (Exhibit 36) Looking ahead, we have a tightening bias on the TIBOR-LIBOR fixing spread. Historically, LIBOR
174
Exhibit 34: As the credit spread charged to domestic companies continued to be tight, TIBOR tends to be maintained at a high level relative to the policy rate.
Average rate charged to corporate in bank lending; %
2.00 1.80 1.60 1.40 1.20 1.00 2004 2005 2006 2007 2008 2009 2010 2011
Source: BoJ, J.P. Morgan
Exhibit 35: LIBOR fixing was higher than TIBOR fixing at the end of QE in 2006 and after the subprime shock in 2007
3M TIBOR and LIBOR fixing; %
1.20 1.00 0.80 0.60 0.40 0.20 0.00 2005 2006 2007 2008 2009 2010 2011 3M LIBOR 3M Euroy en TIBOR
Source: Bloomberg
Exhibit 36: Term TIBOR-LIBOR spreads generally moved in a very tight range with limited flows and sticky spot-fixing spread this year
5Y term TIBOR/LIBOR spread and spot-fixing spread; bp
12 10 8 6 4 2 0 -2 -4 Jan-10
Source: Bloomberg
TIBOR-LIBOR fix ing spread
5Y Term T/L
Jul-10
Jan-11
Jul-11
tended to move faster and be more volatile both in a rise and in a fall. Therefore, the TIBOR-LIBOR fixing spread has had a tendency to narrow in a rise and to widen in a fall. As mentioned in the previous sections, there are some risk scenarios. One is a fall of LIBOR on a cut in IOER. Another scenario is that LIBOR rises under extreme stress, which we think is more likely. Besides, the level of LIBOR is absolutely low and lower than TIBOR. The magnitude of the spread move is likely to be asymmetric, i.e. we expect more tightening if LIBOR increases than widening if LIBOR falls.
Exhibit 37: Term LIBOR 3s/6s spread curve widened and steepened this year
Spot fixing 3s/6s and 5Y and 20Y term 3s/6s basis; bp
30
Spot fix ing 3s/6s LIBOR 5Y term 3s/6s 20Y term 3s/6s
25
20
15
3s/6s spread
The Term LIBOR 3s/6s spread curve steepened this year (Exhibit 37). The spreads moved in a parallel fashion until 2010. However, although the spreads widened, the widening was larger at the longer ends. Widening can be explained by one-sided demand in the 3s/6s market. With relatively tight credit spreads in Japan, samurai issuance continued in 2011. When a non-resident entity issues JPY bonds (e.g. EMTN or samurai) and swaps into USD, it enters a paying position of 3s/6s basis swap (pay 3M vs. rec 6M) as well as a receiving position in a crosscurrency swap (Exhibit 38), driving 3s/6s basis higher and FX basis wider (more ve). Such flows tended to take place in 510Y sectors. Conversely, if a nonresident investor buys a JPY bond with a cross-currency swap as a synthetic USD bond, it has the opposite impact on the spreads. However, there was little such demand this year. Rather, those investors who had such positions, particularly in the super-long sectors, might have unwound them for the purpose of reducing balance sheets. This could be a reason for the widening on longer term 3s/6s spreads. We think the short-to medium-term spreads are biased wider. Tight credit spreads in Japan are still attractive for non-resident entities to raise funding, even after large negative cross-currency basis spreads and extra premium for non-resident entities are considered. Issuance by non-residents is expected to continue in 2012 and such flows would be a widening factor in the medium-term spreads. There is also the possibility that LIBOR will rise if the crisis in Europe deepens. If the scenario realises, the spot fixing 3s/6s spread will widen as it had amid the Lehman shock and lead to a widening of term 3s/6s spreads at the short ends.
10 Jan-10 Jul-10 Jan-11 Jul-11
Exhibit 38: Flows of non-resident entity issuing JPY bonds and swapping into USD drive 3s/6s basis higher and FX basis wider (more ve)
How 3s/6s is involved in swapping JPY into USD
3s / 6s marktet JPY 6M FX basis marktet USD 3M L Issuer JPY 3M L + FX LIBOR basis JPY 6M JPY Fixed JPY Fixed Samurai/EMTN JPY 3M L + 3s/6s basis
Swap marktet
175
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Jorge GarayoAC (44-20) 7325-4820 jorge.garayo@jpmorgan.com J.P. Morgan Securities Ltd
Inflation-linked Markets
2011: Linkers underperformed as nominal yields declined substantially on the back of an escalation of the sovereign risk crisis in the Euro area and a deteriorating global macro outlook 2012 outlook for inflation: Substantial falls in headline and core inflation across the board We believe that the market will focus more on downside risks from a deeper recession than on medium-term inflation, as inflation expectations remain anchored German and UK 10Y real yields to test new lows and fall well below zero in Europe but expected to stay close to 0% in 10Y TIPS while breakevens should be biased lower as the sovereign crisis continues in 1H12 Expect cash breakevens to underperform relative to inflation swaps in early 2012 Euro area: Position for flatter breakeven curve, as inflation should be sticky and longer-dated breakevens remain highly directional UK strategy: Be long intermediate real yields, and short inflation breakevens in the 10Y sector UK tactic: Go short UK real yields into linker syndications as a tactical trading strategy US: Fair-value model suggests falls in inflation and nominal yields, combined with fiscal tightening, should cause breakevens to narrow into early 2012 Beyond 1Q12 longer maturity TIPS breakevens should widen as nominal rates rise Supply: UK and US to see modest increases in gross supply, whereas Euro area issuance should be below average, with Italy being the main question mark
Exhibit 1: Inflation breakevens were supported by energy price inflation in 1Q11 but started to decline in the summer
Inflation breakevens for 10Y benchmarks in US TIPS, Germany and the UK; %
3.50 3.00 2.50 2.00 1.50 1.00 Nov 10 Feb11
UK
Germany
US
May 11
Aug11
Nov 11
Exhibit 2: Headline inflation rose steadily over the past year, with inflation in 1Q11 being lifted by energy prices
Inflation measures, in oya terms; %
6.00 4.00 2.00 0.00 -2.00 -4.00 Nov -06
Euro HICP
US CPI
UK RPI
Nov -07
Nov -08
Nov -09
Nov -10
Nov -11
Exhibit 3: and core inflation also moving up significantly over the period
Core inflation measures, in oya terms; %
4.00
HICP core
US CPI core
UK CPI core
3.00
2.00
2011: Linkers outpaced by nominals in a relatively high inflation environment

2011 was dominated by large declines in nominal yields, as the European sovereign debt crisis accelerated from the summer. Throughout the year, market-implied inflation breakevens were initially driven by a spurt of
176
1.00
0.00 Nov 06 Nov 07 Nov 08 Nov 09 Nov 10 Nov 11
energy price inflation in 1Q11, but they entered a clear downward phase from May onwards (Exhibit 1). From this point, inflation breakevens were driven lower by a prevailing risk-off environment and worries of a worldwide recession as data printed particularly weak globally. The evolution of inflation prints would suggest a positive return year for inflation-linked markets. Headline inflation rose steadily since November 2010, peaking at the end of the year (Exhibit 2). Even though energy inflation played a significant part in 1Q11, core inflation also rose throughout the year (Exhibit 3). This was contrary to our January predictions that a high degree of slack would push core inflation lower (with the caveat that the move in UK core CPI was heavily impacted by VAT effects). However, inflation prints were not the main driver and inflation expectations were pushed lower by the macro landscape. The downgrading of growth expectations in the Euro area and globally from the summer was likely exacerbated by sovereign risk dynamics in the periphery. Commodities reversed their upward trend, and equity markets fell over 15% in the space of two weeks in August. Central banks recognised a shift in the macro fundamentals and risks around them, dismissing the contemporaneous relatively high inflation prints. The ECB eased rates in November after having tightened policy in April and July, whereas the Fed committed to keeping rates unchanged up to mid2013 and initiated Operation Twist. The BoE announced another 75bn of QE purchases and is expected to announce another round of QE in the coming months. Survey-based measures of medium-term inflation expectations remained well anchored throughout the year despite the significant deterioration in the macro outlook. Inflation expectations from 2013 onwards were broadly unchanged according to economists consensus forecasts (Exhibit 4), supporting the message of anchored inflation expectations that central banks have been highlighting thus far. So how did inflation-linked portfolios fare relative to conventional government bonds? Their returns (adjusted for duration differences) were negative in the Euro area and the UK and just about matched those of Treasuries in the US (Exhibit 5). In other words, a strategy of being long inflation breakevens across the curve would have given investors a return of -3.3% and
Exhibit 4: Medium-term inflation expectations are well anchored according to economists surveys, with the Euro area and the US likely close to target and UK CPI expected to be 0.8% above target
Long term inflation forecasts from economists*; % US CPI UK CPI
Euro HICP
Oct10 Oct11 Chg Oct10 Oct11 Chg Oct10 Oct11 Chg 2011 2012 2013 2014 2015 2016 2017-21 1.9 2.2 2.1 2.2 2.2 2.3 2.3 3.1 2.1 2.1 2.3 2.3 2.2 2.2 1.2 -0.1 0.0 0.1 0.1 -0.1 -0.1 2.6 2.1 2.6 3.0 2.6 2.7 2.7 4.4 2.7 2.5 2.7 2.6 2.8 2.8 1.8 0.6 -0.1 -0.3 0.0 0.1 0.1 1.6 1.5 1.7 1.9 2.0 2.1 2.1 2.6 1.8 1.8 1.9 2.0 2.0 2.1 1.0 0.3 0.1 0.0 0.0 -0.1 0.0
* Source: Consensus Economics, October 2011 long-term forecasts.
Exhibit 5: Despite the upward trend in inflation in 2011, buying European linkers vs. conventional bonds would have been unprofitable, with Italian linkers being the main drag for the Euro area; the same strategy in the US would have posted returns close to zero
Total return between 15 November 2010 and 18 November 2011; % Total return
Euro linkers (JPM ELSI) Euro gov comparator* portfolio Long inflation B/E (JPM EBEX) Euro linkers ex Italy ex Greece Italian linkers US TIPS (JPM JUSTINE) USTs comparator* portfolio Long inflation B/E (JPM UBEX) UK Linkers (FTSE) Gilts 7Y+ (JPM GBI) Long inflation B/E (est.**)
-10.7% -5.8% -3.3% -2.7% -23.5% 12.5% 10.4% 0.0% 18.7% 20.4% -5.7%
* Comparator portfolios are based on the same weights used by linker indices for comparable nominal bonds, adjusting for the relative duration. ** Calculated by replicating a strategy of being long FTSE All-Share index linked against being short the JPM Gilts 7Y+ index in a duration-adjusted amount, reweighted weekly.
Exhibit 6: The deterioration in liquidity was an important driver of performance. Our estimated bid-offers widened to levels above those seen in 4Q08 for Italian linkers, and to similar levels in French linkers whereas other linker markets and inflation swaps bid-offers held in remarkably well
Indicative bid/offers across different inflation products*; bp of yield Assumptions 2008 2011
Prior Italy linkers France linkers German linkers Euro HICP sw aps UK linkers UK RPI sw aps US TIPS US CPI sw aps
* Source: J.P. Morgan
4Q 08 1H 11 2H 11 15 15 15 15 7 18 7.5 15 2.5 2.5 2.5 2 2 4 3 5 20 15 10 4 2 4 4 5 50mn size 25-50mn size 25-50mn size 50mn size 10mn size 25mn size $50mn size $25-50mn size
177
2 2 2 2 2 4 2.5 5
-5.7% in the Euro area and the UK, respectively, and a flat return in the US. As in 2008, the deterioration in linkers liquidity was an important driver of performance, particularly in the Euro area. Bid/offers reached similar or larger dislocations than those seen in 2008. In particular Italian BTPeis bid/offers widened on average 20bp vs. 15bp at their worst point in 2008. But even bid-offers for French linkers widened to levels reached in late 2008 (Exhibit 6). Supply in Euro area linkers was heavily affected by the sovereign debt dynamics. Germany cancelled its linker programme in 2Q11 and 3Q11, not tapping the market between April and November. Italys net issuance of linkers in 2H11 was just around 1.5bn (vs. usually 24bn per quarter). Unstable market conditions resulted in reduced auctions, and the Treasury conducted buybacks of heavily dislocated lines. It is interesting to note that US and UK linkers, as well as inflation swaps, held up much better in terms of bidoffers, despite some evident deterioration in liquidity dynamics.
Exhibit 7: J.P. Morgan 2012 inflation forecasts: core measures to decline across the board; headline inflation to fall but average above target in the UK, around target in the Euro area, and well below target in the US
US= US CPI-U nsa and core CPI, Euro = Euro HICP all items, core HICP (excluding all food and energy), UK= UK RPI, CPI and core CPI; oya% US Euro UK
CPI Oct 11 Nov 11 Dec 11 Jan 12 Feb 12 Mar 12 Apr 12 May 12 Jun 12 Jul 12 Aug 12 Sep 12 Oct 12 Nov 12 Dec 12 Av e 2012 3.5 3.5 3.1 2.7 2.4 1.6 1.2 1.0 1.3 1.2 1.0 0.9 1.1 1.2 1.3 1.4
Core 2.1 2.1 2.2 2.1 1.9 1.9 1.7 1.5 1.3 1.1 1.0 1.1 1.1 1.1 1.2 1.4
HICP 3.0 3.0 2.8 2.5 2.4 2.0 1.8 1.9 2.0 1.9 1.9 1.7 1.5 1.5 1.5 1.9
Core 1.6 1.6 1.7 1.7 1.7 1.6 1.5 1.6 1.5 1.6 1.5 1.4 1.4 1.5 1.5 1.5
RPI 5.4 5.2 4.9 4.4 3.7 3.5 3.0 2.9 3.1 3.2 2.9 2.6 2.7 2.6 2.6 3.1
CPI 5.0 4.8 4.2 3.6 3.2 3.2 2.5 2.5 2.7 2.5 2.3 1.9 2.0 1.9 1.9 2.5
Core 3.4 3.2 3.0 2.6 2.1 2.1 1.5 1.6 2.0 1.9 1.6 1.6 1.6 1.5 1.5 1.8
Inflation outlook for 2012

The inflation outlook for 2012 can be characterised by an expected decline in both core and headline inflation (Exhibit 7). In terms of headline inflation, we expect more significant declines, in great part because of energy and food base effects from substantial price increases in the past year. In sequential terms (seasonally adjusted q/q changes in CPI indices), we have already seen a clear declining trend in Euro area headline inflation, and we expect significant falls in the UK and the US in the coming quarters (Exhibit 8). Euro area inflation should drop to an average of 1.9% oya during 2012, in line with the ECBs target. In the UK, we expect headline inflation to fall sharply but average 2.5% (above the BoEs target of 2%) whereas we see US inflation averaging 1.4%. The assumptions behind our forecasts are as follows: Stable oil prices, assuming Brent around $110/bbl, most likely stuck in a $100115/bbl range. We see OPEC supporting prices above $100 (barring a very significant recession, which could push it below that figure), while marginally supportive fundamentals could push it to $115/bbl.
Exhibit 8: Euro area sequential inflation has already initiated a downward trend and should stabilise around 1%; we expect significant falls in the US and the UK in the next few quarters
Euro area HICP inflation Q/Q saar, US CPI inflation Q/Q saar and UK CPI Q/Q saar, actual and J.P. Morgan forecasts; %
7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 -1.00 -2.00 Mar 09 Sep 09 Mar 10 Sep 10 Mar 11 Sep 11 Mar 12 Sep 12 US Euro area Forecasts UK
178
Stable currencies in the major crosses, in line with their recent behaviour this year 50% chance of a technical recession in the UK (base case is one-quarter of GDP contraction, annualised 0.5% growth in 2012), a moderate recession in the Euro area (peak-to-trough contraction in GDP just over 1%) and the US growing at sub-par pace of 1.8% in 2012.
Exhibit 9: Our inflation view is sensitive to assumptions about volatile elements in baskets and model risk for core inflation; we see downside risks coming from a deeper recession (lower oil) and upside risks from tax increases (VAT) or currency depreciation
Impact on headline inflation; % points
10% 10% mov e in oil price^ Euro HICP UK RPI US CPI 0.10 0.20 0.50 $100-120 J.P. Morgan v iew range in a deeper recession Near term to food prices Dow nside dow nside mov e in food prices^ 0.15 0.10 0.15 10% drop in currency *^ 0.25 0.65 0.25 Core models error +/- 0.7 +/- 0.7 +/- 0.5 Risk of deeper recession / Misjudge slack
State controlled prices / VAT increase** 0.25
One of the main forces that poses significant downside risks to inflation is the prospect of a more severe and prolonged recession in the Euro area, which could significantly impact the region and the rest of the world. A global spillover could potentially bring oil prices well below the $100/bbl level and push headline inflation two to three tenths lower, purely through energy components, as well as impact long-term inflation expectations. This is not our central scenario, however, and there are also upside risks if core inflation remains sticky on the way down if we misjudge the slack in the economy. As always, the behaviour of currencies could also have a significant impact on imported inflation. In the Euro area, ongoing fiscal retrenchment in key countries may prove inflationary over the near term, as governments are likely to increase taxes for the consumer (in the form of VAT, administered prices, and other duties). Our forecast does not incorporate any of these factors, and therefore we should be conscious of them. We show a summary of some of these factors sensitivities in Exhibit 9. We also discuss some of these risks in some detail in the individual market sections.
Central v iew of stable currencies EUR at risk
Most likely increases in Euro area
(-)
(-)
(+)
(-) / (+)
(+)
* Based on drop in Trade Weighted currency. ** Euro area based on combined 1% moves in Italy, Spain and France, assuming a 65% pass through on 65% of baskets, for more granular information see Exhibit 26. ^ For simplicity we assume broadly equal betas on the downside and upside. (-) Downward impact on inflation (+) Upward impact on inflation.
Exhibit 10: Investors expect central banks to do a good job at avoiding deflation; the Fed and the BoE are seen as giving significant weight to growth, while the ECB is more inflation-oriented but remains the most credible in terms of keeping inflation close to target
Investor perception of weights (median response) given to growth and inflation for each central bank*; %
Grow th / employ ment 100%
Inflation
Are we approaching deflation? Not quite

The quick deterioration in the macro data and a gloomy outlook for coming years has raised questions about the possibility of entering a very low inflation world. In our view, a zero-inflation world would come with a significant decline in medium term inflation expectations. We are far from that, as inflation expectations have remained firmly anchored in the three main markets. The prospect of a global recession is pushing expectations lower in the Euro area, but not to a worrying degree.
20 80% 60% 40% 20% 0% US Euro area 80 35 65
30
70
UK
* See: J.P. Morgan inflation expectations survey, November 2011 by Garayo & Chordia
179
Exhibit 11: Inflation expectations 1 year out do not point to immediate disinflation; investors were likely much more worried about low inflation in July 2010 than in our latest survey
Percentage responses to the question: Where do you expect core inflation to be a year from today? from the J.P. Morgan inflation expectations client survey, November 2011 and previous surveys (from July 2010); % applies to responses from investors across different asset classes US Euro area UK
Nov 2011 survey Below 0.5% 0.5-1.5% 1.5-2.5% 2.5%+ July 2010 survey Below 0.5% 0.5-1.5% 1.5-2.5% 2.5%+ Mean (Nov 11) Chg v s. prev Mean (Jul 11) Mean (Mar 11) Mean (Nov 10) Mean (Jul 10) 20% 53% 24% 4% 1.86 -0.01 1.87 1.80 1.40 1.10 22% 55% 77% 20% 4% 1.61 -0.16 1.77 1.80 1.35 1.00 39% 18% 12% 30% 57% 2.34 -0.06 2.40 2.40 1.90 1.60 4% 24% 58% 15% 6% 42% 48% 38% 14% 37% 47% 4% 13% 84%
Exhibit 12: The Euro area shows the highest probability of deflation among investors, but it is still counterbalanced by an equal probability of high inflation; probabilities remain heavily skewed to the upside in the UK and the US
% probability; J.P. Morgan inflation expectations client survey* November 2011
35 30 25 20 15 10 5 0 US
Prob of deflation
Prob of high inflation Jul 11
Nov 11 surv ey
surv ey
Euro
UK
US
Euro
UK
* Probability perception of deflation/high inflation as per the question below: What is, in your view, the likelihood of getting deflation (CPI annual inflation below 0%) or high inflation (CPI annual inflation above 4%) for two years out of the next five years in each of these regions?.
Exhibit 13: The implied probability of deflation in 2 or 5 years time is not significantly different from that of inflation being above 4%; the UK prices in high probability of high inflation in 5 years time; over the near term US is pricing a relatively higher probability of deflation
% probability of inflation <0% or above 4% in 2 and 5 years time* Prob Inf<0% p.a. Prob Inf>4% p.a. 5Y ZC (bp of notional)
With central banks watching these carefully, there is no reason to believe in a changing inflation environment, which would only be consistent with the world economy sinking into a prolonged depression (not our current expectation). As we showed in our J.P. Morgan November Inflation Expectations Survey (11 November 2011), an overwhelming 45% of investors expect inflation to be close to target in the Euro area over the medium term. In the case of the BoE and the Fed, they have both been extremely vocal about the risks of low inflation and continue to be seen as giving a lot of weight to growth at the current cross roads (Exhibit 10). Inflation expectations 1 year out remain remarkably high according to our latest Inflation Expectations survey (November 2011), even if there has been a shift lower in the Euro area (Exhibit 11). Our own expectation is that Euro area inflation will average 1.9% in 2012, which is likely to keep expectations relatively high. Furthermore, the risks of low inflation over the medium term seem to be pretty much counterbalanced by perceived risks of very high inflation. This is most clear in the US and the UK, where investors attach a particularly high probability of inflation being above 4% in 2 out of the next 5 years (Exhibit 12). Admittedly, investors perceived probability of a deflation in the Euro area has
180
2Y Euro HICP UK** US 8.5% 10.7% 14.3%
5Y 12.8% 10.9% 14.6%
2Y 5.2% 8.9% 12.4%
5Y 11.8% 35.9% 23.5%
0% floor 35 52 39
4% cap 44 67 74
* Probability is derived from pricing caps and floors spreads using the J.P. Morgan vol surface calibrated with current zero coupon style caps and floors as well as y/ y caps and floors. ** For UK RPI, we report the probability of inflation > 4.8% and <0.8%, roughly equivalent to UK CPI > 4% or <0% p.a. if we applied the historical spread between RPI and CPI of 0.8%.
increased recently (to 15%), but this is counterbalanced by an equal probability of very high inflation. Looking at inflation volatility markets, the implied probability of deflation in 2 or 5 years time is not significantly different from that of inflation being above 4% in our survey (Exhibit 13).
Euro area market dynamics Sovereign risk to dominate inflation breakevens in 2012
Euro area breakevens have been intrinsically linked to the sovereign risk dynamics during 2011, with a very clear correlation between the risk-off/risk-on dynamics and linker breakeven performance (Exhibit 14). This dynamic is unlikely to change in 2012 and the performance of peripheral spreads is likely to continue to dominate while inflation fundamentals will take the back seat. Italian linkers to continue to feel the pressure The main question is whether the underperformance of BTPei linkers vs. BTPs will continue, and whether it will spread to French linkers vs. OATs as we have seen recently. The large declines in Italian BTPei breakevens relative to inflation swaps have been significantly larger than what we saw in the 2008 financial crisis, while breakevens for French OATeis have already moved by a similar degree vs. 2008 (Exhibit 15). This has led to negative breakevens in Italian linkers out to 2019 and a complete decoupling from inflation expectations. Our central view is that Italian breakevens will continue to underperform as spreads widen further. There are risks that Italy may be downgraded and if this is severe enough it could impact inflation indices. Based on our understanding of current index rules of one of the most widely followed benchmarks, a downgrade to BBB will result in investors reallocating BTPei holdings (Exhibit 16). These numbers are based on our estimate that 2030% of investors are benchmarked, holding around 20bn of the 77bn of current market value. With such a large percentage of the market potentially being dropped from indices, we believe other nonbenchmarked investors may also be reluctant to hold onto their BTPei investments, resulting in underperformance ahead of the event and exacerbating the impact. We note the following could well limit any potential impact to Italian linkers: 1) The index rules could be changed if a majority of investors consider them to be very harmful for the market.
Exhibit 14: Inflation breakevens have been extremely correlated with equity markets over the past year, rising in risk-on and falling in riskoff environments
Euro Stoxx 50 and DBRi 2020 inflation breakevens vs. risk-on and risk-off *; Index points %
3200 3000 2800 2600 2400 2200 on on off 2000 1800 Nov10 10Y B/E Euro STOXX 50
2.40 2.20 2.00 1.80 1.60 1.40 1.20 Nov11 off
off
Jan11
Mar11
May11
Jul11
Sep11
* Risk-on and Risk-off periods defined with reference to sustained trends in Euro STOXX 50 index See European Derivatives for definition.
Exhibit 15: BTPei linker breakevens have underperformed vs. HICP swaps by more than twice as much as seen in late 2008; French breakevens have already moved by a similar degree as in 2008
BTPei 2019 breakeven minus HICP swaps, OATei 2020 breakevens minus HICP swaps; bp
0 -50 -100 -150 -200 -250 2008 2009
OATei 20 b/e minus HICP sw aps
Financial crisis late 2008 2010
BTPei 19 b/e minus HICP sw aps
2011
Exhibit 16: In the event that Italian linkers leave linker indices, we would see a reallocation out of BTPeis by benchmarked investors; France should be the main beneficiary in terms of inflows, given that it constitutes 54% of Euro benchmarks
Relative market value of constituents of Euro linker indices and expected activity from benchmarked investors; bn Euro linker index * Index ex Italy Index ed activ ity **
MV Italy France () France (F) Germany Total 77.1 74.4 75.3 52.5 279
% split 28% 27% 27% 19% 100.0%
MV ex Ita 0% 37% 37% 26%
Chg -28% 10% 10% 7%
20% -15.4 5.7 5.7 4.0
on
30% -23.1 8.5 8.6 6.0
* Based on Euro area government linkers with maturity over 1Y excluding Greece ** Expected selling/buying under two assumptions for percentage held by benchmarked investors in bn. 181 () Linked to Euro HICP ex tobacco (F) Linked to French CPI ex tobacco.
2) Investors could switch to a new linker benchmark that includes Italy regardless of the rating. This is more of an administrative process, and would only be likely adopted by Italian domestic investors (which we estimate own around 50% of the outstandings). Taking all risks into account, we see a risk that Italian real yields go significantly higher. If nominal yield spreads were to widen by 250bp (broadly consistent with our projection of 7.75% for 10Y Italy-Germany spread, see Euro cash), Italian breakevens could trade very negative out to the 10Y sector, with real yields between 9.50% and 11% in the sub-10Y sector possible. Our model uses recent betas of breakevens to peripheral spreads, which are then adjusted according to dirty price dynamics of linkers. Linkers whose dirty prices reached levels of 6070 have seen buying flows, resulting in a better performance in breakeven terms (as has been the case in BTPei 2041 and BTPei 2035).Exhibit 17 shows the moves in inflation breakevens during the first two weeks of November vs. the dirty prices of linkers, we think this behaviour could continue. The beta of Italian breakevens to peripheral spreads has been approximately 40% out to the 15Y sector over the past 6 months (i.e 4bp of breakeven falls for every 10bp of widening of spreads, Exhibit 18) and around 25% for the very long end. Two factors drove the BTPei real yield underperformance relative to nominal BTPs. First, linker real yields rose reflecting liquidity concerns and balance sheet de-leveraging, and, second, the ECB bought conventional BTPs in its SMP, focused on maturities out to the 10Y sector of the curve. While the illiquidity factor is unlikely to result in further rises in real yields, the SMP has continued to expand, and we expect increased buying volumes in BTPs. We see a very small probability that the SMP starts to buy linkers or BTPs at the very long end of the curve, so we would continue to expect high betas with nominals in sub-10Y breakevens. Another factor we must take into account is the deflation floor embedded in the principal of BTPeis, particularly relevant in a negative inflation breakeven environment. The fact that principal payments of linkers cannot go below par creates a theoretical floor for breakevens. This floor is relatively close for the BTPei September 2016, which only has around 3% of accrued inflation but reached the breakeven lows of around -0.50% in mid-November. We can calculate the breakeven level which is consistent with the principal hitting the deflation floor, which is somewhere between
182
Exhibit 17: Italian linkers with a high dirty price saw significant declines in breakevens in November, whereas those with low dirty prices outperformed We must allow for this when projecting performance going forward
Inflation breakeven change 1 November18 November vs. cash dirty price of BTPei linkers on 1 November; bp
100 50 B/E change 0 -50 -100 -150 60
41
35
26 23 21
1916 17
14
70
80 90 100 Dirty price, points
110
120
Exhibit 18: BTPei breakevens have declined 4bp per every 10bp widening in the nominal yield spread to Germany and are likely to decline further given our view on peripheral spreads
BTPei 2021 breakeven vs. nominal yield spread in 10Y nominal benchmarks over the past 6M; %
2.5 2.0 1.5 1.0 0.5 0.0 1.0
y = -0.43x + 2.79 R2 = 79%
BTPei 21 breakeven
2.0 3.0 4.0 5.0 Italy -Germany nom y ld spread, %
6.0
-0.60% and -0.70% (which on a compounded basis would completely offset the 3% of inflation accrual). Exhibit 19 projects the nominal cash flows and future HICP index levels applied to BTPei 2016 using the current inflation breakeven expectation, as well as under the scenario where we reach the threshold at which the principal floor kicks in. In this scenario, the final HICP index level expected is at the same level as the base CPI value of 109.5227. Putting all the information together, we have modeled the expected breakeven moves for all BTPeis under a significant increase in BTP nominal yield spreads to Bunds (ranging from 220bp to 290bp) in Exhibit 20. Our
beta for breakevens is based on the factors highlighted above. The BTPei 2016 is expected to hit the threshold breakeven of -0.65%, at which point the deflation floor on the principal starts to kick in, and therefore outperforms other lines. BTPei 14s and BTPei 17s would suffer most under our assumption while we would expect more limited falls in breakevens at the very long end of the curve. While we have not modeled alternative widening scenarios, we should expect broadly proportional relative performance. Clearly a scenario where Italian spreads tighten (unlikely in our view) would result in normalisation in Italian real yields and significant performance, both in absolute and relative terms. French linkers should prove more resilient than Italian to a worsening of the sovereign risk dynamics Is France likely to follow the footsteps of Italian linkers? As Italian and Spanish yields spreads widened significantly, French lines started to underperform in November as France widened vs. Germany. However, we believe there are a series of factors that put French linkers in a better position than Italian ones. Sales of Italian linkers (28% of the linker market) could result in switches into French (54% of the market if we include both French and Euro HICP-linked lines). With German linkers only around 20% of the market and trading at negative real yields, French linkers should be reasonably attractive alternatives. Looking at the relative weights of linkers within indices, the bonds that should attract more
Exhibit 19: The breakeven floor-threshold for BTPei is not far from current levels given current market volatility; breakevens should not go below -0.66% as this would imply hitting the floor on the principal
Cash flow calculation under current breakeven assumption and threshold assumption for BTPei Sep 2016 (Base CPI value 109.5227); Nominal cash flows and Index points
Date 21/11/2011 15/03/2012 17/09/2012 15/03/2013 16/09/2013 17/03/2014 15/09/2014 16/03/2015 15/09/2015 15/03/2016 15/09/2016 IRR Current B/E= -0.20%p.a. Threshold= -0.66% p.a. Cash Flows HICP index Cash Flows HICP index -83.988 1.083 1.082 1.081 1.079 1.078 1.077 1.075 1.074 1.073 103.088 6.64% 112.796667 112.999 112.855 112.715 112.574 112.432 112.293 112.152 112.013 111.872 111.732 -82.176 1.082 1.078 1.075 1.071 1.068 1.064 1.061 1.057 1.054 101.050 6.71% 112.79667 112.863 112.482 112.111 111.736 111.361 110.996 110.623 110.258 109.890 109.5227
inflows are (in this order) OATei 20, OATi 17, DBRi 16 & DBRi 20, and OATi 13. We expect domestic demand to remain strong on the back of French CPI hedging (at least in 1Q12) and interest from Libor-based domestic investors to resurface. Inflows into Livret A accounts have been particularly strong in 2011 (Exhibit 21), and we predict the Livret A rate will increase to 2.50% in February 2012 (before falling back to 1.75% later in the year, see Exhibit 22). This means that we are likely to see some French CPI hedging demand in the early stages of 2012. Domestic Libor-based demand for linkers should provide some support to the market in the event of significant
Exhibit 20: Possible scenario for BTPei breakevens if BTP nominal yield spreads widened by an average 250bp
Current linker metrics for Italian BTPeis and projected levels under nominal yield widening assumptions
Nom Yld Beta real Dirty Line 15/09/2016 15/09/2017 15/09/2019 15/09/2021 15/09/2023 15/09/2026 15/09/2035 15/09/2041 price 82.58 78.19 70.96 68.29 63.74 Cum Infl 3% 6% 4% 2% 5% Real Yld B/E Infl Threshold w idening to nom B/E* -5.51% -0.61% -1.90% -0.80% -0.42% -0.79% -0.13% -0.58% -0.15% assume 288 275 260 250 250 240 230 220 220 y lds 1.40 1.10 1.36 1.24 1.24 1.14 1.12 1.10 1.08 Beta to B/E 40% 24% 36% 24% 24% 14% 12% 10% 8% Ex pected Ex pected Ex pected Real Yld 11.1% 10.1% 10.7% 10.0% 9.7% 9.5% 9.3% 7.7% 7.5% B/E Infl -1.64% -0.61% -1.23% -0.49% -0.29% 0.26% 0.47% 2.14% 2.04% Mov e (bp) -115 -27 -94 -60 -60 -34 -28 -22 -18
15/09/2014 103.23 17% 7.12% -0.49% 7.13% -0.34% 6.87% 0.11% 6.63% 0.31% 6.72% 0.74% 5.16% 2.22% 85.51 12% 7.20% -0.30%
73.21 10% 6.81% 0.59% 70.20 15% 5.27% 2.36%
* Threshold breakeven level at which the principal deflation floor would be hit.
183
cheapening of linkers in asset swap terms, though the timing of such flows is difficult to predict.
Exhibit 21: Livret A inflows have been significant during 2011, we expect continued inflows in the first months of 2012
Inflows into Livret A accounts, centralised by CDC; bn
5 4 2008* 3 2 1 0 -1 -2 Oct-09 Oct-10 Apr-09 Apr-10 Jan-09 Jan-10 Jan-11 Apr-11 Jul-09 Jul-10 Jul-11 18.3 2010 7.8bn YTD 16.2bn
Outlook for real yields, breakevens, and linker asset swaps

Real yields can go more negative Real yields for German linkers are negative out to the 7Y sector and a mere 30bp in the DBRi 2020 line. Real yields for French linkers are slightly higher given the current credit risk dynamics. The main driver for real yields has been a deteriorating macro outlook, combined with the flight-to-quality resulting from the sovereign risk crisis. We think these factors are likely to continue to drive real yields into more negative territory while the ECB easing the real refi rate should also contribute to lower real yields at the front end (with our view not fully priced in, see European Derivatives). With a forecast of 10Y Bunds reaching 1.25% by 2Q12 and Italian spreads widening 250bp (see Euro Cash), we believe real yields will go negative in the 10Y sector. If we use the beta of real to nominal yields of the past 12 months, the DBRi 2020 should decline around 40bp, to negative -0.10% real yield, while 10Y breakevens would fall by a similar extent, to around 1.10%. Upside risks to our inflation call The J.P. Morgan inflation assumption is consistent with inflation staying a bit stickier than the path implied by inflation swaps (Exhibit 23). Moreover, it does not factor in upside risks from rises in state-administered prices (VAT and duties), which we see as a potential upside risk for the beginning of 2012. With Italian, Spanish, and French public finances under close scrutiny, we believe that any shortfalls in austerity plans are likely to be met with rises in taxes or state-controlled prices. Italy and Spain have just formed new governments, which will be very much focused on providing a deficitreduction plan, while France is being closely watched by markets. We have already seen a rise in administered-prices inflation in Italy and Spain, where these correspond to around 8% of the respective inflation baskets (Exhibit 24). In Exhibit 25, we set out the impact state-controlled prices may have on Euro HICP inflation, looking at VAT rises, VAT reclassifications, general rises in administered prices (which include services such as refuse collection, water supply, as well as pharmaceutical products), and
184
Exhibit 22: as we expect the Livret A rate to go up to 2.50% in February before moving down to 1.75% in the following fixing (given the expected downward trajectory in inflation)
Livret A rate*, French CPI ex tobacco inflation oya, previous month average of 3M Euribor and EONIA; %
2007*
5.00 4.00 3.00 2.00 1.00 0.00 -1.00
Av g of 3M Euribor & EONIA (1M) Liv ret A Rate Current: 2.25%
Forecasts
Inflation
Jan07
Jan08
Jan09
Jan10
Jan11
Jan12
* Livret A rate is set by Bank of France every 15 January and 15 July with reference to an average of inflation and rates (also depicted in the Exhibit), floored at inflation plus 25bp.
Exhibit 23: Inflation swaps assume a larger fall in inflation than the J.P. Morgan forecast, even if this does not include upside risks from administered prices, state-controlled duties and VAT
Euro HICP ex-tobacco J.P. Morgan forecasts and implied by HICP swaps; % oya
3.0 JP Morgan 2.5 2.0 1.5 1.0 Oct11 Dec11 Feb12 Apr12 Jun12 Aug12 Oct12 Dec12 EURO HICP sw aps
duties on energy prices. Our conclusion is that the potential impact from these measures on Euro HICP can be quite substantial, somewhere between 0.2% and 0.5%, on inflation. Inflation breakevens offer value at the front end Traditional fundamental models for inflation breakevens would fail to explain recent performance. Breakevens were driven by nominal yields, peripheral spreads and equities more than by inflation fundamentals. As a measure of value in inflation breakevens, we prefer looking at the excess nominal returns that sub-5Y linkers would provide over conventional bonds, using a blend of our inflation scenario to 2013 and the path implied by inflation swaps thereafter (Exhibit 26). Linkers offer above-average returns vs. conventionals in the case of France (except OATei 12), but even the German OBLi 13 offers a substantial 65bp pickup vs. Bunds. Even if the current illiquidity dynamics justify positive excess returns in linkers, we think front end breakevens offer good value at these levels. Long end breakevens to be driven by nominal yields and a flight to quality dynamic Favour breakeven flatteners Inflation breakevens further out are much more likely to be driven by the performance of Bund yields and a flight-to-quality dynamic. The inflation
Exhibit 24: Administered prices inflation has sky rocketed in Italy and Spain, as one would expect with countries engaging in fiscal tightening, thus we expect more pressure on this front
Administered price inflation as defined by Eurostat*; % oya
10% 8% 6% 4% 2% 0% -2% -4% 2002 2004 2006 2008 2010 France Italy Spain
* Items correspond to around 8% of basket in Italy & Spain, 13.5% in France.
breakevens implied by the French HICP-linked lines are only a touch lower than German breakevens (Exhibit 27), and we see risks that they underperform in relative terms, driven by a general widening of French nominal yields vs. Bunds.
Exhibit 25: Fiscal austerity programmes can impact short-term inflation significantly, with numerous challenges putting pressure to reduce deficits: optimistic growth forecasts and increased visibility will force countries to hike VAT, administered prices, and other duties. Italy, Spain, and France could push Euro HICP by up to half a percentage point
Impact on Euro HICP inflation from hypothetical changes in government-controlled measures
Admin Euro area VAT rate HICP w eight Germany France Italy Spain Netherlands Greece Euro area 26% 21% 18% 13% 5% 4% %) 19.0 19.6 21.0 18.0 19.0 23.0 19.8 prices >> 12.9 13.6 8.1 8.0 13.6 9.8 11.3 % duties on gasoline 59.2 57.3 55.8 49.0 60.3 59.3 56.9 1% higher VAT* 0.11 0.09 0.08 0.05 0.02 0.02 0.37 0.22 VAT 0.06 0.06 0.04 0.03 0.01 0.01 0.21 0.12 Admin reclass** 2.5% rise 0.08 0.07 0.04 0.03 0.02 0.01 0.24 0.13 0.07 0.07 0.01 0.06 Rise Total potential 0.25 0.22 0.16 0.16 0.05 0.03 0.88 0.55 gas^ (standard w eight prices*** duty on Euro HICP inflation sensitiv ities
France + Italy + Spain impacts combined
>> Weight of items whose prices are fully or mainly administered by the State, as defined by the European Commission. * Effect from a 1% rise in standard VAT rate. Assumes a pass through of 65% on 65% of the total basket, or around a 0.40% increase in domestic inflation. ** Assuming a VAT reclassification of 2% of basket items from the reduced rate of VAT to a standar rate of VAT. *** Assuming a 2.5% rise in all administered prices in country baskets (average administered price inflation has been 3.5% over the last 7 years). ^ Assuming convergence in duties as a % of overall price of gasoline to the Euro area average, affecting the item Fuels for personal transport equipment in the basket.
185
Overall, being short breakevens is not necessarily the best risk-adjusted trade, given the expected volatility with sovereign risk dynamics. Therefore, we prefer expressing this view as a curve trade through inflation breakeven flatteners. This is consistent with our expectation of significant flattening in the nominal yield curve (in 2s/10s), plus our view that inflation can surprise on the upside in 1H12. There are different variants in which we can express this view: 1)1s/10s flatteners in HICP swaps (more of a play on inflation surprising on the downside), 2) overweighting OBLi 13 vs. BTPei 19 in breakeven terms (incorporating a negative view on Italian breakevens), or 3) BTPei 16s vs. BTPei 21 (incorporating the value of the deflation floor in BTPei 16). Inflation swaps are likely to outperform cash breakevens as peripheral spreads continue to widen, and in general, we wouldnt fade the relative historical cheapness in the first months of 2012. A further cheapening may provide good medium-term opportunities to buy linker breakevens vs. swaps. Regarding Euro HICP-linked vs. French CPI linkers, the latter have underperformed significantly in November and present opportunities for investors to overweight OATis vs. OATeis over the medium term. Whereas OATis have traded historically at higher breakevens than OATeis, they are currently trading around similar breakeven levels, around 20bp too cheap vs. a regression since January 2010 in the sub-10Y sector (Exhibit 28). While this may reflect the relatively lower liquidity of French CPI-linked lines, we expect a decent amount of Livret A hedging to take place early in 2012, which should favour French CPI lines. Any demand for French linkers from Libor-based investors should also favour French CPI-linked lines, given they provide a more attractive pickup.
Exhibit 26: Sub 5Y linkers will provide reasonable pickup vs. conventionals under our inflation scenario, which suggests breakevens offer value
Current levels, Nominal IRR under J.P. Morgan inflation assumption and pickup relative to conventional bonds; % and bp Nom Rtn Pickup vs.
Lines Euro HICP lines Germany OBLI 2.250 0413 Germany DBRI 1.50 0416 France France France France France OAIE 3.000 0712 OAIE 1.600 0715 OAIE 2.500 0713 BTNS 0.45 16 IL OAIE 1.000 0717
Real Yld Inf B/E -0.60 -0.35 -1.52 1.30 0.28 1.59 1.75 0.86 1.20 2.43 0.91 1.12 0.87 1.04
(JPM*) 0.97 1.31 0.84 3.22 2.14 3.50 3.70
Noms (bp) 65 35 7 88 55 91 86
France CPI lines
* Projected nominal return using J.P. Morgan forecast out to December 2012 and inflation forwards implied in Euro HICP swaps thereafter with J.P. Morgan seasonal model.
Exhibit 27: Inflation breakevens for French linkers are trading close to those of German lines, and could under perform, though we do not expect them to behave like Italian breakevens
Inflation breakevens for Euro HICP linked lines (x-axis denotes redemption year); %
2.50 2.00 1.50 1.00 0.50 0.00 -0.50 -1.00 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Italy France Germany
Exhibit 28: French CPI-linked breakevens are cheap vs. French Euro HICP-linked breakevens
Residual from regressing OATi 2019 vs. OATei 2020 breakeven, since January 2010; %
Supply/demand dynamics
Euro area linker supply: risk of very low supply from Italy, lower volumes than in 2011 Supply of inflation-linked bonds in 2011 was below our initial expectations (41bn vs. our initial forecast of 51bn), as Italy issued only around 1.5bn of linkers (net of buybacks) in 2H11. This was clearly a reflection of the limited demand for Italian linkers during the sovereign debt crisis that began unfolding in the summer. France continued to issue at regular auctions, supplying around
20 10 0 -10 -20 -30 Jan10 Apr10 Jul10 Oct10 Jan11 Apr11 Jul11 Oct11
186
3.5bn during this period, while Germany did not hold any auctions between April and November. With sovereign risk dynamics likely to continue to impact markets, we expect supply in 2012 to be a below-average year. The total supply to the market will very much depend on Italys strategy towards linker issuance in a dislocated market. While we are calling for wider BTP spreads, the timing of such widening may impact the issuance strategy from the Italian Treasury, which will remain as flexible as possible. We have therefore forecast Italian linker supply to be of the order of 12bn, assuming they continue to tap the market in an opportunistic manner. We expect Germany to issue 9bn while we have penciled 18bn for France (with a bit more issued in Euro HICP lines through a new 5Y benchmark in Q1). There are clear downside risks to Italian linker issuance, which could result in a bit more issued from France and Germany. Overall, we expect the supply of linkers to be at best around 39bn, slightly below what was issued in 2011 (Exhibit 29). On the demand side, inflation-hedging at the long end of the curves has been very limited in 2011, and we do not expect a huge change in this dynamic in 1H12. The pension reform agenda in the Netherlands can be a significant development for inflation-hedging, but we do not expect a quick implementation of the proposed measures, which give more leeway for Dutch pension funds to be more active in inflation hedging programmes.
Exhibit 29: Euro area linker supply forecasts for 2012: lower volumes than an already low 2011
Gross supply (net of buybacks); bn
Euro HICP France (Euro) Italy (Euro) Germany (Euro) France CPI Total % of total vs 2011 By Maturity Short (<3yrs) Medium (3-7yrs) Long (7-12yrs) Ultra long (13yrs+) Q1 10.5 5.3 3.3 2.0 2.8 13.3 32% (3.5 ) Q2 8.0 2.8 3.3 2.0 2.8 10.8 26% (4.6 ) Q3 7.5 1.5 3.0 3.0 0.8 8.3 20% +4.6 Q4 6.0 1.5 2.5 2.0 0.8 6.8 16% +0.7 Total 32.0 11.0 12.0 9.0 7.0 39.0 93% (2.8 ) % 82% 28% 31% 23% 18% vs 2011 +0.8 +1.6 (1.8 ) +1.0 (3.6 ) (2.8 )
1.3 6.8 2.5 2.8
1.5 4.3 2.8 2.3
1.3 0.0 2.5 4.5
1.0 2.0 2.3 1.5
5.0 13.0 10.0 11.0
11% 28% 22% 24%
+5.0 (5.7 ) (5.6 ) +3.6
Exhibit 30: The 5s/30s inflation curve should be steeper in US CPI swaps compared to Euro HICP swaps
5s/30s HICP swap and US CPI swap curves; bp
140 120 100 80 60 40 20 0 Nov 09 May 10 Nov 10 May 11 Nov 11 Euro HICP 5s/30s US CPI 5s/30s
Cross market: relative flattening vs. the US inflation curve

Our view for Euro area breakevens contrasts with our outlook for US TIPS, where we see short dated breakevens most at risk and expect the breakeven curve to steepen (around 40bp in 5s/30s by 2Q 12, see US TIPS section). This is on the back of significant expected falls in US inflation prints and further impact from fiscal tightening on short dated breakevens.
Thus we generally recommend overweighting Euro area breakevens at the front end vs. US TIPS, whilst doing the opposite in 10Y+ maturities. In swaps space, the 5s/30s curve in HICP swaps is broadly at the same level as in US CPI swaps, whereas on average the latter has traded 25bp steeper (Exhibit 30).
187
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Jorge GarayoAC (44-20) 7325-4820 Francis Diamond (44-20) 7325-3541 J.P. Morgan Securities Ltd
UK inflation-linked markets
2011 review
Index-linked gilts underperformed nominals during 2011. Despite the fact that both CPI and RPI inflation both broke the 5% level (5.2% and 5.6%, respectively), the linker market was driven by the collapse in nominal yields as investors sought the safe haven of UK gilts as the situation in peripheral Europe deteriorated. On the real yield curve, 5Y real yields fell the most but were also the most volatile (Exhibit 31), while at the long end of the curve, 30Y and 50Y real yields have moved broadly in lockstep. Real yields are barely above 0bp at the long end, and although this part of the curve is much less affected by fundamental valuations and macro drivers than the short end, there has been a high correlation between 30Y and 10Y real yields. Breakevens are lower on the year with the movement driven entirely by the fall in nominal gilt yields. The resumption of QE purchases in October had little impact on cash breakevens, despite the fact that linkers are not included in the QE purchase basket. This was not the case when QE was first started in March 2009 when 10Y breakevens initially fell 70bp following the announcement, only to reverse this over the following weeks. 2011 saw the issue of new ILG29, ILG34, and ILG62 gilts via syndication with the latter, attracting an order book close to 10bn.
Exhibit 31: 5Y real yields outperformed in 2011 and correlation with nominal yields was high across the curve
Performance of UK index linked gilts from 1 January 201117 November 2011; bp unless stated Std. R-sq v s. R-squ Real y ield Current High Low Chg. Dev * 10Y real v s. nom
ILG16 ILG22 ILG32 ILG40 ILG55
-140 -30 5 5 4
6 87 91 88 72
-153 -36 3 5 3
-132 -96 -63 -61 -48
5 5 4 4 3
93% 100% 97% 95% 88%
96% 95% 91% 88% 73%
Exhibit 32: Our forecast for lower headline inflation in 2012 is driven by falls in all the main component
Contribution to headline CPI by component, %oya
2.5 2.0 1.5 1.0 0.5 0.0 -0.5 Jan-11 May -11 Sep-11 Jan-12 May -12
Food Energy Core goods Serv ices
Sep-12
2012 inflation view

2011 saw CPI inflation rise to the highest levels since September 2008, driven by higher food and energy prices. Core CPI rose above 3.5% oya during 1H11 before subsiding. Looking forward into 2012, we think headline and core inflation will fall markedly to 2.7% oya and 2.0% oya, respectively by the end of 2Q12 (see Exhibit 7 in the earlier section Inflation outlook for 2012). The drivers are lower food and energy prices on the back of base effects, a downshift in core goods prices driven by a weaker growth outlook and the impact of the VAT hike falling out of oya inflation (Exhibit 32). We forecast the contribution of core goods inflation to headline CPI to be slightly negative by the end of 2012. The main risks to our inflation forecast stem from downside risks to growth. Our baseline view is that the UK generates anaemic growth of 0.5% oya in 2012 but there is clearly the risk that output is flat or negative over the year, dragging core goods and services inflation
188
Exhibit 33: An increase in QE gilt purchases will push 5Y real yields lower we forecast a 50bp fall
Estimated change in 5Y real yields based on two factor model* under various scenarios for additional QE and level of RPI by mid 2012; bp Mid 2012 RPI, %oya
4.50 Additonal QE by 50 mid 2012, bn 75 100 125 150 -20 -35 -45 -60 -75
4.00 -10 -25 -40 -55 -70
3.50 -5 -20 -35 -45 -60
3.00 5 -10 -25 -40 -55
2.50 10 -5 -20 -35 -45
2.00 20 5 -10 -25 -40
* we model 5Y real yields as a function of the base rate adjusted for QE purchases** and the level of RPI. 5Y = 0.48 * effective policy rate 0.15* RPI. R-squ: 81%, std. error: 50bp ** we assume 25bn of QE lowers base rate by 0.3%. Current effective rate is 2.8%.
down. Aside from the usual monthly volatilities surrounding the CPI comments, it is difficult to envisage a scenario in which inflation does not fall from current levels. The impact on inflation of QE increasing to 425bn as per our forecast is difficult to gauge, but the BoE estimates that the impact of 200bn QE on headline
CPI has been between 0.75% and 1.50%, and we dont discount that a further increase in QE could prevent inflation falling as low as we forecast. The monetarist argument that excess reserves will generate inflation will only become relevant once economic recovery is achieved and demand for credit increases a situation which is several years away, in our view.
Exhibit 34: Long positions in 5Y and 10Y real yields look attractive compared to the forward real yield curve
Implied carry and slide for being long real par rates over 3M, 6M, 9M, and 12M horizon; bp
30 25 20 5Y 10Y 30Y 50Y 21 17 14 11 7 6 1 3
28 22
Real yields view be long 10Y real yields

Real yields are at historically low levels, but we think that they can fall further in 2012, particularly in the sub10Y sector of the curve. The QE increase in October 2011 and the very dovish set of forecasts in the November Inflation Report suggest that the BoE is focused on managing downside risks to growth, with more QE likely in the near future in our view. Our model of 5Y real yields as a function of the base rate adjusted for QE purchases and the level of RPI (to measure demand for inflation protection) suggests that 5Y real yields can fall around 40bp-50bp by mid-2012 (Exhibit 33) from current levels. Given the strong correlation between 5Y and 10Y real yields we expect ILG22 real yield to fall around 30-40bp by mid 2012. We note that, long positions in 5Y and 10Y linkers offer attractive carry and slide vs. the forwards (Exhibit 34). During 2010, 30Y real yields traded in a very pronounced 60bp90bp range (ILG40) and many end investors tailored their investment strategies around playing this range. This has not been the case this year as 30Y yields have trended downwards (Exhibit 35). One of the key questions for 2011, in our view, is the direction of 30Y real yields. Will 30Y yields rise back up to the 2010 levels or will a new trading range be established around current levels? To answer this question, we need to take a look at the drivers of 30Y real yields, namely nominal 30Y yields and the supply/demand outlook. We note that that correlation between 30Y real and nominal yields has not been constant over the past few years, particularly during periods of QE gilt purchases when the relationship has tended to be weaker (Exhibit 36). This correlation also tends to increase when nominal yields are trending higher or lower (Exhibit 37). We can model 30Y real yields as a function of 30Y nominal yields and the size of pension fund liabilities (Exhibit 38) where an increase in liabilities tends to result in lower real yields. The intuition here is that as
15 10 5 0 3M
6M
9M
12M
Exhibit 35: Long-end real yields trended lower in 2011 after rangetrading for most of 2010
Par 30Y real yield; %
1.20 1.00 0.80 0.60 0.40 0.20 0.00 Feb 10 Aug 10 Feb 11 Aug 11
Exhibit 36: The correlation between 30Y real and nominal yields is weaker when the BoE is conducting QE gilt purchases
Beta and r-squared between 30Y real par and 30Y nominal par yields, January 2009October 2011
0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 Beta R-squ QE Non-QE
189
liabilities, which are inflation-linked, increase, then pension funds are more likely to buy index-linked gilts to hedge these, although we note that changes in liabilities may not always result in renewed hedging flows as they can also increase due to increases in longevity and tend to increase as the point of payment increases. This model suggests that 30Y index-linked gilt yields currently look cheap (Exhibit 39), and our forecast for a 100bp fall in 30Y gilts would imply 30Y real yields rallying 30bp to around the -25bp level, although we think that in practice any move to this level would be very slow and gradual. The 0bp level may represent a psychological barrier around which investors may attempt to fade any move to negative 30Y real yields. At the other extreme, we think any move in 30Y real yields back up to the 1% level is highly unlikely to occur over 2012 as 30Y nominal gilt yields would need to rise to around 5% based on the current betas.
Exhibit 37: The relationship between 30Y real yields and nominals is weaker when nominal yields are range-bound
Statistics from regressing 30Y par real yields vs. 30Y par nominal yields Period 30Y y ield Beta Rsq
19/11/09 - 15/2/10 16/2/10 - 16/12/10 10/2/11 - 18/11/11
sell-off range Rally
0.8 0.2 0.5
64% 37% 89%
Statistics from regressing 30Y par yield vs. 30Y nominal par yield and the PPF measure of total liabilities for DB pension funds, monthly data, March 2003Oct11 Variable Coefficient T-stat
Exhibit 38: 30Y real yields can be explained by nominal yields and the size of pension fund liabilities
30Y nominal y ield (%) PPF liabilities (bn) R-squ Std. error (bp)
0.308 -0.003 81% 21
3.6 -18.2
Exhibit 39: and currently look some 30bp cheap
Breakeven view
Breakevens have fallen over the past year by 50bp in the 10Y and 30Y sectors, and by nearly 100bp in the 5Y sector, despite the rise in CPI inflation to 5.2% and RPI inflation reaching the 5.6% level. In fact, over the past few years 5Y cash breakevens have show a weak relationship with spot inflation prints unless RPI is below the 2% level (Exhibit 40) and over the past couple of years, nominal yields have been the dominant drivers of breakevens (Exhibit 41). Our breakeven view for 2012 is as a straightforward one lower nominal yields will drive breakevens lower in the 5Y and 10Y sectors of the curve. Our forecast for a 75bp fall in 10Y gilt yields during 1H12 (to the 1.50%) level would broadly translate into a 30bp fall in ILG22 breakevens, to the 2.30% level. Our expectations for CPI and RPI inflation may, at the margin, also contribute to lower breakevens. In the 30Y sector, we note that non-linearities are starting to appear in the relationship between breakevens and nominal yields (Exhibit 42), which is likely a result of demandspecific drivers. Pension funds and ALM-driven investors tend to have specific trigger levels at which they look to hedge inflation which may limit the magnitude of any down move in 30Y breakevens. The 5s/10s inflation curve has been broadly range-bound, particularly in the RPI swap space, and we think this can continue. We think it is far too early to position for sustained steepening on the back of unanchored inflation expectations from an increase in QE. The monetarist
190
Actual and predicted level* of 30Y par real yields, monthly data March 2003 October 2011; %
2.50 2.00 1.50 1.00 0.50 0.00 -0.50 Mar-03 Mar-05 Mar-07 Mar-09 Mar-11 0bp lev el Predicted Actual
* predicted using model in Exhibit 38.
Exhibit 40: 5Y breakevens are generally non-directional with inflation when RPI is above 2%
5Y par cash breakeven* regressed against RPI print, 20052011**; %
4.00 3.50 3.00 2.50 2.00 1.50 1.00 -2.0
y = 0.31x + 2.3 R 2 = 29%
y = 0.1x + 2.55 R 2 = 3% -1.0 0.0 1.0 2.0 3.0 RPI, %oy a 4.0 5.0 6.0
* Derived from BoE fitted par curve. ** Excludes period from December 2008February 2002.
arguments that, for additional QE, translates into higher inflation is not easy to demonstrate empirically, and we think that inflation expectations run the biggest risk of becoming unanchored when the macro backdrop has improved. This scenario is several years away, in our view, and we see little benefit in positioning for this outcome now.
Exhibit 41: Breakevens have been driven by nominal yields over the last couple of years. We expect the strong relationship to continue
Statistics from regressing breakevens vs. nominal yields over the last 12M and 24M 12M 24M
Beta ILG22 ILG32 ILG40 ILG55 0.4 0.5 0.5 0.6
R-squ 87% 91% 92% 93%
Beta 0.4 0.5 0.5 0.7
R-squ 68% 77% 81% 84%
Cash breakevens vs. RPI swaps

The past few months have seen a marked underperformance of index-linked cash breakevens compared with RPI swaps, with the relative spread now at the widest levels ever for ILG55 and at its widest level since March 2009 for ILG37 (Exhibit 43). This spread essentially reflects the relative ASW levels between index-linked gilts and conventional gilts, and indicates that linkers offer an attractive pick-up vs. Libor compared with conventional gilts. Going forward, we think this relative spread can widen further, and we would not advocate fading this apparent cheapness. We think Libor-based investors may look to take advantage of the relative cheapness in linkers compared with nominals, but over the past 6 months, this relative spread has been driven by nominal gilt ASW levels and European peripheral spreads, particularly in the 10Y+ sector (Exhibit 44). We think cash breakevens will continue to underperform swap breakevens, given our view that peripheral spreads will widen further in the coming months and that nominal ASW levels may fall further (i.e. gilts outperform swaps), driven by additional QE and lower nominal gilt yields. For investors who do want to fade this cheapness, we think the sub-10Y part of the curve is the best place to position, given the lower correlation with peripheral spreads and nominal ASW levels.
Exhibit 42: but the long end of the BE curve is becoming non-linear
ILG40 regressed against nominal 30Y yields, last 3M; %
3.50 3.40 3.30 3.20 3.10 3.00 2.90 3.0
18 Nov
y = 0.64x 2 - 3.9x + 9.0 R 2 = 83%
3.2
3.4
3.6
3.8
4.0
30Y nominal y ield, %
Exhibit 43: Cash breakevens are at cheap levels vs. RPI swaps across the curve
Cash breakeven RPI swap for selected linkers; bp
150
ILG17
ILG37
ILG55
100
50
Market technicals Issuance

Our fiscal outlook and issuance forecasts for gilts are described in the United Kingdom section, and for FY12/13, we expect linker issuance to break through the 40bn level (Exhibit 45). The low level of spot and forward real yields, and the continued backdrop of strong demand from institutional investors warrant an increase in index-linked gilt supply, in our view. Total gilt supply is expected to rise to 190bn, and we forecast linker issuance remaining around 23% of total gilt supply. As is typically the case, we expect the bulk of issuance to
-50 Oct 08 Jun 09 Feb 10 Oct 10 Jun 11
come in the long end (15Y+) of the real yield curve, with taps to ILG34, ILG40 and ILG62 likely. We also see a reasonable chance of a new linker in the 35Y sector of the curve, as well as the possibility of a short-dated linker in the 10-12Y sector, although this is less likely now that the ILG29 has been issued.
191
The results of the consultation process on whether the DMO should issue CPI-linked gilt should be available in the coming months. In our view, the decision as to whether CPI-linked gilts will be issued will be very much driven by anticipated end-investor demand, particularly pension funds and other investors who will be required to provide CPI-linked payments. We expect around 15% of pension schemes with indexation linked to RPI will be able to or will want to make the switch from RPI to CPI indexation, as mandated by new legislation. However, given that pension fund schemes hedge only around 60% of their inflation risk, they may prefer to continue to use RPI-linked instruments and run the CPI-RPI basis risk. If CPI-linkers were to be issued, we think they would likely be in the 15-30Y maturity sector of the curve where demand is likely to be the highest. We would also expect a flexible issuance programme to enable the DMO to tailor issuance needs in either the CPI-linked or RPIlinked space to meet demand, with the allocation of supply between the two instruments likely to vary during the fiscal year.
Exhibit 44: and peripheral spreads and nominal ASW levels explain the bulk of this cheapening for 10Y+ linkers
R-squ from separately regressing* cash BE RPI spread vs. nominal ASW and peripheral spreads, last 6M; %
100% 80% 60% 40% 20% 0% ILG17

* Single factor regressions.
Nom ASW
Peripheral spread
ILG22
ILG37
ILG55
Exhibit 45: We expect over 40bn of linker issuance in FY12/13

Index linked gilt total issuance and supply as % of total gilt issuance bn %
50 Amt, bn 40 30 Proportion, %
40% 30% 20%
Real yield behaviour into syndications

For the past couple of fiscal years, linker syndications have risen to around 40%50% of total index-linked gilt sales, and we expect a similar proportion of linkers to be issued via syndications in FY12/13, equating to 18bn 22bn of syndicated linker issuance, based on our forecast. Since the DMO introduced syndications as a permanent feature of the issuance programme in April 2009, there have been 9 index-linked gilt syndications, raising a total of 38.75bn, with an average size of 4.3bn. These syndications have given rise to some interesting behaviour on the linker curve. Looking at the performance of the closest maturity bond to the linker being issued, we note that, on average, real yields have risen 8bp in the 15 business days prior to the syndication (Exhibit 46). This is then quickly reversed in the five days following the syndicated supply. This phenomenon has occurred into 7 out of the 9 syndications, with the largest increase close to 25bp. We think that this can be explained by the fact that 1) syndications are often multiples of the size of typical index-linked gilt auctions and represent large concentrated amounts of linker duration supply into the market. This creates uncertainty around the actual potential demand for the bond; and 2) the actual details of the bond maturity are not announced until a few weeks before the syndication date.
192
20 10 0
2012 (f) 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
10% 0%
Exhibit 46: Real yields rise prior to linker syndications on average

Average real yield of closest maturity linker around index linked gilt syndications*, July 2009October 2011; bp
70
65
60
55 -20 -15 -10 -5 0 5 10 15 20 Business day s around linker sy ndications

* Dates and linkers used: 23/07/09 ILG42, 24/09/09 ILG50, 27/01/10 ILG40, 26/05/10 ILG50, 27/07/10 ILG40, 27/01/11 ILG55, 24/05/11 ILG34, 26/07/11 ILG34, 25/10/11 ILG62.
We expect this trend to continue and we think investors can exploit this dynamic by selling the closest maturity bond to the linker being issued via syndicate in 15 business days ahead of linker syndications. Instigating a trading rule in which the closest maturity linker to the syndicated bond is sold on the 15th day prior to the syndication and unwound on the day of the syndication itself would have generated a total of 24bp of P/L in yield terms in 2011, and on average, 6bp per syndication over the past three years (Exhibit 47). Demand for linker syndications is almost entirely from domestic investors, and the size of the book as well as the actual amount of the bonds issued can be affected by large asset allocation orders. Predicting the size of the potential demand for linker syndication in advance is not easy, but looking at the relative valuations between equities and real yields may indicate potential demand. The idea here is that pension funds and asset managers are more likely to switch from equities into linkers when equities are relatively rich compared with linkers on a real-yield basis. This is because pension funds asset/liability mix is not managed based on specific funding level triggers but more holistic, long-term funding plans, and on aggregate, funds are reluctant to crystallise equity losses. We can measure the relative valuations by looking at the ratio of equity E/P (as a measure of equity real yield) vs. the real yield on long linkers. Looking at this valuation metric and its evolution in the few days prior to
Exhibit 47: creating attractive trading opportunities for investors
Total and average P/L from trading rule which goes short closest maturity linker on T-15 and unwinds the trade on T*; bp
35 30 25 20 15 10 5 0 2009
Total P/L
Av erage P/L
2010
2011
* T = date of linker syndication.
syndications there appears to be a loose relationship i.e. syndication demand is stronger if equities have richened vs. linkers, although this is skewed by the strong demand for the ILG62 syndication in October 2011. It is worth noting that the syndication of ILG62 coincided with the richest levels of equity/linkers for 2011, and we think that looking at the performance of this ratio ahead of future linker syndications may provide some indication of potential investor demand.
193
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Jorge GarayoAC (44-20) 7325-4820 Kimberly Harano (1-212) 834-4956 J.P. Morgan Securities Ltd, J.P. Morgan Securities LLC
US inflation-linked markets
Up, down, then back around
The path of breakevens over 2011 looks remarkably similar to the 2010 experience. First, breakevens steadily widened over the first few months of the year thanks to a surge in energy prices (Exhibit 48). In late spring, however, European concerns resurfaced, the US economy appeared to hit a speed bump, and the resulting risk-off trade pushed breakevens narrower. Finally, in the fall, the Fed was forced to step in with Operation Twist, and this action, along with some positive news out of the October EU summit, helped breakevens reverse some of their narrowing. On a total return basis, the sizable increases in headline CPI helped TIPS outperform nominals early in the year, but TIPS sharply underperformed Treasuries in 3Q11 as nominal yields declined more than real yields (Exhibit 49). Real yields, in turn, refreshed their all-time lows this year (Exhibit 50). Five-year yields traded negative for most of 2011, hitting a low of -1.12% in early November. As a result, the Apr-16 TIPS became the first Treasury security issued with the new minimum 0.125% coupon rate, despite a clearing yield of -0.18%. Even 10-year real yields turned negative in 3Q11, reaching -0.15% at the lows.
Exhibit 48: Breakevens widened early in the year thanks to a surge in energy prices, but narrowed sharply in the summer as growth prospects waned and risk aversion rose
5-year, 10-year, and 30-year TIPS breakevens (all left axis) versus rolling front WTI futures contract price (right axis); bp $/bbl
300 30Y 250 10Y
Oil prices
120 110 100 90
200
150
5Y
80
70 100 Nov 10 Jan 11 Feb 11 Apr 11 Jun 11 Jul 11 Sep 11 Nov 11
Exhibit 49: TIPS outperformed nominals on a total return basis in 1H11, but sharply underperformed in 3Q11
Quarterly total returns on J.P. Morgan TIPS index (JUSTINE) and GBI-US index by maturity bucket; duration of the indices, and TIPS excess returns*; %
Nominal returns Nominal dur'n TIPS returns TIPS dur'n TIPS-nom returns* 1-10Y 10Y+ 1-10Y 10Y+ 1-10Y 10Y+ 1-10Y 10Y+ 1-10Y 10Y+ 1Q11 2Q11 3Q11 -0.1% 2.3% 3.8% -0.9% 3.3% 23.5% 3.9 3.9 4.0 13.0 13.2 14.7 2.3% 2.9% 1.8% 1.3% 5.0% 11.7% 4.7 4.9 5.1 14.4 14.3 14.8 2.4% 0.0% -3.0% 1.3% 2.3% 1.4% -12.0% 4.0%
QTD 0.1% -0.3% 4.1 14.8 1.4% 3.8% 5.1 15.2 * Calculated as TIPS return-(Tsy duration)/(TIPS duration)*(Tsy return).
TIPS supply and demand in 2012

After Treasury announced modifications to the TIPS calendar at the November 2010 refunding, 2011 became the first year with a TIPS auction every month. At this years November refunding, Treasury wrote that TIPS are an important part of Treasurys overall debt management strategy and indicated that it expects to continue to gradually increase gross issuance of TIPS in 2012. Based on these comments, we look for modest increases in auction sizes across the curve, for a total of $146bn of gross issuance in 2012 (Exhibit 51) compared to $131bn of expected issuance in 2011. With $46.3bn of redemptions in 2012 and net Fed purchases of zero, effective net issuance in 2012 will likely rise to $100bn versus $79bn (net of redemptions and Fed purchases) in 2011. Alternatively, we can evaluate changes in TIPS supply on a duration-weighted basis, since the Fed will remain a source of duration demand in 2012, even if its holdings of TIPS will likely remain unchanged in par notional terms. Over two rounds of purchases and sales
194
Exhibit 50: Real yields refreshed their lows in 2011

5-year and 10-year TIPS real yields; %
1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 Nov 10 5Y 10Y
Feb 11
Jun 11
Sep 11
operations, the Fed has bought about $2.3bn of 10-year TIPS equivalents per month. If we assume the Fed maintains this pace, then it will likely purchase about $14bn 10-year Treasury equivalents in 2012. However, these purchases will only partially offset the expected
rise in TIPS duration supply. As Exhibit 52 illustrates, even with Fed demand, we expect TIPS duration supply to rise to $140bn in 2012 from $108bn 10-year TIPS equivalents in 2011. Although supply should increase in 2012, we expect end-user demand to rise to meet it. As Exhibit 53 shows, while the average offering size for TIPS has steadily increased over recent years, so has end-user demand, with direct and indirect bidders taking down even greater fractions of (even larger) issues. The effect of larger auction sizes can also be seen in primary dealer transaction data: as Exhibit 54 shows, transaction volumes in TIPS have steadily increased since Treasury committed to increase auction sizes and improve liquidity in late 2009. Although increased transaction volume by itself is not necessarily indicative of improved liquidity, anecdotal reports suggest that the increased frequency and size of auctions have improved the depth of the TIPS market and attracted a broader base of investors1. Outside of auctions, we can gauge investor demand for TIPS based on flows into inflation-protected mutual funds. In our 2011 Outlook, we noted that mutual fund flows have historically been correlated to changes in energy prices, with higher prices leading to greater flows. In addition, since the Fed began its QE programs, longterm inflation concerns have increaseda development
Exhibit 53: If you sell it, they will come
Offering size for TIPS auctions averaged by calendar year, versus average of sum of direct and indirect bidder percentage*; $bn %
Exhibit 51: We expect gross issuance of TIPS to increase modestly in 2012

5yr TIPS Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 TOTAL 14 44 76 26 12 14 12 8 14 16 12 8 12 10yr TIPS 30yr TIPS 14 10
J.P. Morgan forecast for gross issuance; reopenings shaded in grey; $bn of real principal
Subtotal
36
36
40
34 146
Exhibit 52: and duration supply will rise substantially even with Operation Twist
Gross issuance of TIPS versus duration supply net of Fed purchases*; $bn of real principal in 10-year TIPS equivalents
90 80 70 61 60 51 50 41 41 40 1H10 2H10 1H11 2H11 1H12 2H12
* Assumes pace of duration buying over October and November 2011 ($2.3bn 10-year TIPS equivalents per month) is maintained.
Gross issuance Dur'n supply net of Fed 77
86
72
68 68
54 47
54
11 10
Avg auction size
End user %
55%
50% 9 8 7 40% 6 5 2006 2007 2008 2009 2010 2011

* 2011 average is over January-November. Source: US Treasury
45%
35%
that has been captured in the rise in the 5Yx10Y payer swaption skew. As a result, we found that the level of the swaption skew has helped explain flows into inflationprotected mutual funds over the past three years. With the skew holding fairly constant this year, however, energy prices were the dominant driver of mutual fund flows (Exhibit 55). As our oil strategists look for modest gains in oil prices next year (see below), we expect to see continued inflows into inflation-protected mutual funds.
See US Treasurys Makeover on TIPS Ushers in Boom Era; China Buys, Min Zeng, Wall Street Journal, 11/15/11
195
In sum, although we look for TIPS supply to increase next year, we expect demand at auctions and retail demand to remain robust. Indeed, our J.P. Morgan US Fixed Income Investor Survey indicates that 24% of respondents plan to increase their exposure to TIPS next year, versus just 7% that plan to reduce exposure (see US Cross Sector Overview).
Exhibit 54: Transaction volumes in TIPS have steadily increased since Treasury committed to improving liquidity
Rolling 3-month moving average of primary dealer transaction volume in TIPS (daily average for each week); $bn
12
10
The outlook for US inflation

Last year, heading into 2011, we looked for headline and core inflation to end the year below 1%. Instead, monthly gains in headline inflation averaged 0.6% over the first five months of the year, reaching 3.9% (year-over-year ago) by September. Similarly, core CPI surged to 2.1% in October from 0.65% in December 2010. Looking ahead, we expect both headline and core inflation to moderate. As Exhibit 56 shows, the upside surprises in core CPI were driven primarily by strong gains in apparel and vehicle prices as well as some firming in owners equivalent rent (OER). The gains in the first two categories were driven by temporary factorshigh cotton prices and supply shocks from Japanwhich we do not expect to recur in 2012. In addition, with incomes constrained by the weak labor market, gains in OER should remain muted. Thus, we look for core CPI to decline to 1.0% and headline CPI to fall to 1.4% by year end. Exhibit 57 shows our forecast for the monthly changes in headline CPI. We look for prices to rise modestly over 1H12 and be roughly flat over 2H12. Two factors that pose risks to our inflation forecast are
Exhibit 56: The upside surprise in core CPI this year was driven by gains in OER, apparel and vehicle prices, which we do not expect to recur in 2012
Year-over-year-ago percent change in owners equivalent rent, apparel prices, and vehicle prices; %
4 Oct 06
Mar 08
Jul 09
Nov 10
Source: Federal Reserve Bank of New York.
Exhibit 55: Flows into inflation-protected mutual funds were driven by energy prices this year
Monthly flows into inflation-protected mutual funds* versus 1-year percentage change in oil futures prices, lagged 1 month; $bn %
2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 Mutual fund flows
1Y chg in oil prices
40 35 30 25 20 15 10 5 0 -5
Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep 10 10 10 10 10 11 11 11 11 11 11 11 11 11
* Includes both weekly and monthly reporters. Source: EPFR Global
6% 4% 2%
energy prices and the dollar. Our commodity strategists currently look for modest declines in Brent and WTI oil prices early in the year followed by increases in 2H12 (Exhibit 58). In addition, our currency strategists look for the dollar to weaken modestly over the next year.
Inflation expectations and dispersion

0% -2% OER -4% Oct 08
196
Apparel Jun 10 Dec 10
Vehicles Jul 11
May 09
Nov 09
How does our forecast for inflation compare to investors expectations? To answer this question, we examine the November 2011 results of our J.P. Morgan Inflation Expectations Survey, which surveys a wide range of investors across the globe. As Exhibit 59 shows,
investors expectations for near-term core inflation are higher than they were a year ago. Our November 2011 survey shows that over half of investors expect core inflation to be between 1.5%-2.5% next year, while a similar number expected inflation to range between 1-2% in 2011. In addition, the tails of the distribution have shifted: last year, 10% of investors expected core inflation to be below 0.5% while 8% expected inflation to exceed 2.5%. The corresponding figures for this years survey are 4% and 15%, respectively, indicating that investors now appear to be more concerned about upside risk in inflation than downside risk. The message is similar for medium-term expectations. Sixty percent of respondents think headline inflation will be above target or significantly above target over the medium term, while 8% think it will be below target (Exhibit 60). In contrast, in last years survey 57% of respondents thought headline inflation would be above target, while 13% thought it would be below target. It is interesting to note that the tails of the distribution have fallen: last year, about 30% of investors expected inflation to be either below target or significantly above target, while only 20% of investors held those extreme views this year. This decline in the dispersion of medium-term inflation expectations can be seen in other survey measures. For example, Blue Chip collects economists forecasts for various economic indicators over the long term. As Exhibit 61 shows, the dispersion of economists forecasts for medium-term inflation (over the next five years) rose sharply around the financial crisis, but has now declined to levels last seen in early 2008.
Exhibit 57: We look for headline inflation to moderate in 2012
J.P. Morgan forecast for monthly % changes in headline CPI and year-over-year ago rate; % %
0.3
0.20
0.23
0.26
Monthly change
0.24 0.16 0.09 0.08 0.15 0.05
3.0 2.5 2.0 1.5
0.2 0.1
0.02
0.0
-0.02
-0.1 -0.2 Year-over-year ago rate
-0.06
1.0 0.5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 12 12 12 12 12 12 12 12 12 12 12 12
Exhibit 58: We expect modest increases in oil prices and a weakening of the dollar over the course of 2012
Our forecast for WTI (period average) and the J.P. Morgan USD index (nominal narrow effective exchange rate) Current 1Q12 2Q12 3Q12 4Q12 Brent oil futures; $/bbl 107.56 108.00 105.00 115.00 120.00 WTI oil futures; $/bbl 97.41 90.00 90.00 100.00 110.00 JPM Dollar index; level 81.4 82.4 80.5 80.5 80.0
Exhibit 59: Most investors expect higher core inflation than they did at this time last year
Distribution of expectations for US core inflation over the next year from the J.P. Morgan Inflation Expectations Client Survey as of November 2011 and November 2010; %
40 Nov 11 30 20 10 11 0 01 3 7 19 18 20 12 9 5 4 6 4 Nov 10 31 33 24
The outlook for breakevens

Last year, we modified our long-term model for breakevens to use an illiquidity cost metric as one of the explanatory variables in the model, instead of trying to model liquidity-adjusted breakevens. This year, we make two additional modifications to our model. First, given that the Feds quantitative easing programs have had substantial impacts on inflation expectations and breakevens, we have added the size of the Feds securities holdings as an additional explanatory variable. Second, instead of using Nov-27 P-STRIPS asset swap spreads minus a barbell of 10-year and 30-year Treasury asset swap spreads as our metric for the cost of illiquidity, we are now using the spread between 5-year pre-refunded muni yields (taxable-equivalent) and 5-year Treasury yields. Pre-refunded munis are fully
0.5-0.99%
1.0-1.49%
1.5- 1.99%
collateralized with Treasuries, so they have effectively the same credit risk as Treasuries, but they trade cheaper because of lower liquidity. Thus, the pre-refunded muni/Treasury spread serves as a metric for liquidity. The other factors in our model are unchanged. We model 3-month forward breakevens to take out the impact of near-term carry, and we use four additional
197
2.5-2.99%
-0.49 - 0%
Above 3%
0-0.49%
2-2.49%
Below 0.5%
factors: 1) the level of nominal yields to account for the directionality between breakevens and nominal yields; 2) the unemployment rate, as a measure of slack in the economy; 3) the one-year-ahead forecast for the budget surplus as a percentage of GDP, as a proxy for fiscal policy-driven inflation pressures, and 4) realized headline CPI, since inflation expectations tend to be anchored to realized inflation itself. Exhibit 62 presents the statistics for our model for 5-, 10-, and 30-year breakevens. We can make several observations about breakevens based on this model. First, as expected, breakevens are positively correlated with nominal rates, with breakevens widening as nominal rates rise. Interestingly, long-end breakevens appear to be most sensitive to nominal rates, widening about 3.6bp for every 10bp rise in long-end nominal yields, while front-end breakevens are less sensitive to rates. On the other hand, front-end breakevens are most sensitive to the current level of the unemployment rate and realized inflation, as these factors have a greater impact on near-term inflation expectations. Forward-looking budget surplus expectations have a similar impact across the curve, with every 1% reduction in the annual deficit (roughly $160bn) causing breakevens to narrow about 9-10bp. We can use these betas to estimate the impact of future deficit reduction proposals on breakevens. For example, if market expectations shift to expect an additional $1tn of cuts over 10 years (above and beyond baseline expectations), that would equate to $100bn of cuts per year, or about 0.5% of GDP. Based on the betas in Exhibit 62, these expected cuts would lower the fair value for breakevens by about 4.5-5bp. As for the Feds balance sheet, we find that increases (due to quantitative easing) lead to wider breakevens,
Exhibit 62: Our updated model for breakevens
Nominal yields (% ) 5Y 10Y 30Y Beta T-stat Beta T-stat Beta T-stat 20.97 4.5 25.67 6.1 35.66 9.0
Exhibit 60: Fewer investors expect extreme outcomes for mediumterm inflation compared to last year
Distribution of expectations for US headline inflation over the next 2-5 years from the J.P. Morgan Inflation Expectations Client Survey; % 60 Nov 11 Nov 10 48 50 41 40 33 30 30
20 10 0 8
13
12
16
below target
around target
above target
sig. above target
Exhibit 61: and this decline in the dispersion of inflation expectations can be seen in other surveys
Average of top 10 responses minus average of bottom 10 responses for longrange consensus projection for year-over-year ago percent change in headline CPI (average rate for the next five years); %
1.6 1.4 1.2 1.0 0.8 0.6 Mar 03
Jul 04
Dec 05
Apr 07
Sep 08
Jan 10
May 11
Source: Blue Chip Economic Indicators
Statistics for 3-month forward breakevens regressed against nominal yields, the unemployment rate, the 3-month average of 1-year ahead budget surplus expectations as a percentage of GDP, year-over-year ago headline CPI, the size of the Feds balance sheet, and our illiquidity cost metric*; monthly data over 12/2002-10/2011
Unemployment Budget surplus as rate (% ) % of GDP (% ) -29.3 -4.2 -21.4 -5.2 -21.6 -7.1 -8.7 -2.4 -9.6 -4.0 -8.8 -4.4
Headline CPI, yoy rate (% ) 15.9 6.8 11.4 7.2 6.8 5.0
Fed's securities holdings ($bn) 0.034 4.4 0.029 5.6 0.023 5.1
Illiquidity cost metric (bp) -0.46 -7.1 -0.23 -5.1 -0.16 -3.8
Intercept (bp) 237.7 172.7 161.4
R-sq 85% 83% 85%
* 5-year muni pre-re yield (taxable equivalent) minus 5-year hot-run Treasury yield 198
with the effect strongest on the front end. Using the betas from our model, we can also estimate the impact of QE1 and QE2 on breakevens. Given that the Feds security holdings have grown by about $2.14tn since late 2008, when the Fed first began purchasing MBS, we estimate that the fair value of TIPS breakevens have risen by 72bp, 62bp, and 49bp, in the 5-, 10-, and 30-year sectors, respectively. Our model also implies that 5-year breakevens will benefit the most if the Fed embarks on QE3, with every $200bn in net purchases increasing the fair value of breakevens by 6.7bp. Five-year breakevens also appear likely to benefit the most if liquidity improves, or underperform the most if liquidity deteriorates. This suggests that in flight-toliquidity periods, the breakeven curve is likely to steepen, ceteris paribus. Putting it all together, we project the level of breakevens in 2012 based on our forecasts for the underlying variables and our expectation that our illiquidity cost metric declines to its average level over the past year. We also assume that the Fed does not expand its balance sheet via QE3, though there is upside risk to our fair value projections for breakevens if it does. As Exhibit 63 shows, we expect breakevens to be hit by quite a few negative factors: the unemployment rate will remain at elevated levels, budget deficit expectations should trend lower as fiscal tightening takes hold, and headline CPI should decline sharply from its current
Exhibit 65: Because of seasonality in inflation, the value of the inflation stub for TIPS varies over time
Exhibit 63: Our targets for breakevens in 2012
J.P. Morgan forecast for the drivers of forward breakevens and projected level of breakevens based on the model in Exhibit 15; bp Factors Current 1Q12 2Q12 4Q12 5Y UST yield, % 0.92 0.75 1.25 1.25 10Y UST yield, % 2.01 1.70 2.50 2.50 30Y UST yield, % 3.00 2.70 3.60 3.60 Unemployment rate, % 9.1 9.0 9.0 9.0 Budget surplus as % ge of GDP, % -7.5 -6.4 -6.1 -5.4 Headline CPI, oya % 3.72 1.46 1.10 1.40 Fed's security holdings ($bn) 2625 2625 2625 2625 Illiquidity metric, bp 91.8 80.3 68.7 45.7 Breakeven Targets Current 1Q12 2Q12 4Q12 3Mx5Y breakevens, bp 174.9 120 130 140 3Mx10Y breakevens, bp 205.3 160 175 180 3Mx30Y breakevens, bp 218.4 175 205 205
Exhibit 64: TIPS asset swaps have been negatively correlated with the level of breakevens
10-year TIPS asset swap spread regressed against 10-year TIPS breakevens and Nov-27 P-STRIPS asset swap spread; all data are 1-week averages; weekly data over past three years; bp
100 80 60 40 20 0 -20 0 50
Y = -0.31(10Y BE)+0.46(20Y P-STRIPS) + 79.0 R-sq = 77%
Value of the inflation stub at mid-month points over the calendar year for TIPS with different maturity months; % Inflation stubs for TIPS lines (by maturity month) For month Seasonal factor, starting with lag applied Jan Feb Apr Jul 15 Jan -0.28% 0 -0.28% -0.25% 0.48% 15 Feb -0.11% 0.27% 0 0.03% 0.76% 15 Mar 0.14% 0.38% 0.10% 0.14% 0.87% 15 Apr 0.26% 0.24% -0.04% 0 0.73% 15 May 0.31% -0.02% -0.30% -0.27% 0.47% 15 Jun 0.16% -0.33% -0.61% -0.58% 0.16% 15 Jul 0.06% -0.49% -0.77% -0.74% 0 15 Aug -0.10% -0.55% -0.82% -0.80% -0.06% 15 Sep -0.16% -0.45% -0.72% -0.70% 0.04% 15 Oct -0.09% -0.29% -0.57% -0.54% 0.19% 15 Nov -0.07% -0.20% -0.48% -0.45% 0.28% 15 Dec -0.13% -0.13% -0.41% -0.38% 0.35%
Note: We estimate the seasonal factor as the difference between non-seasonally adjusted changes in headline CPI minus seasonally-adjusted changes, averaged by calendar month over 2001-10. This series is lagged by two months to account for the lag in TIPS indexation and interpolated to get mid-month values.
100 150 200 10-year TIPS breakeven; bp
250
300
level. In addition, we look for nominal rates to plunge in 1Q12 as the European crisis worsens, and then rise over the rest of the year. Thus, early next year, negative factors will dominate, and we expect breakevens to narrow sharply. As the year progresses, however, breakevens should widen out as rates rise. Thus, based on this analysis, we recommend underweighting TIPS versus nominals over the next few months. Looking beyond 1Q12, we would look to initiate breakeven wideners.
A model for TIPS asset swaps

How are TIPS asset swaps likely to perform in 2012? To answer this question, we present a simple model for TIPS asset swaps in Exhibit 64. Since TIPS asset swaps have no inflation risk due to the embedded inflation swap,
199
they should trade similarly to asset swaps on other lessliquid sectors of the Treasury market. Thus, we use 20year P-STRIPS asset swaps as one of the explanatory variables in our model. In addition, although TIPS asset swaps should not have inflation risk, we find that they have been correlated with TIPS breakevens over recent years, with asset swaps narrowing (i.e., TIPS richening versus swaps) when breakevens widen. This is likely because demand for TIPSwhich will impact asset swapshas been correlated to inflation expectations and thus breakevens. Given our expectation that breakevens will narrow over the next few months and that liquidity will likely remain constrained, we look for TIPS asset swaps to widen modestly, though longer-maturity asset swaps will likely retrace that widening over the course of next year.
Exhibit 66: We estimate that the April-July seasonality differential is worth about 36bp of yield for 2-year TIPS
Difference in value of the inflation stub for various pairs of TIPS (in price terms) as given in Exhibit 23, and in yield terms for various maturities
Price impact (% ) 2Y 5Y 10Y 20Y
Jan-Apr 0.25% 12.2 4.8 2.5 1.3
Jan-Jul -0.49% -23.9 -9.4 -4.8 -2.5
Apr-Jul -0.73% -36.1 -14.2 -7.3 -3.8
Yield impact, by maturtiy (bp)
Exhibit 67: Projected value of seasonality differentials for TIPS with different maturities over time
Difference in value of the April-July inflation stub for bonds with different maturity years; bp of yield
0 -10 -20 -30 -40 -50 -60 May 10 2013 2015 2017 2019 2021 2031 Aug 10 Dec 10 Mar 11 Jun 11 Sep 11
Trading seasonality in TIPS

Since TIPS are indexed to headline CPI, which is not seasonally adjusted, TIPS are exposed to seasonality. Furthermore, for a given TIPS issue, the value of inflation seasonality varies at different points in the year, even if we assume the seasonal pattern in inflation remains the same over time. To see this, we consider a hypothetical TIPS with a January 15 maturity date. If we look at this TIPS on November 15, we can think of inflation over the remaining life of the bond as n full years of November 15-November 15 inflation (with a net seasonality component of zero), plus an inflation stub for a November 15-January 15 period. Similarly, if we look at the same bond on March 15, we can think of it accruing n-1 full years of March 15-March 15 inflation (with a net seasonality component of zero), plus a March 15-January 15 inflation stub. In Exhibit 65, we show the value of those inflation stubs for the different TIPS maturity months over the course of a calendar year. Here we estimate the seasonal factors by taking the difference between non-seasonally adjusted changes in headline CPI minus seasonally-adjusted changes and averaging the differences by calendar month over 2001-10. We also lag the series by two months to account for the lag in TIPS indexation, and then we interpolate to get mid-month values. As the table shows, some TIPS lines benefit from seasonality more than othersin particular, the inflation stubs for July TIPS are almost always positive, while the stubs for February TIPS are almost always negative.
200
Exhibit 68: The flattening of the Apr-13/Jul-13 TIPS yield curve has been roughly in line with what we would expect given the change in the value of the seasonality differential
Apr-13/Jul-13 TIPS real yield curve versus projected value of the April-July seasonality differential; bp of yield
15 10 5 0 -5 -10 -15 -20 -25 -30 May 10 Aug 10 Nov 10
Value of seasonality differential
-20 -25 -30 -35
Apr-13/Jul-13 TIPS curve
-40 -45 -50 -55 Mar 11 Jun 11 Sep 11
Although the inflation stubs for individual TIPS vary over time, the difference between stubs for TIPS with different maturity months (i.e., January versus April)
Exhibit 69: TIPS breakevens have closely tracked equities over the past two years
10-year TIPS breakevens versus S&P 500; bp level
280 260 240 220 10-year TIPS breakevens
1400
1300
TIPS is the benchmark 5-year and enjoys greater liquidity, while the Jul-16 TIPS is an old 10-year issue that is less liquid. As a result, the fact that the spread between these two issues is almost zero likely indicates that the liquidity premium for Apr-16 TIPS counters the seasonality disadvantage. In addition, at the very front end of the TIPS curve, investors near-term inflation forecasts are likely to have a greater influence on spreads than backward-looking estimates of seasonality. With these constraints in mind, we think it is still possible to find instances where investors can take advantage of seasonal differentials. For example, we estimate that Jul-16 TIPS should trade about 11bp rich to Jan-16 TIPS based on seasonals, but they are currently about 1bp cheap. In addition, this curve currently looks too steep relative to the broader Jan-15/Jan-19 curve, and liquidity differentials should be a minor issue, since both
Exhibit 70: Going short breakevens versus the S&P 500 when breakevens look wide has been more profitable than going long breakevens when breakevens look narrow
1200 200 180 160 140 1000 Nov 09 Feb 10 Jun 10 Sep 10 Dec 10 Mar 11 Jul 11 Oct 11 S&P 500 1100
stays constant, assuming that the seasonal pattern is unchanged. Thus, we can estimate the difference in the value of seasonality for various pairs of TIPS, as shown in Exhibit 66. As the table shows, April TIPS are most adversely affected by seasonality, accruing 0.25%-pt less inflation than January TIPS, and 0.73%-pt less inflation than July TIPS. We can also translate this seasonality differential into yield terms by dividing by duration. As the table shows, the April-July seasonality differential is worth 36bp of yield for 2-year TIPS, but only about 4bp of yield for 20-year TIPS. To see how the value of this seasonality differential (in yield terms) changes as bonds age, we project the differential for April and July TIPS with various maturity years in Exhibit 67. For longer-maturity TIPS, the change in the value of the seasonality differential over time is small: for example, for TIPS maturing in 2019, the differential was worth 8.4bp of yield 18 months ago, and is currently worth 9.5bp. In contrast, for TIPS maturing in 2013, the differential was worth about 25bp of yield 18 months ago, and is currently worth 51bp. We can see the impact of this seasonality effect on the Apr13/Jul-13 TIPS real yield curveas Exhibit 68 shows, the flattening of this curve over the past 18 months has been roughly in line with what we would expect given the change in the value of the seasonality differential. In practice, it may be difficult to isolate the value of seasonal differentials, since the spread between two issues will also be impacted by the slope of the overall curve as well as liquidity differentials. For example, even though April TIPS have a seasonal disadvantage compared to July TIPS, in the 2016 sector, the April
Statistics for long or short breakeven trades initiated when the residual* reached a certain level and hedged with the opposite position in S&P 500**; 11/09-10/11
Long BE if # Avg P/L # Hit Avg gain Avg loss residual (bp) < trades (bp) winners ratio (bp) (bp) -10 -15 -20 -25 Short BE if 5 10 15 20 125 85 42 11 # 130 70 32 10 -0.6 -0.5 1.9 6.4 Avg P/L (bp) 6.8 9.3 16.3 18.0 64 45 26 10 # 82 49 27 10 51% 53% 62% 91% Hit 63% 70% 84% 100% 9.5 8.7 8.8 8.3 (bp) 15.8 17.3 20.2 18.0 -11.2 -10.8 -9.4 -11.8 (bp) -8.6 -9.4 -4.4 N/A
Avg gain Avg loss
residual (bp) > trades
winners ratio
* Backward-looking 3-month residual of the level of breakevens versus the level of S&P 500. ** The hedge ratio is given by the backward-looking 3-month beta of weekly changes in breakevens versus weekly changes in the S&P 500. Note: P/L calculated as change in breakevens minus (hedge ratio)*(change in S&P) plus 1month ex-ante breakeven carry for long breakeven positions, and the inverse of that quantity for short breakeven positions.
Exhibit 71: Going long breakevens when carry is attractive and breakevens look narrow has been profitable
Statistics for long 5-year breakeven trades (held for one month) initiated when exante 1-month carry was above a certain level and the rolling 6-month residual of 5year breakevens regressed against 5-year nominal yields and WTI oil futures prices was below a certain level; 1/09-10/11
Carry Residual (bp) > 3 3 5 5 (bp) < 0 -2 0 -2 # trades 88 77 56 47 Avg carry Avg P/L (bp) 8.0 7.7 10.4 10.3 (bp) 6.9 6.5 6.2 5.6 # 56 48 34 28 Hit 64% 62% 61% 60% Avg 19.1 19.4 17.7 17.4 Avg loss (bp) -14.4 -14.9 -11.6 -11.9
201
winners ratio gain (bp)
issues are old 10-year TIPS. Thus, we think overweighting Jul-16 TIPS versus Jan-16 TIPS is currently an attractive way to take advantage of seasonality differentials. Such trades, based on the theme of identifying seasonality mispricings, will be a key element of our TIPS strategy in 2012.
Exhibit 72: and this strategy works even better with inflation swaps
Statistics for receiving inflation/paying fixed positions in 5-year or 3-year inflation swaps (held for one month) when ex-ante 1-month carry was above a certain level and the rolling 6-month residual of 5-year or 3-year inflation swaps regressed against 5-year or 3-year nominal yields and WTI oil futures prices was below a certain level; 1/09-10/11
Carry Residual (bp) > (bp) < 5-year swaps 3 3 5 -6 -14 -6 # trades 63 25 45 29 42 25 27 20 Avg carry Avg P/L (bp) 7.6 7.2 9.1 10.9 16.5 17.2 18.3 21.2 (bp) 15.2 15.7 12.6 10.3 13.0 14.8 14.7 12.7 # Hit Avg Avg loss (bp) -9.4 -14.7 -7.8 -9.1 -8.1 -7.8 -8.9 -8.9 winners ratio gain (bp) 54 22 38 23 32 20 22 15 86% 88% 84% 79% 76% 80% 81% 75% 19.3 19.9 16.4 15.4 19.5 20.4 20.1 19.9
Relative value trading themes for 2012

Trading breakevens versus the S&P 500 Although we expect breakevens to narrow in early 2012 as the European crisis worsens, one way we can protect against being wrong is to trade breakevens on a hedged basisnamely, by trading breakevens versus equities. Over recent years, the correlation between breakevens and risky assets has been striking, with breakevens widening in risk-on trades and narrowing in risk-off trades. In particular, breakevens have closely tracked the S&P 500 over the past two years (Exhibit 69). Given this strong correlation, we would expect trading breakevens versus a beta-weighted amount of equities when breakevens look either very narrow or very wide relative to equities to be an attractive trading theme. We test this hypothesis by back-testing trades over the past two years as follows: each day, we calculate the backward-looking 3-month residual of the level of breakevens versus the level of S&P 500. If the residual is below (above) a certain threshold, we initiate a long (short) breakeven position and hedge with the opposite position in the S&P 500. The hedge ratio is given by the backward-looking 3-month beta of weekly changes in breakevens versus weekly changes in the S&P 500. Exhibit 70 presents the results of our analysis. We would expect the strategy to work equally well when breakevens look either too wide or too narrow, but we find that over this period, going long breakevens hedged for the S&P when breakevens look narrow has generally been unattractive, unless breakevens look extremely narrow. In contrast, going short breakevens on a hedged basis when breakevens look too wide has been profitablethese trades have had high hit ratios and attractive P/Ls. This asymmetry could be a result of our observation that equities and inflation expectations have been more correlated during deflationary periods than inflationary periods in recent years (see US Cross Sector Overview). Thus, one trading theme we like for 2012 is opportunistically selling breakevens versus a hedged amount of S&P 500 futures when breakevens look too rich versus equities.
202
7 -6 3-year swaps 7 7 9 11 -6 -10 -8 -8
Earning carry when breakevens look narrow A second trading theme we like is opportunistically going long breakevens when carry is attractive and breakevens look narrow relative to fair value. To see this, we back-tested a strategy under which long breakeven trades were initiated for a 1-month period when 1) 1month carry exceeded a certain level and 2) the backward-looking 6-month residual of breakevens regressed against nominal yields and oil prices fell below a certain threshold. In Exhibit 71, we present statistics for these trades over January 2009-October 2011. As the exhibit shows, this strategy generated hit ratios around 60% and average 1-month P/L of 5-7bp. We found that this strategy was even more effective using inflation swapsspecifically, receiving inflation/paying fixed when carry was attractive and inflation swap rates looked too low relative to fair value. As Exhibit 72 shows, using inflation swaps produced significantly higher hit ratios and average P/Ls than the strategy using breakevens. Even if we account for the wider bid-offer in inflation swaps, the derivatives version of this strategy still appears more attractive than the cash version. Thus, a second trading theme we like is receiving inflation/paying fixed in swaps when carry is attractive and inflation swap rates look too low.
Medium term themes for inflation markets Euro area

Favour breakeven flattening trades Even if inflation should fall, we see upside risks to prints in 1H12 vs. expectations, while the long end likely to be driven by sovereign risk dynamics which should flatten nominal yield curves. Favour 1s/10s flatteners in HICP swaps, overweight OBLi 13 vs. BTPei 19 breakevens or BTPei 16s vs. BTPei 21 breakevens. Overweight French CPI-linked vs. Euro HICPlinked lines into early 2012 French-CPI linked should perform well in 1Q12 when we expect further hedging from Livret A accounts and support from Libor-based domestic investors. Italian real yields to be under pressure Real yields can reach 10% on the back of our peripheral spread views as well as risks that Italian linkers are dropped from linker benchmarks.
US
Look for TIPS breakevens to narrow sharply over the next few months and then widen modestly over 2012 Breakevens will hit a number of headwinds in early 2012: we expect nominal yields to plummet, fiscal policy to tighten, and headline inflation to fall. Thus, our model projects that breakevens should narrow over the next few months, though breakevens should widen over the remainder of the year as nominal rates rise. QE3 also poses an upside risk to our targets. Overweight TIPS with attractive seasonals versus TIPS with unattractive seasonals when mispricings exist We recommend taking advantage of seasonality differentials among TIPS that are close in maturity when mispricings exist. Opportunistically trade breakevens versus the S&P 500 when breakevens look too wide With correlations expected to remain high as long as European risks remain significant, trading breakevens hedged for moves in the S&P 500 is likely to prove attractive, as it has in recent years. Look to tactically trade breakevens and inflation swaps for carry We find that strategies involving receiving inflation when carry is attractive and swap rates look low have been profitable over recent years and recommend that investors use these strategies to earn carry.
UK
Be long real yields in intermediate part of the curve Our model predicts 5Y real yields to fall 30-40bp from current levels. Long positions in 5Y and 10Y linkers offer attractive carry and slide vs. the forwards. Short 10Y breakevens Nominal gilts will continue to be the major driver of breakevens. Our predicted 75bp fall in 10Y gilt yields during 1H12 (to the 1.50%) level translates into a 30bp fall in ILG22 breakevens, to the 2.30% level. Go short real yields into linker syndications as a tactical trading strategy A trading rule selling the closest maturity bond to the linker being issued via syndicate 15 business days ahead of syndications has been profitable.
Cross market themes

Generally overweight US inflation breakevens vs. Euro area equivalents in 10Y+, vs. doing the opposite at the front end This is based on our relative inflation views, the risk of sovereign dynamics impacting Euro breakevens more than US equivalents.
203
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Sally AuldAC (61-2) 9220-7816 sally.m.auld@jpmorgan.com J.P. Morgan Securities Australia Limited
Australia
J.P. Morgan economists expect the Australian economy to grow 3.0% in 2012, but risks are biased to lower growth outcomes given the fragile global environment. Against this backdrop, J.P. Morgan economists expect the RBA to deliver 50bps of easing in Q1 next year While the domestic economic outlook might suggest the potential for modestly higher yields, the experience of 2011 suggests that domestic drivers may not be the main influence on term yields As such, the bearish outlook we retain on Europe means that we recommend extending duration on dips in AUD fixed income. Look to lighten duration in 2H12, given the risks that the market prices in a more optimistic global outlook We target the 3s/10s bond curve spread to make new highs in 2012. Investors trading long duration should implement cash bond curve flatteners in order to gain some portfolio protection We think swap spreads will be biased wider in 2012, with the spread curve likely to flatten. We like owning cash spread product against swap rather than bond, and prefer NSWTC in the semi-government space Look to buy the NSWTC Nov-2035i against the ACGB Sep-2030i. The ACGB Aug-2020i looks cheap against the UST Jul-2021i; real yield spreads look wide given AUS-US growth differentials
Exhibit 1: Australias (export weighted) trading partner growth forecast to be below trend in 2011 and 2012
Calendar year % growth
7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 1995 1998 2001 2004 2007 2010 Av erage
Exhibit 2: Movements in the RBA cash rate generally correlate well with global GDP growth
%-pts Qtrly. annualised %
2.5 1.5 0.5 -0.5 -1.5 -2.5 -3.5 -4.5 -5.5 Mar-98 RBA cash rate, annual change, lhs Global GDP, rhs, pushed fw d 2q Dec-00 Sep-03 Jun-06 Mar-09 Dec-11
8.0 6.0 4.0 2.0 0.0 -2.0 -4.0 -6.0 -8.0 -10.0
Economy expected to improve in 2012, but global headwinds are fierce

2011 was a year of disappointment on the macroeconomic front, at least relative to expectations. Consensus forecasts now expect that Australia will likely record growth of 1.5-2.0% in 2011, well below trend. This was a lower number than most expected at the beginning of the year, but the disappointment was largely due to the impact of severe floods in Queensland and a slower-than-anticipated recovery from the natural disaster.
The outlook for Australia in 2012 is more optimistic, with the ongoing recovery in coal exports expected to see solid growth momentum in 1H12. J.P. Morgan expects growth of 3.0% for the calendar year; the RBA expects 4.0% growth, although we would argue that this looks a little ambitious given risks surrounding the European (and global) outlook. While Australias direct trading links with Europe are small (around 8% of Australian exports go to Europe), the indirect links via Asia are reasonably strong. As the RBA noted in its most recent quarterly Statement on Monetary Policy: The possibility of a sharp economic deterioration in Europe represents a downside risk for the Australian economy. Given strong trade links with Asia, it is likely that Australia would be less directly affected than some other countries by a deterioration in Europe, although the economy would still be affected through falls in asset
204
prices and weaker household and business confidence. Commodity markets could be expected to weaken and growth in domestic incomes would be lower. The clear message is that Australia will not be immune to a severe economic contraction in Europe. Against this backdrop, J.P. Morgan economists are forecasting the RBA to deliver modest easing in early 2012 and expect a cash rate of 3.75% by March next year. The risks to this view are clearly biased to the downside, given fragility in the global growth environment and risks around an escalation in the European crisis. Indeed, our estimate of Australias trading partner growth in 2012 suggests little change from 2011 (Exhibit 1), with trading partner growth at 4.2% (below the average of 4.9%). It is also worth noting that movements in the RBA cash rate generally correlate well with the global growth outlook (Exhibit 2). Of course, the main medium-term influence on the Australian macro-economic outlook continues to be China, and while our economists have recently shaved their growth forecasts for China, they still expect a reasonably positive outcome in 2012 (growth of 8.3%). As Exhibit 3 illustrates, Australian and Chinese GDPs have been well correlated over the past 5-6 years, with growth in China leading Australian growth by around 3 months. Although our expectation for Chinese economic growth would be historically consistent, with Australian growth around 2.0%, the consensus expects strong capex growth and a rebound in flood-affected exports to temporarily drive a divergence between Australian and Chinese growth outcomes (hence the disconnect between Australian and Chinese growth forecasts in Exhibit 3). Overall, our forecasts for Australian growth next year represent a cautiously optimistic outlook, especially given global risks and our forecasts for another year of below-trend trading partner growth. Accordingly, we believe risks are biased to the downside to our growth forecasts. This is especially the case given a reasonably subdued pulse in the domestic dataflow of late.
Exhibit 3: Chinese and Australian GDP growth have been well correlated over the past 6 years
Dotted lines represent J.P. Morgan forecasts oya % oya %
16.0 14.0 12.0 10.0 8.0 6.0
Chinese GDP, lhs, pushed fw d 1q Australian GDP, rhs
6.0 4.9 3.8 2.7 1.6 0.5
Dec-04 Apr-06 Aug-07 Dec-08 Apr-10 Aug-11 Dec-12
Exhibit 4: The market is pricing in over 150bp of rate cuts from the RBA
AUD OIS curve, %
4.75 4.50 4.25 4.00 3.75 3.50 3.25 3.00 2.75 2.50 Current Feb-12 Apr-12 Jun-12 Aug-12 Oct-12
the months ahead. But the market is already pricing in this risk; Exhibit 4 illustrates that the OIS curve is pricing in 165bp of rate cuts over the next 12 months. Exhibit 5 illustrates that since early August, the market has oscillated between pricing in 100bp and 165bp of rate cuts from the RBA (over a 12-month horizon). This is a distinct change from the period JanuaryAugust 2011, when the market priced anywhere from 0-50bp of rate hikes. Investors may be tempted by the attractive carry in paid positions in the very front end of the Australian curve; for example, a paid position in AUD 1Y swap currently offers 30bp of carry and roll over a 3-month horizon. But assuming the RBA does not validate current market pricing anytime soon and with no end-game solution in
205
Can the RBA validate front-end pricinganother 165bp of rate cuts?

As noted above, J.P. Morgan economists expect the cash rate to reach 3.75% by March 2012. We still think the risks to this view are to the downside, given the situation offshore. And while the RBA may not yet believe it has begun a traditional easing cycle, we would expect the RBAs policy stance to retain a degree of optionality in
Europe on the horizon, we see little reason for current market pricing to change materially into 2012 given: 1) Ongoing concerns surrounding Europe, China and global growth; 2) Data suggesting that while the Australian economy is not on the verge of a recession, nor is it showing any sign of a significant uptick in momentum; 3) Technical aspects to the OIS market which currently bias yields lower (domestic banks are structural receivers of OIS given their need to hedge deposit books); and 4) The fact that the RBA could if needed cut rates by a lot, relative to other developed market central banks. Interestingly, the cycle trough in 3Y yields has often been a good indicator of the subsequent trough in the cash rate (2001 and 2008). Exhibit 6 illustrates that with 3Y government bonds currently yielding 3.25%, this would be consistent with a further 125bp of easing from the RBA. It also implies that there is probably 4050bp of risk premium in the very front end of the AUS yield curve at present.
Exhibit 5: and has been pricing in cuts since early August

bp of tightening* priced in over the next 12 months
75 25 -25 -75 -125 -175 Jan-11 Apr-11 Jul-11 Oct-11
* negative number implies easing
Exhibit 6: Even term yields suggest the possibility of more easing the level of 3-year bond yields suggest the risk of rate cuts, if history is any guide
%
8.0 7.0 6.0 5.0 4.0 3.0 2.0 Jan-00
3-y ear bond y ield RBA cash rate
The outlook for outright yields loweryield highs and lower-yield lows in 2012 as domestic fundamentals continue to take a back seat
Broadly speaking, the trend over the course of 2011 was towards lower yields in Australia, with a rally in global bond markets, some disappointment on the domestic and global growth fronts, and a November rate cut from the RBA being the main drivers of lower yields. A persistent bid from offshore investors also helped to suppress yields (see our analysis below for more detail on this theme). Australian 3Y bond yields have traded a 215bp range over the course of 2011, and at 3.23% at the time of writing, sit at the lows of the year (Exhibit 7). Remarkably, the calendar year yield high for 3Y bonds has been very similar over the past 3 years, at around 5.35%. Exhibit 7 illustrates the structurally lower trading range for 3Y yields post the financial crisis; with offshore risks mounting, we think it is likely that Australian yields post new lows in 2012.
206
Feb-02
Mar-04
Apr-06
May -08
Jun-10
Exhibit 7: 3-year government bond yield calendar year trading ranges highlight a structurally lower trading range post crisis and the risk of new yield lows in 2012
Trading range for 3Y government bond yields over the course of 2011; %
7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Even the 10Y bond has traded within a 175bp range over the course of the year, with a yield high for the year of 5.75% also close to the yield highs seen in 2009 and 2010 (Exhibit 8). With outright yields at the lows for the year (and not far from the yield-lows seen since 2009) and with J.P. Morgan economists expecting a pick-up in growth momentum next year, it might be intuitive that shortduration strategies are the correct view on Australian bonds in 2012. However, the message from 2011 has been that domestic fundamentals have not been the main driver of yields over the course of this year indeed, Exhibit 9 illustrates the divergence between domestic fundamentals and the level of 3Y bond yields. Against this backdrop, we think there are a number of reasons why investors are not likely to be rewarded for shortduration strategies in the coming year: 1) Our forecast for US 10Y yields does not envisage a significant selloff, at least for the first half of calendar year 2012. After an early rally to 1.7%, we expect the UST 10Y yield to oscillate around the 2.5% level for much of the year, limiting the extent to which Australian bond yields are likely to rise. 2) Domestic real money accounts are already running short-duration positions and have done so for much of 2011 (Exhibit 10), and as such, we believe there will be little appetite to shorten the duration of portfolios further from current levels. 3) Our European rates strategy team remains extremely cautious on the outlook for the European peripherals in 2012. We think developed market bond yields will struggle to selloff significantly in such an environment. Furthermore, this backdrop will reinforce the attractive risk/reward in owning AAA-rated Australian sovereign risk at (relatively) attractive yield levels. 4) Following on from point 3, we expect offshore demand for Australian government bonds to remain solid in 2012. As we note below, 71% of ACGBs are owned by foreign investors (as of June 2011).
Exhibit 8: The 10Y bond has traded within a 175bp range over the course of the year, with a yield high for the year of 5.75% also close to the yield highs seen in 2009 and 2010
Trading range for 10Y government bond yields over the course of 2011; %
7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Exhibit 9: Australian bond yields have not traded in line with domestic fundamentals in 2011
% oya %
7.0 6.0 5.0 4.0 3.0 2.0 Mar-99 Dec-01 Sep-04 Jun-07 Mar-10 3y bond y ield, lhs Real GDP + CPI (2q av e.), lagged 2 qtrs, rhs
9.0
6.0
3.0
0.0
Exhibit 10: but this has not prevented domestic real money managers from trading short duration for most of 2011
Actual portfolio duration relative to benchmark index, %
110 106 102 98 94 90 86 Nov -06 Dec-07 Jan-09 Feb-10 Mar-11
207
Of course, our fair value models for both 3Y and 10Y government bonds continue to suggest that current yield levels for AGCBs should be regarded as rich. Exhibit 11 shows our fair value for 10Y bond yields. For 2012, we see fair value around 4.75%. But with higher-than-usual levels of uncertainty persisting offshore and ongoing demand for yields from foreigners, the valuation richness of the ACGB market may be a constant theme in 2012. In 2011, investors have not been paid to position portfolios on domestic fundamentals alone, and we suspect this will continue to be the case in 2012 given our global strategy views. Indeed, movement in a weighted European peripheral 10Y swap spread has accounted for almost 86% of the moves in AUD 1y/1y swap over the course of 2011. Our preferred duration strategy for 2012 would be to use any selloff in bond markets to enter long-duration portfolio structures. Indeed, we expect the next couple of months to provide good opportunities for investors to initiate long-duration positions. Short-risk trades are likely to be unwound going into the year-end as money managers, hedge funds, and dealers alike pare back their balance sheet. Towards 2H12, we would look to lighten duration given the risk that the market could start to price in a more optimistic global outlook. Against the US, we see Australian 10Y bonds as cheap at present. Our medium-term model of the AUS-US 10Y bond spread suggests that the fair value for the spread is probably around the 175bp level for 2012 (Exhibit 12). The spread is currently trading at 205bp, and we suspect that AUD 10Y bonds will be well bid on any widening of this spread. Increasingly, relative (and not absolute) yield levels are becoming the driver of cross-border bond market flows.
Exhibit 11: Our fair value model suggests current 10Y government bond yields are rich. For 2012 we see fair value around 4.75%
10Y government bond yield, actual and model*; %
16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 Sep-88 Mar-93 Sep-97 Mar-02 Sep-06 Mar-11 Actual Model +/- 2 s.d.
* model is 10yr yield = 2.52 + 0.60*90-day rate + 0.31*moving avg. of nominal GDP 4.49*% of ACGBs held offshore; R-squared is 0.92, standard error is 0.72ppts.
Exhibit 12: Australian 10Y bonds still look cheap vs. US 10Y notes on our medium-term model
Quarterly data for spread between 10Y Australian government bonds and US Treasuries, actual and model*; bp
300 Actual 250 200 150 100 50 0 Jun-96 Apr-99 Feb-02 Dec-04 Oct-07 Aug-10 Model
* model is 10yr spread = 58.13 + 5.32*real GDP growth spread (pushed fwd. 4q) + 30.98*1y swap yield differential (%-pts); R-squared is 0.83, standard error is 0.29ppts. Latest read for the actual spread is 30-Sep-11.
Offshore investment becomes a key feature of Australian fixed income markets and is here to stay
One of the features of the Australian government bond market in recent years has been strong demand for ACGBs by offshore investors. Indeed, foreign buyers have accounted for much of the additional supply in ACGBs (Exhibit 13). Offshore holders now own 71% of ACGBs, just shy of all-time highs (Exhibit 14). The solid demand for ACGBs in recent years reflects a number of factors:
Exhibit 13: Foreign buying of ACGBs has accounted for much of the additional supply in recent years
AUDbn
200 ACGBs (offshore holdings) 150 Total ACGBs outstanding
100
50
0
208
Dec-88 Jun-92
Dec-95 Jun-99
Dec-02
Jun-06
Dec-09
1) Reserve diversification by central banks into AUD; 2) A largely non-leveraged investor base with longer investment horizons (central banks and offshore pension funds) and a small level of holdings in absolute terms pre-crisis ($34.4bn as of December 2007); 3) Wide yield differentials with most developed market economies and outperformance by the Australian economy (suggesting potential for AUD appreciation); and 4) A strong fiscal situation in Australia. When we look at the relative rankings of AAA-rated sovereigns1, Australia should definitely be considered a strong AAA-rated sovereign. And when one considers the additional yield pick-up, it is not difficult to see why offshore capital has found a home amongst ACGBs (Exhibit 15). As we have noted above, ongoing foreign demand for ACGBs is one reason we do not expect a significant selloff in AUD rates. Indeed, we can estimate the extent to which changes in offshore participation can impact the level of bond yields. As we note in the footnote to Exhibit 11, we include the proportion of ACGBs held by offshore investors as an explanatory variable in our medium-term, fair-value model for Australian 10Y bond yields. Our other inputs are a moving average of nominal GDP growth (to capture the influence of domestic economic fundamentals on long-end yields) and the 90-day rate (to capture the influence of short-term interest rates on the level of term yields). The model statistics are outlined in Exhibit 16. The model suggests that a 1-ppt rise in offshore ownership will see 10Y yields decline 4.5bp. This suggests that the rise in offshore holdings though 2010 and 2011 has lowered 10Y yields by around 34bp, all other things being equal. We think that much of rise in offshore demand for ACGBs has been driven by central bank reserve diversification. This would be consistent with recent data
1
Exhibit 14: Offshore holders now own 71% of ACGBs, just shy of alltime highs
Proportion of ACGBs held by offshore investors; %
80 70 60 50 40 30 20 10 0 Sep-88 Jun-93 Mar-98 Dec-02 Sep-07
Exhibit 15: Australia is one of the stronger AAA-rated sovereigns, yet it has the highest 10Y bond yields among all AAA-rated peers*
10-year sovereign bond yield vs. standard deviations away from AAA median**
4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 -0.6
Weaker AAA NZ AUS
Stronger AAA
UK
GER CHF -0.3 0.0 0.3 0.6 0.9 1.2
Std. Dev s. aw ay from AAA median

* we include New Zealand even though Moodys is the only rating agency to give the sovereign an AAA credit rating. ** based on a model based on real GDP growth and 1Y swap yield differentials.
Exhibit 16: Our models for 10Y yields suggest a 1% increase in offshore ownership will result in yields declining 4.5bp
Model statistics for our 10Y ACGB model Model Statistics
Intercept Nominal GDP (moving ave.) 90-day rate % of offshore holdings R-squared Standard error (ppts) Value 2.5 0.6 0.3 -4.5 0.92 0.7
For more details on the analysis behind the ranking of AAA-rated sovereigns see S. Auld, The Antipodean Strategist, 18 August 2011
from the IMFs composition of foreign exchange reserves, which suggests that non-G4 currencies now account for 5% of reserves, up from just over 2% in
209
2009. And even with much of the new allocation to AUD now out of the way, quarterly net buying of ACGBs still remains well above average. But the other driver of offshore demand for ACGBs continues to be the wide yield differential both in nominal and real terms between Australian government bonds and bonds in most developed market economies. Exhibit 17 shows that the 10Y real yield spread between Australia and the US has a leading relationship with changes in offshore holdings of ACGBs. If anything, the chart suggests that current levels of foreign demand look a touch low relative to rate spreads. And with the outlook for G4 government bond markets consistent with relatively low yields for a long time, we expect the real money carry trade to support foreign sponsorship of the ACGB market.
Exhibit 17: Real yield spreads are a key driver of offshore demand for ACGBs
%-pts %-pts
30 25 20 15 10 5 0 -5 -10 -15 Sep-97
Annual change in offshore holdings of ACGBs, lhs 10y AUS-US real y ield spread, rhs, 6 mths fw d
3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0
Aug-00
Jul-03
Jun-06
May -09
Curve spreads: the 3s/10s spread has scope to steepen further

Generally speaking, the shape of the yield curve over time in Australia is determined by movements in the RBAs cash rate. Exhibit 18 illustrates the point investors who think the RBA will validate market pricing should clearly position for a steeper curve spread. Indeed, our fair value model for the 3s/10s bond spread suggests that the curve has traded flat relative to fair value for some time (Exhibit 19). There are possibly two reasons for the persistent cheapness of the curve it could be a reflection of strong demand for AUD duration from offshore (as discussed above), or the negative carry inherent in curve steepeners may prevent investors from positioning for a steeper curve. Given our expectation of rate cuts from the RBA, we believe the 3s/10s bond spread will make new highs for the cycle in Q1 next year and target a move to +105bps. The carry and roll associated with cash bond steepeners is very negative. Accordingly, we think there is better value in entering forward starting steepeners; forward curve spreads are very flat, suggesting that investors biased towards steeper curves should easily beat the forwards. The least carry and roll negative forward starting steepener which offers some roll-up to spot is the 3s/7s swap curve steepener, 1Y forward. The 1Y forward spread is around 4bps flatter than the spot spread, and the carry and roll on the trade costs 4.8bps over a 3-month
210
Exhibit 18: The policy cycle tends to be the medium term driver of movements in the 3s/10s bond curve in Australia
%-pts %
1.6 1.2 0.8 0.4 0.0 -0.4 -0.8 Aug-98
AUS 3s10s curv e, pushed fw d. 3 months, lhs RBA cash rate, rhs, inv erted
2.00 3.00 4.00 5.00 6.00 7.00 8.00
Jun-01
Apr-04
Feb-07
Dec-09
Exhibit 19: The 3s10s bond curve still looks a touch flat relative to our fair value model
3s/10s bond curve actual and model*; %-pts
1.75 1.25 0.75 0.25 -0.25 -0.75 -1.25 Oct-01 Apr-04
Actual Model
Oct-06
Apr-09
Oct-11
* model is 3s/10s bond spread = 2.20 + 0.07*1s6s AUD OIS curve (%-pts) - 0.41*3y bond yield + 0.18*US2s10s bond spread; R-squared is 0.88, standard error is 0.15ppts.
horizon. The 1Y forward spread is currently 66bps; we target a move to +90bps.
Exhibit 20: Even though the 3s/10s bond curve has shown some correlation to peripheral spreads this has been reasonably limited
3s/10s bond curve vs. 10Y weighted peripheral swap spread*, %-pts and bp
Investors already trading long duration should look to use cash bond flatteners as a means of portfolio hedging
Given the prospect of difficult trading conditions in 2012, we think it is wise to think about trades which are 1) positive carry, and 2) not so correlated with moves in European peripheral spreads. In the Australian market, carry-efficient curve trades offer good opportunities for investors trading long duration looking for some portfolio diversification or protection in the event that peripheral spreads narrow. Exhibit 20 illustrates that the 3s/10s curve has been reasonably invariant to moves in peripheral spreads over the course of the year. This is probably contrary to expectations most would have expected the curve to bull-steepen in a risk-averse environment and when the RBA is easing policy. As we note above, we think the most logical explanation for this is constant demand for AUD duration from offshore. The positive carry associated with cash bond flatteners means that curve flattening trades can offer some protection for investors trading long duration. In the cash space, the most carry-efficient flatteners involve shorting April 2012 bonds to buy April 2023 or April 2027 bonds (around 15bp of carry over a 1-month horizon); based off the relationship in Exhibit 20, this suggests the flattener offers protection against around 215bp of 10Y peripheral spreads narrowing (over 1-month).
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 200 300 400 500
y = 0.0007x + 0.2384 R 2 = 0.5181
600
700
800
10y w eighted peripheral sw ap spread

* A weighted peripheral spread computed against Germany for Ireland, Portugal, Italy, Spain and Greece (weighted by the size of their outstanding bond markets)
Exhibit 21: AUS 3Y and 10Y swap spreads have drifted wider over the course of the year
Swap spread measures; bp
80 70 60 50 40 30 20 10 Jan-11
3y sw ap spread 10y sw ap spread
Swap spreads should be biased wider in 2012, 3s/10s spread curve likely to flatten
3Y and 10Y swap spreads have drifted wider over the course of 2011 (Exhibit 21), and we would expect the higher trading range to persist into 2012. We expect spreads to make new wides next year, and target 3Y spreads to trade a 50bp to 85bp range, and 10Y spreads to trade a 55bp to 90bp range in 2012. We expect spreads to retain strong directionality next year, with spreads widening on a rally and narrowing on a selloff. We still believe that a paid position in AUD 3Y swap spreads is a good way in which to position for still-wider peripheral spreads in Europe; Exhibit 22 illustrates the strong correlation between the two series. Aside from
May -11
Sep-11
Exhibit 22: AUS 3Y swap spreads exhibit strong directionality with 5Y peripheral spreads*
bp bp
1000 800 600
5-y ear w eighted peripheral spread to bund, lhs 3-y ear sw ap spread, rhs
65
50
35 400 200 Jan-11 Apr-11 Jul-11 Oct-11 20
* as defined in the footnote in Exhibit 20.

211
reflecting heightened banking system risk, wider swap spreads should also reflect strong demand for Australias AAA-rated sovereign paper in times of market stress (especially stress that is driven by sovereign debt concerns). And if peripheral spreads move wider and markets remain illiquid, then AUD corporate and Kangaroo issuance will likely be limited in 2012, implying the absence of issuance-related, receive-side flow in the swap market. Indeed, year-to-date issuance in Australia has been the lowest since 2008 for all spread products (Exhibit 23). Despite our view of wider swap spreads, there are a couple of factors which might work to cap the move wider in spreads next year. The first is further rate cuts from the RBA. Generally speaking, a lower cash rate has been associated with narrower 3Y swap spreads (less demand for fixed-rate mortgage products implies less demand from mortgage hedgers). But we expect this effect to be limited given the likelihood of an escalation of the European crisis. The second is the potential for AUD/JPY-related receiving in the long end (10Y sector) of the swap curve; generally speaking, a lower AUD/JPY exchange rate generates receiving of longer-dated swaps due to the hedging needs inherent in the issuance of callable power reverse dual currency notes2. Taking into account all the varied influences on swap spreads next year, we are biased towards a flatter 3s/10s swap spread curve.
Exhibit 23: In 2011, $A issuance volumes are likely to be the lowest since 2008
AUDbn
120 100 80 60 40 20 0
Financials
SSA
Corporate
2008
2009
2010
2011 YTD
Exhibit 24: The market has made a clear distinction between European and non-European supra-national paper this year
Asset swap levels (semi/quarterly basis), bp*
200 170 140 110 80 50 20 -10 Apr-2012
ASIA IADB RENTEN
EIB IFC
EUROF KFW
Dec-2014
Sep-2017
Jun-2020
Mar-2023
Macro-prudential regulation continues to be the major structural influence on relative value in AUD cash spread product
In terms of cash spread product, regulation continues to be the major structural influence upon the spread product market in Australia. In 2011, the major development was greater clarity on APS210, the regulatory standard which defines the type of assets that can be held in domestic bank liquidity books. In addition, the Australian bank regulator (APRA) and the RBA announced the establishment of a committed secured liquidity facility (CLF) in order to facilitate compliance by Australian
2
*Levels as of 17-November-2011
Authorised Deposit-Taking Institutions (ADIs) with the new macro-prudential framework outlined by Basel III. The purpose of this facility is to enable an ADI to establish a committed secured liquidity facility with the RBA, sufficient in size to cover any shortfall between the ADIs holdings of high-quality liquid assets and the liquidity coverage ratio requirement (LCR). APRA has determined that the only assets that qualify as Level 1 (high quality liquid) assets are cash, balances held with the Reserve Bank of Australia, and Commonwealth Government and semi-government securities. APRA has announced that (at this point in time) there are no assets that qualify as Level 2 assets. This implies that other AUD AAA-rated assets, such as supra-national bonds, covered bonds and RMBS, will only qualify towards the LCR via the CLF.
See S. Auld, Why does a higher AUD/JPY FX rate generate paying in AUD long end swap yields and AUD/USD cross currency basis spreads?: Explaining the impact of Callable Power Reverse Dual Currency Notes, 21 April 2011).
212
The ADI will pay a 15bp fee to access the CLF. Given current market spreads and RBA haircuts/repo charges, we estimate that spread product particularly European supra-national paper, RMBS and covered bonds look cheap to a bank liquidity portfolio. In terms of AUD AAA-rated paper eligible for the CLF, an optimally structured liquidity portfolio should overweight RMBS and covered bonds3. More generally, we remain cautious on spread products, given the risk of further spread widening in peripheral Europe. In the supra-national space, the market is very clearly making a geographic distinction, as investors and dealers reduce exposure to European names in favour of Washington names (Exhibit 24). As long as European issues remain largely unresolved, we expect this divergence to remain with non-European AAA-rated assets well sought after (and note that European supranationals EIB and KfW represent 35% of 2011 YTD Kangaroo issuance so far this year).
Exhibit 25: 5-year semi-government spread to swap vs. relative credit ranking NSWTC looks cheap relative to TCV
Asset swap vs. average z-score away from AAA mean; bp
24.0 16.0 8.0
SAFA (Apr-15)
NSWTC (Apr-16) 0.0 TCV (Nov -16) -8.0 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 WATC (Apr-15) 0.1 0.2 0.3 0.4
Av erage Z-score aw ay from AAA mean
Exhibit 26: We like owning the NSWTC April-15 against the TCV November-16 given the cheapness of the asset swap box
Spread in asset swap measures between the NSWTC April 15 and TCV November 16; bp
Semi-government paper we favour owning NSWTC bonds against swap and ACGBs
Given our views on swap spreads, we like owning cash spread product relative to swap rather than government bonds. When we look at AAA-rated semi-government paper using an assessment of relative credit metrics, both SAFA and NSWTC paper look cheap against swap (Exhibit 25). However, given recent commentary from the South Australian Premier, we see a significant risk that the state of South Australia could loose its AAA rating. As such, we prefer owning NSWTC paper against swap. In contrast, WATC and TCV paper looks rich/fair value against swap. This relative value assessment holds across both the 5Y and 10Y sectors, but we prefer to exploit this relative value in the 5Y sector, given our view on the swap spread curve (likely biased flatter) and like owning
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 -1.0 -2.0 May -11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11
the NSWTC April-15 against the TCV November164(Exhibit 26). As we noted above, we remain cautious on spread product given 1) the likelihood of an escalation in the European crisis and 2) the prospect of disappointment on State government budget outcomes (given a weaker domestic growth backdrop). But we note that the latter point works in favour of NSWTC paper, given the strong possibility of an announcement concerning the full privatisation of New South Wales electricity assets. One implication of any such policy announcement is that the
More detail on our relative value methodology can be found in S. Auld, Strategy Update: Ranking the AAA-rated State Government Issuers Where is the risk/reward? 4 October 2011.
4
For further detail on this issue, see S. Auld, The RBA's committed liquidity facility and initial margin changes: A low fee means that European supras, covered bonds and RMBS look cheap, 16 November 2011.
213
sale of electricity assets could mean significantly reduced funding requirements for NSWTC; currently, around 30% of the NSWTC net borrowing requirement is used to fund electricity assets. All other things equal, this should aid out-performance of NSWTC bonds, especially in the mid- to longer-dated maturities, and drive flattening of the NSWTC bond curve5.
Exhibit 27: Australian real yields have fallen significantly across the curve over the course of 2011
Real yields across Australian linkers; %
3.00 2.75 2.50 2.25 2.00 1.75 1.50 1.25 1.00 0.75 2015i 2020i 2025i 2030i 3-Jan-11 18-Nov -11
Australian inflation-linked bonds: the outperformer in 2011

Australian inflation-linked bonds have performed very well over the course of 2011, with real yields declining considerably over the year (Exhibit 27). The rally has been most pronounced at the front end of the curve, helping the real yield curve to bull-steepen. This is consistent with both downward revisions to the Australian growth profile over the course of the year and a global rally in inflation-linked bonds. Interestingly, Australian 10Y real yields are close to their all-time lows (since the data series began in the mid1980s) (Exhibit 28). Given that real yields are theoretically meant to encapsulate expectations about GDP growth over the long run, this performance is perhaps a little counter-intuitive (especially given the consensus expectation of strong, medium-term and longterm growth prospects in Australia thanks to its close ties with China). We think one potential explanation for the collapse in real yields is Australias disappointing productivity performance. In the long run, productivity growth is an important determinant of an economys GDP performance. Exhibit 29 illustrates that labour productivity growth and 10Y real yields are reasonably well correlated. The implication of this is that ongoing trend declines in productivity might imply a structurally lower trading range for longer-dated real yields going forward. But one important point to remember is that despite the impressive rally in Australian real yields this year, spreads to offshore markets continue to remain wide (Exhibit 30). As we have noted above, this real yield differential continues to be a driver of offshore demand for ACGB products in both the nominal and inflationlinked space.
For more detail, see S. Auld, Strategy Update: NSW Electricity Privatization - implications for the NSWTC bond curve, 2 November 2011.
214
5
Exhibit 28: The recent rally has pushed Australian 10Y real yields to historic lows
Monthly data for 10Y real yields; %
6.5 5.5 4.5 3.5 2.5 1.5 Jul-86 Mar-91 Nov -95 Jul-00 Mar-05 Nov -09
Exhibit 29: The trend decline in productivity may be driving structurally lower real yields in Australia
oya % %-pts
6.5 5.5 4.5 3.5 2.5 1.5 0.5 -0.5 -1.5 -2.5 -3.5 Mar-00
Labour productiv ity , lhs Annual change in 10y real y ield, rhs
1.20 0.80 0.40 0.00 -0.40 -0.80 -1.20
Aug-02
Jan-05
Jun-07
Nov -09
Indeed, the growth differential between Australia and the US correlates well with the 10Y real yield spread; based on J.P. Morgan growth forecasts for Australia and the US, the 10Y real yield spread looks too wide at present (Exhibit 31). Accordingly, we think the better AUS-US spread contraction trade is to own the ACGB Aug-2020i against the UST Jul-2021i. Australian breakevens have not performed well over the course of the year. This was largely contrary to expectations, where the consensus anticipated rising inflation in Australia. However, a softer growth outcome in 2011 and a lower inflation forecast track from the RBA has meant that market inflation expectations have struggled to rally. This has been mirrored in J.P. Morgans November 2011 Inflation Expectations Survey, where expectations of the chance of above-target inflation in Australia have declined consistently over the course of the year. The survey suggests that it was changing perceptions of domestic investors (as opposed to offshore investors) which, in turn, have driven the move towards more benign inflation expectations. J.P. Morgan economists now expect core inflation to remain within the target band until the end of 2012. Headline inflation should decline markedly in the near term (as base effects associated with the impact of the Queensland floods roll off), but will move higher as the impact of the carbon tax works its way through the economy. J.P. Morgan forecasts for inflation are outlined in Exhibit 32. We think one of the better opportunities in the AUD inflation-linked market is to be found in NSWTC inflation-linked bonds. NSWTC has been the predominant issuer in the semi-government inflation market, largely due to its requirement to fund stateowned utility assets. However, a recent discussion paper suggests the strong probability that the NSW Government will consider the full privatisation of its electricity assets (generation and transmission and distribution assets). In our view, the main consequence should the government decide to proceed with privatisation is that there is likely to be far less supply in NSWTC inflationlinked bonds as a large part of longer-dated inflation linked issuance is used to fund electricity assets. And given the recent under-performance of the 2035i maturity, we think there is good value in owning NSWTC 2035 linkers vs. bonds. For investors able to
Exhibit 30: Despite the rally in Australian real yields, 10Y cross-market real yield spreads remain wide
Monthly data for 10Y real yield spreads to TIPS, UK linkers and Canadian Real bonds; %-pts
3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0
Spread to US Spread to UK Spread to Canada
Mar-00 Nov -01 Jul-03 Mar-05 Nov -06 Jul-08 Mar-10
Exhibit 31: The 10Y real yield spread between Australia and the US looks too wide, given their GDP differential
10Y AUS-US real yield and real GDP growth spreads (J.P. Morgan forecast dashed line); %-pts %-pts
2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 Mar-99
10y AUS-US real y ield spread, lhs Real GDP spread, rhs, pushed fw d 4q
7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 -1.0 -2.0
May -02
Jul-05
Sep-08
Nov -11
Exhibit 32: We expect core inflation to remain within the target band in 2012. Headline inflation should decline markedly in the near term on base effects, but will move higher as the impact of the carbon tax works its way through the economy
Australian headline and core inflation, actual and J.P. Morgan forecasts; oya %
6.0 5.0 4.0 3.0 2.0 1.0 0.0 03Q1
Headline Core
05Q3
08Q1
10Q3
13Q1
215
trade asset swaps, we think that the 2035i also looks attractive on an asset-swap basis.
Exhibit 33: The benchmark ACGB curve, maturity and issue size
Outstanding amount by line in ACGBs; AUDbn
18.0
Australian bond supply update

The Australian Office of Financial Management (AOFM) expects that gross Treasury bond issuance will be around AUD51bn in the current 2011/12 financial year. In net terms, supply will be around $37bn. So far this year, the AOFM has issued $44.7bn of ACGBs, with net issuance at $33.5bn due to the maturity of the Jun-11 benchmark. Four new benchmark bonds have been introduced, including two new long-end bonds with maturity dates of 21 April 2023 and 21 April 2027. The latter bond extended the benchmark ACGB curve by a further four years. Exhibit 33 shows the ACGB benchmark curve and benchmark issue size as of 31 October 2011. There is currently $14.4bn of ACGB inflation-linked bonds on issue across four benchmark lines, and we expect a further $1.2bn to be issued by June 2012. We expect gross (and net) issuance of $2.0bn in 2012. As of October 2011, the AOFM had no plans to extend the inflation-linked benchmark curve, but in the May budget, it committed to keeping the size of the inflation-linked market at 10% to 15% of the total ACGB market. At present, the outlook for bond supply in 2012 is difficult to ascertain. One reason is that there is some uncertainty about whether or not the Federal Government will be able to meet its commitment to returning the budget to a small surplus in the 2012/13 fiscal year. J.P. Morgan economists are doubtful, and do not envisage the budget returning to surplus until 2014/15. Exhibit 34 highlights the difference between Federal Government and J.P. Morgan forecasts for the budget in the coming fiscal years. Over the forecast horizon, the cumulative difference is $29bn. We should get greater clarity on the Governments intentions when it releases the next MidYear Economic and Fiscal Update before the year ends. Depending on which set of forecasts we use, the outlook for ACGBs (nominal and inflation-linked) on issue through to mid-2014 varies from $215bn to $250bn (Exhibit 35). Note that this estimate does not include Tnotes (usually around $10bn-15bn on issue). Either way, it is likely that Australia will require the debt ceiling to be raised at some point in either 2012 or 2013. Section 5 of the Commonwealth Inscribed Stock Act (1911) states, The total face value of stock and
216
15.0 12.0 9.0 6.0 3.0 0.0

21-Apr-27 21-Apr-23 15-Jul-22 15-May-21 15-Apr-20 15-Mar-19 21-Jan-18 21-Jul-17 15-Feb-17 15-Jun-16 21-Oct-15 15-Apr-15 21-Oct-14 15-Jun-14 15-Dec-13 15-May-13 15-Nov-12 15-Apr-12
Exhibit 34: Federal and J.P. Morgan government budget forecasts J.P. Morgan economists are less optimistic on the governments ability to return to surplus in 2012/13
Budget balance; AUDbn
10 0 -10 -20 -30 -40 -50 -60 2009-10 (a) 2010-11 (a) 2011-12 (f) 2012-13 (f) 2013-14 (f) J.P. Morgan Gov ernment
Exhibit 35: Supply of ACGBs depends heavily on budget forecasts for coming fiscal years
Outstanding amount of ACGBs (nominal and inflation-linked), forecast under Federal Government and J.P. Morgan assumptions; AUDmn
260,000
Using Federal Gov ernment budget forecasts Using J.P. Morgan budget forecasts
195,000
130,000
65,000
0 Dec-70 Jun-78 Dec-85 Jun-93 Dec-00 Jun-08
securities on issue under this Act and the Loans Securities Act 1919 at any time must not exceed $250 billion. As of 11 November 2011, total Commonwealth Government Securities on issue amounted to $215.1bn. Looking to the end of the forecast horizon, there is still some debate about the future size of the ACGB market. This is because if the Federal Government realises its forecasts for a return to surplus by 2012/13, then this would imply a negative net bond supply at some point. In the 2011/12 Budget, the Government noted the recommendation from a panel of financial market participants and regulators that the ACGB market be kept at a constant size of nominal GDP (around 12% to 14%) (Exhibit 36). If the Government decides to accept this recommendation, then the AOFM will be required to manage a portfolio in order to invest surplus cash.
Exhibit 36: There is a proposal to keep the size of the ACGB market as a constant proportion of nominal GDP
Size of ACGB market as a proportion of nominal GDP; %
30.0 25.0 20.0 15.0 10.0 5.0 0.0 1983 1988 1993 1998 2003 2008 2013 Proposed target lev el: 12%-14% of GDP
217
Trading themes
Extend duration on dips The lesson in 2011 for investors in AUD fixed income was that events in Europe rather than domestic fundamentals had a major impact on government bond yields. Our European strategists remain bearish on Europe, and for this reason we are biased towards longduration strategies in AUD fixed income in 2012. While this is a somewhat difficult call given that yields are close to their lows for the year, we note that domestic investors are still running short of benchmark duration; US 10Y yields should decline further and strong demand from offshore investors should continue to support the market. Look to lighten duration in 2H12, given the risk that markets price in a more optimistic global outlook.
Spread product We are biased towards wider swap spreads next year and believe that paying 3Y swap spreads is a good insurance trade to hold into 2012. If risk aversion dominates markets, then we would expect the asset swap curve to flatten. In the cash spread product, we remain cautious. Domestic positioning still feels overweight spread product and as such, we think there will be better buying opportunities. We favour owning NSWTC paper to swap and AUD European supras against USD European supras. Inflation We like owning the ACGB Aug-2020i against the UST Jul-2021i; this spread looks too wide given the expected growth spread between Australia and the US. This trade should also perform given our view that the nominal AUS-US 10Y bond spread is too wide versus our medium term forecasts. Our preferred trade in the inflation space is to own the NSWTC Nov-2035i inflation-linked bond against the ACGB Sep-2030i bond. Any decision to privatise NSW electricity assets will be extremely positive for longer-dated NSW inflation-linked paper. AUS-US 10Y spread Our medium-term model suggests the 10Y spread should be biased narrower from current levels.
Curve The 3s/10s bond curve should steepen in 2012, given our economists expectation of further rate cuts from the RBA. But cash bond steepeners have negative carry and roll; instead, we prefer to express this view via forward starting swap steepeners. There is very attractive carry and roll in cash bond curve flatteners. Accordingly, we think these are an effective portfolio hedge for investors already running long-duration portfolios as they provide protection against the peripheral spread narrowing.
Exhibit 37: AUD interest rate forecasts

Percent Current RBA cash rate 90-day rate 3Y yield 10Y yield 3s/10s curve (bp) 3Y swap yield 10Y swap yield 3Y swap spread 10Y swap spread AUS-US 10Y spread (bp) 4.50 4.60 3.22 3.99 77.0 3.84 4.77 62.0 75.0 203.0 Dec-11 4.25 4.00 3.25 4.15 90.0 4.00 4.95 75.0 80.0 210.0 Mar-12 3.75 3.65 2.75 3.80 105.0 3.65 4.65 90.0 85.0 210.0 Jun-12 3.75 3.90 3.30 4.30 100.0 4.05 5.05 75.0 75.0 180.0 Sep-12 3.75 4.00 3.50 4.35 85.0 4.10 5.05 60.0 70.0 175.0 Dec-12 3.75 4.00 3.70 4.40 70.00 4.20 5.05 50.0 65.0 170.0
218
New Zealand
Despite the fact that both the NZD 2Y and 10Y swap yields are close to all-time lows, we think New Zealand yields will likely reach new lows again in 2012. Accordingly, we prefer to extend duration on any selloff This position should work well for 1H12, and we would look to lighten duration into the second half of the year due to the risk of a move towards higher yields globally We think the NZD OIS curve looks cheap relative to the AUD OIS curve, especially looking at Jun12 RBA and RBNZ OIS dates Generally, our view is that the 2s/10s swap curve looks too flat, especially if the RBNZ does not tighten policy next year. We like entering 2s/5s swap curve steepeners, 3 months forward. An alternative expression of this view is to pay the belly of the 2s/5s/10s swap butterfly, 3 months forward Given our view that 1) New Zealand swap yields are likely to reach new lows, and 2) that 10Y asset swap spreads already look too wide, we would look to underweight the long end of the NZGB curve against swap.
Exhibit 1: The New Zealand economy is yet to return to its pre-crisis peak
Real GDP, Index is December 2007 = 100
100 99 98 97 96 Dec-07 Sep-08 Jun-09 Mar-10 Dec-10
Exhibit 2: and if the Scandinavian experience is any indication, household deleveraging could take a long time
Household debt to GDP, %
103 95 87 79 71 63 55 47 New Zealand, lhs Scandinav ian av erage, rhs
55 50 45 40 35 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 Years from start of recession 4 5 6 7
The economic recovery that wasnt

New Zealand has faced a plethora of headwinds over the course of 2011. Indeed, we estimate that the economy grew by just 2.3% this year, a small improvement on the 1.7% growth recorded in 2010. Indeed, Exhibit 1 illustrates that in level terms, the New Zealand economy is yet to return to its pre-crisis levels. The New Zealand economy has been buffeted by ongoing household sector deleveraging, an elevated NZD, weaker-than-expected trading partner growth (especially Australia) and the impact of the Canterbury earthquake (and the subsequent delayed recovery efforts). Indeed, we expect household deleveraging to be a permanent feature of the New Zealand economic backdrop in the years ahead the experience of the Scandinavian countries in the early 1990s is quite illustrative for New Zealand too (Exhibit 2).
Exhibit 3: Low number of building consents suggest that postearthquake-related activity is yet to have a material impact on the economy
Number of building consents issued
3500 3000 2500 2000 1500 1000 500 Jan-90
Seasonally Adjusted Trend
Oct-93
Jul-97
Apr-01
Jan-05
Oct-08
219
The one bright spot for the New Zealand economy in 2011 was the Rugby World Cup (RWC), although sentiment data since the RWC suggests that any boost was very short-lived. Accordingly, we only expect a modest uptick in growth in 2012, with calendar growth likely to be 2.8% during the year. It has become evident that ongoing seismic shocks are continuing to delay any rebuilding efforts in the earthquake-affected regions and as such, the boost to broader economic activity is likely to be more of a 2013 story. Exhibit 3 shows that so far, there has been no notable uptick in building consents issued. The RBNZ responded to the severe earthquake in Canterbury by cutting the OCR by 50bp to 2.5%, back to the lows seen during the financial crisis (Exhibit 4). Initially, the RBNZ characterised this as an emergency setting and was keen to withdraw the extra stimulus as soon as possible. It has since become clear that this setting of monetary policy looks increasingly appropriate given the lack of momentum in the domestic economy and mounting risks offshore (Exhibit 5). We see a risk that RBNZ rhetoric will turn more dovish as we move into 2012. J.P. Morgan economists do not expect the RBNZ to begin normalising policy until September 2012. Clearly, risks are biased to unchanged policy for longer and possibly, a new low in the OCR. Indeed, the RBNZs decision in early November to delay the introduction of some macro-prudential regulatory changes (the core funding ratio for New Zealand banks) because of offshore instability may go some way to reflecting policy makers current mindset.
Exhibit 4: The RBNZ returned the policy rate back to lows seen during the financial crisis
OCR; %
9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 Mar-99 Dec-01 Sep-04 Jun-07 Mar-10
Exhibit 5: but if global conditions deteriorate further, then the case for further rate cuts will start to mount there is a strong correlation between moves in the OCR and J.P. Morgans Global PMI
%-pts Index
2 1 0 -1 -2 -3 -4 -5 -6 Sep-99 Feb-02 Jul-04 Dec-06 May -09 Oct-11 Annual difference in cash rate, lhs Global PMI, pushed fw d 6 months, rhs
62 57 52 47 42 37 32
Front-end pricing: A new low in the NZ policy rate in 2012?

Front-end pricing in New Zealand has been very volatile over the course of 2011, with the market pricing in anywhere from 20bp of rate cuts over a 12-month horizon to 110bp of rate hikes over the same period (Exhibit 6). Currently, the market is expecting the RBNZ to deliver 20bp of rate cuts by June 2012. In contrast, the local market economists consensus strongly favours the next move in rates as up in 2012. While we believe that offshore (and domestic) fundamentals would need to deteriorate markedly from here before the RBNZ will countenance the need for a lower policy rate, we are also cognizant that illiquid
220
Exhibit 6: Monetary policy expectations have been quite volatile in New Zealand this year
Tightening* priced in over a 12-month period; bp
120 100 80 60 40 20 0 -20 -40 Jan-11 Apr-11 Jul-11 Oct-11
* Negative number implies easing
global funding markets could also pressure the RBNZ to ease policy again. Exhibit 7 illustrates that there is a reasonable disconnect between the level of the OCR and bank lending rates. More expensive bank funding will only pressure these spreads wider. In a relative value sense, we think the front end of the NZ OIS curve looks cheap against the AUD OIS curve. For example, the June 2012 RBNZ meeting date OIS implies 20bp of rate cuts, while the June 2012 RBA meeting date OIS implies almost 160bp of rate cuts. If the RBA has cut 160bp by the middle of 2012, then the RBNZ will have likely cut by more than 20bp. Similarly, if markets normalise, then there is scope for a more rapid selloff in the AUD front end than in the NZD front end.
Exhibit 7: Lending rates have not fallen to the same extent as the OCR in New Zealand, meaning bank funding stress could also be a key driver of RBNZ policy moves in 2012
Various bank lending rates and OCR; %
11 10 9 8 7 6 5 4 3 2 Mar-99 Oct-01 May -04 Dec-06 Jul-09 OCR Variable business lending rate Floating mortgage rate
The outlook for outright yields new lows in 2012

New Zealand yields have largely moved lower over the course of the year, with a new historical yield low in the 2Y swap rate made this month. As Exhibit 8 illustrates, price action has been consistent with the bullish trend seen in 2Y rates over the past couple of years. This theme has also been replicated in the longer end of the swap curve, with the 10Y yield reaching a new historical yield low in November. The move lower in yields has been in sympathy with the broader macroeconomic environment in New Zealand and offshore (disappointing growth outcomes), and a growing expectation that central bank policy could turn more dovish. In addition, we think the recent move lower has been a function of positioning, with most dealers looking for higher yields into the year-end and not prepared for a rally. In the short term, some of the domestic sentiment indicators suggest the possibility of a bounce in the NZD 2Y swap yields (Exhibit 9). However, as we noted in our outlook for Australian yields, domestic fundamentals have not necessarily been a reliable guide to interest-rate market movements in 2011. And given the risks to the global backdrop, we would look to use any short-term sell-off in New Zealand yields as an opportunity to extend duration. In the 10Y part of the curve, a similar theme has prevailed. The NZD 10Y swap has made new, all-time lows recently but still looks to be trading largely in line with our estimate for fair value (Exhibit 10).
Exhibit 8: NZD 2-year swap yields have made new historical lows recently, consistent with the bull trend in place since late 2009
2Y swap yield, %
9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 Jan-07 May -08 Sep-09 Jan-11
Exhibit 9: Similar to Australia, domestic indicators suggest the risk of higher yields but this can easily be swamped by offshore developments
Index %-pts
45.0 35.0 25.0 15.0 5.0 -5.0 -15.0 -25.0 -35.0 -45.0 -55.0 Jun-04
Domestic economic activ ity ex pected, lhs, pushed fw d. 1q Annual change in 2y sw ap, rhs
3 2 1 0 -1 -2 -3 -4 -5
Oct-06
Feb-09
Jun-11
221
Our view on duration in New Zealand next year looks for lower yields, at least in 1H12. This is likely to be driven by: 1) The prospect of lower UST yields in early 2012; 2) The strong likelihood of noticeably more dovish commentary from the RBNZ; 3) The risk that the domestic data flow in New Zealand starts to disappoint; and 4) Ongoing flight to quality into New Zealand fixed income.
Exhibit 10: Fair value model for the NZD 10Y swap yield according to our forecasts, new yield lows are likely in the new year
10Y swap yield; %
8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 Mar-98 Oct-00 May -03 Dec-05 Jul-08 Feb-11 Actual Model
Curve spreads should be biased steeper in 2012

The NZD 2s/10s swap curve has traded a wide range over the course of 2011, reaching a peak of 214bp in late April and a trough of 127bp in late September. The curve is currently towards the lower end of this range, at around 152bp. Generally speaking, our view is that the curve looks too flat at current levels, especially given the risks surrounding the OCR in the coming year. Exhibit 11 suggests that the 2s/10s swap curve is already priced for a modest tightening cycle in 2012. We think the better curve trade in New Zealand is to position for steeper forward curves in 2012. At present, we think the forward swap curves look too flat (especially 2s/5s and 2s/10s) and like paying the 2s/5s swap curve 3-months forward. Indeed, our models suggest that the 2s/5s forward curve spread is pricing in a cash rate close to 3.5% in just six months time, an expectation well above our economists view and current front-end pricing (Exhibit 12). These trades are attractive because they roll positively up to spot, especially in 2s/10s (the 1Y forward drop to spot is around 12bp). Investors who prefer a more defensive curve steepening position should look to short the front end against a forward-curve steepener; either the front end is too rich or the forward curve spreads are too flat. The shape of the forward curves (in terms of what they imply for the path of the policy rate) also suggests that the 2s/5s/10s swap butterfly, 3-months forward, is starting to look too low. We like paying the belly of the fly on any move below flat (currently, the spread is market at 1bp).
Model is 10-year swap yield = 0.44 + 0.10*90-day rate + 0.65*US 10-year swap yield + 0.46*moving average of nominal GDP growth; R-squared = 0.75, standard error is 0.33ppts.
Exhibit 11: The 2s/10s curve looks flat, even if the RBNZ executed another 50bp of rate hikes next year
%-pts %
3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 Mar-99
2s10s curv e, lhs NZ OCR, rhs, inv erted
0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5
Jan-02
Nov -04
Sep-07
Jul-10
Exhibit 12: The cash rate implied by 2s/5s forward curve is too high relative to our forecast or market pricing. Therefore given the negative correlation* between the 2s/5s curve and cash rate, we favour 2s/5s swap curve steepeners 3M forward
Cash rate over various horizons; %
4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 Current
Cash rate implied by 2s5s forw ard curv e Cash rate, JPM forecast Cash rate implied by market pricing
3m
6m
1y
2y
* Model used for the curve-implied cash rate: 2s5s swap curve = 1.11 - 0.29*real cash
222
rate; R-squared is 0.86, standard error is 0.22ppts.
Swap spreads look close to their wides in the long end of the curve
New Zealand swap spreads have been quite volatile over the course of this year, with the 2021 bond trading at almost 40bp over the swap earlier in the year (Exhibit 13). Since then, NZGBs have generally outperformed swap, and swap spreads look to be pushing back to the wides seen earlier in the year. Generally speaking, NZGB swap spreads tend to exhibit strong directionality with 2Y swap yields. Indeed, our model suggests that movements in 2Y swap yields can explain 86% of the variation in 10Y swap spreads. This is because the swap market is generally seen as the more liquid market and will usually outperform bonds in a rally and underperform bonds in a selloff, as market participants look to express duration views through swaps rather than bonds. Our model of 10Y swap spreads (Exhibit 14) suggests that swap spreads look wide given the level of 2Y swap yields. Given our view that 1) New Zealand swap yields are likely to reach new lows, and 2) that 10Y swap spreads already look too wide, we would look to underweight the long end of the NZGB curve against swap on any outperformance of NZGBs in Q1 next year. Given the relative issue size ($6.6bn of Mar-19s and $10.4bn of May-21s), we would prefer to express this view via the 2019 maturity. Look to position for a narrower swap spread on the May-19 bond on any move to the +30bp level (currently 24bps).
Exhibit 13: NGZBs have generally outperformed swaps since April 2011, when they traded 40bp over, and look to be pushing back to their wides
Swap spread; bp
50 40 30 20 10 0 -10 -20 -30 -40 -50 Jan-11
NZGB Apr-13 NZGB May -21
Apr-11
Jul-11
Oct-11
Exhibit 14: 10Y swap spread model spreads look too wide given the level of yields
10Y swap spread actual and model*; bp
160 140 120 100 80 60 40 20 0 -20 -40 -60 Oct-03 Jul-05 Apr-07 Jan-09
Actual Model +/- 2 s.d.
New Zealand bond supply update

The New Zealand Debt Management Office (NZDMO) announced that gross bond supply in the 2011/12 fiscal year will be $13.5bn. In net terms, this amounts to a supply of $5.9bn, given the redemption of the Nov-2011 bond. Pre-funding in the prior fiscal year (2010/11) has meant that gross bond supply remains unchanged despite significant costs associated with the post-earthquake rebuilding. So far this year, the NZDMO has issued $6.45bn of benchmark bonds and is running ahead of schedule (on a pro-rata basis) for the fiscal year. The NZDMO lengthened the maturity of the NZGB curve this year with the issuance of a 5.5% April 2023 maturity. The NZGB benchmark curve is illustrated in Exhibit 15; in addition, benchmark line sizes for the April 2013 and April 2015 bonds were increased to a maximum of $12.0bn this year. In New Zealands
Oct-10
* Model 10Y swap spread = -87.55 + 23.90*2Y swap yield; R-squared = 0.86, standard error = 12.95bp
Exhibit 15: NZGB benchmark curve

Outstanding by line; NZDbn
12.0 10.0 8.0 6.0 4.0 2.0 0.0 15-Dec-17 15-Mar-19 15-Apr-13 15-Apr-15 15-May-21 15-Apr-23
.
223
inflation-linked bond space, the NZDMO has postponed plans for an inflation-linked issue (2023 maturity) until 1H12. There has been no new issuance into the sole inflation-linked February 2016 maturity so far this year. Total bonds on issue (nominals) in New Zealand now total $47.4bn (we exclude non-market bonds in our calculations those held by the RBNZ and EQC). We expect that gross (and net) issuance in 2012 to likely total $12.3bn, taking bonds on issue close to $60bn by the end of 2012 (Exhibit 16).
Exhibit 16: Issuance of NGZBs in 2012 will take outstanding amount to close to $60bn
Total outstanding of NGZBs; NZDbn
80
60
40
20
0 Jan-00 Apr-03 Jul-06 Oct-09 Jan-13
224
Trading themes
Extend duration on dips Despite the fact that both the NZD 2Y and 10Y swap yields are close to all-time lows, we think New Zealand yields will likely reach new lows again next year. Accordingly, we prefer to extend duration on any selloff. This position should work well for 1H12, and we would look to lighten duration into the second half of the year due to the risk of a move towards higher yields globally. We think the NZD OIS curve looks cheap relative to the AUD OIS curve, especially looking at the June 2012 RBA and RBNZ OIS dates. Curve Generally, our view is that the 2s/10s swap curve looks too flat, especially if the RBNZ does not tighten policy in 2012. Carry + roll on 2s10s steepeners is largely neutral. We think the forward 2s/5s swap curves look too flat relative to policy expectations. Implement 2s/5s steepeners, 3-months forward (3-month carry + roll of 2.5bps). A more defensive expression of this trade would be to hold steepeners against paid 3-month OIS positions. We like paying the belly of the 2s/5s/10s swap curve, 3-months forward. This trade has positive carry + roll of 3.0bp over a 3-month period.
Spreads Given our view that 1) New Zealand swap yields are likely to reach new lows, and 2) that 10-year asset swap spreads already look too wide, we would look to underweight the long end of the NZGB curve against swap on any out-performance of NZGBs in Q1 next year. Given the relative issue size ($6.6bn of Mar-19s and $10.4bn of May-21s), we would prefer to express this view via the 2019 maturity. Look to position for a narrower asset swap spread on the May-19 bond on any move to the +35bp level.
Exhibit 17: NZD interest rate forecasts

% Current RBNZ OCR 90-day rate 3-year yield 10-year yield 10-year swap spread (bp) 2-year swap yield 10-year swap yield 2s10s swap curve (bp) NZ-US 10-year spread (bp) 2.50 2.66 2.82 3.91 27.5 2.71 4.23 152.0 190.0 Dec-11 2.50 2.65 2.85 4.05 20.0 2.65 4.25 160.0 200.00 Mar-12 2.50 2.65 2.50 3.80 25.0 2.30 4.05 175.0 210.0 Jun-12 2.50 2.75 3.25 4.40 15.0 2.95 4.55 160.0 190.0 Sep-12 2.75 3.00 3.40 4.40 10.0 3.00 4.50 150.0 180.0 Dec-12 3.00 3.25 3.50 4.40 10.0 3.10 4.50 140.0 170.0
225
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Aditya ChordiaAC (44-20) 7777-9841 aditya.x.chordia@jpmorgan.com J.P. Morgan Securities Ltd
Interest rate forecasts

Euro area Refi rate Euribor 18-Nov-11 Mar-12 1.25 0.75 3M 1.47 1.40 2Y 0.46 0.35 1.10 0.75 5Y 1.97 1.55 10Y 2.61 2.15 30Y 2s/10s 151 120 10s/30s 64 60 2s/30s 215 180 2Y 110 130 5Y 96 115 10Y 67 80 19 25 30Y Jun-12 0.50 1.45 0.30 0.70 1.25 1.95 95 70 165 145 130 90 30 Jun-12 0.50 1.10 0.50 1.00 1.50 2.25 100 75 175 120 95 55 5 Sep-12 0.50 1.20 0.40 0.85 1.50 2.25 110 75 185 135 120 85 25 Sep-12 0.50 1.05 0.55 1.05 1.75 2.40 120 65 185 110 85 45 0 Dec-12 0.50 1.00 0.50 1.05 1.75 2.50 125 75 200 130 115 80 25 Dec-12 0.50 1.00 0.65 1.15 1.95 2.60 130 65 195 100 65 25 -10 United States 18-Nov-11 Mar-12 Jun-12 Sep-12 Dec-12 Fed funds 0 - 0.25 0 - 0.25 0 - 0.25 0 - 0.25 0 - 0.25 Libor 3M 0.49 0.60 0.50 0.40 0.35 2Y 0.28 0.17 0.30 0.30 0.30 5Y 0.92 0.75 1.25 1.25 1.25 10Y 2.01 1.70 2.50 2.50 2.50 Govt curve 30Y 3.00 2.70 3.60 3.60 3.60 2s/10s 173 153 220 220 220 10s/30s 99 100 110 110 110 2s/30s 271 253 330 330 330 2Y 48 53 38 38 38 5Y 40 45 34 34 34 Swap spreads 10Y 16 27 18 18 18 30Y -30 -23 -23 -23 -23 Australia Cash rate Govt curve
Govt curve
Swap spreads
United Kingdom 18-Nov-11 Mar-12 Base rate 0.50 0.50 Libor 3M 1.02 1.05 0.48 0.50 2Y 1.13 1.00 5Y 10Y 2.26 1.80 3.19 2.70 30Y Govt curve 2s/10s 178 130 10s/30s 93 90 2s/30s 271 220 2Y 96 110 5Y 71 85 Swap spreads 10Y 31 45 -9 0 30Y Japan O/N call rate 2Y 5Y 10Y 20Y 30Y 2s/10s 10s/30s 2s/30s
10Y
4.50 3.99
3.75 3.80
3.75 4.30
3.75 4.35
3.75 4.40
New Zealand Cash rate Govt curve 10Y
2.50 3.91
2.50 3.80
2.50 4.40
2.75 4.40
3.00 4.40
Govt curve
0.05 0.12 0.31 0.94 1.70 1.93 82 99 181
0.05 0.15 0.30 0.90 1.75 1.95 75 105 180
0.05 0.15 0.35 0.95 1.80 2.00 80 105 185
0.05 0.15 0.40 1.10 1.90 2.10 95 100 195
0.05 0.15 0.45 1.15 2.00 2.15 100 100 200
226
Recent curve movements

Govt. curve 2Y 5Y 10Y 30Y 2s/5s 2s/10s 10s/30s Swap curve 2Y 5Y 10Y 30Y 2s/5s 2s/10s 10s/30s EUR 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 0.46 0.56 1.61 1.78 0.31 1.92 1.10 1.17 2.28 2.66 0.86 2.82 1.97 1.89 3.03 3.35 1.69 3.50 2.61 2.64 3.76 3.79 2.46 3.97 63 61 67 88 44 109 151 133 142 157 128 207 64 75 73 44 39 99 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 1.60 1.50 2.18 2.35 1.37 2.47 2.09 1.98 2.82 3.08 1.84 3.23 2.68 2.55 3.44 3.63 2.40 3.79 2.82 2.74 3.85 3.86 2.60 4.02 49 48 64 73 41 88 108 105 126 128 97 165 14 19 42 23 14 42 GBP 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 0.48 0.56 0.83 1.34 0.47 1.55 1.13 1.07 1.84 2.25 0.95 2.55 2.26 2.26 3.38 3.68 2.10 3.88 3.19 3.55 4.28 4.35 3.12 4.58 64 50 101 91 41 135 178 169 255 234 152 276 93 129 90 67 62 144 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 1.54 1.28 1.45 1.86 1.08 5.05 1.86 1.78 2.51 2.97 1.64 3.22 2.60 2.60 3.55 3.80 2.48 4.05 3.11 3.34 4.08 4.13 3.04 4.29 32 50 106 111 -257 125 105 132 211 194 -162 214 51 74 53 32 20 85 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 96 67 48 48 42 96 71 50 40 36 29 71 31 21 6 6 2 37 -9 -21 -20 -22 -30 -6 -25 -17 -8 -12 -26 -7 -65 -46 -43 -41 -65 -27 -40 -42 -26 -29 -46 -22
Swap spreads 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 2Y 110 92 55 53 47 112 5Y 96 81 49 39 37 100 10Y 67 65 41 26 21 78 30Y 19 9 9 4 0 22 2s/5s -14 -11 -6 -14 -20 2 2s/10s -43 -27 -14 -27 -44 -3 10s/30s -48 -56 -32 -21 -67 -18
Govt. curve 2Y 5Y 10Y 30Y 2s/5s 2s/10s 10s/30s Swap curve 2Y 5Y 10Y 30Y 2s/5s 2s/10s 10s/30s
USD 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 0.28 0.25 0.46 0.79 0.15 0.85 0.92 0.97 1.75 2.22 0.77 2.40 2.01 1.93 3.16 3.45 1.71 3.72 3.00 2.92 4.38 4.51 2.76 4.76 64 71 130 144 57 155 173 167 270 267 151 289 99 99 123 106 95 148 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 0.78 0.58 0.69 0.96 0.42 1.05 1.34 1.26 2.02 2.42 1.07 2.62 2.18 2.11 3.26 3.56 1.90 3.83 2.70 2.70 4.07 4.29 2.56 4.54 56 69 133 147 53 158 140 154 257 260 135 279 52 58 81 73 51 93
JPY 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 0.12 0.14 0.17 0.21 0.12 0.25 0.31 0.37 0.43 0.49 0.29 0.62 0.94 1.03 1.13 1.25 0.94 1.35 1.93 1.91 2.01 2.18 1.85 2.26 19 23 26 28 17 38 82 89 96 104 82 114 99 89 88 93 86 101 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 0.38 0.35 0.38 0.39 0.31 0.46 0.46 0.49 0.56 0.63 0.42 0.78 0.95 1.03 1.16 1.31 0.95 1.40 1.72 1.80 2.01 2.17 1.72 2.25 7 14 17 24 7 32 57 68 78 91 57 98 77 77 85 86 75 88 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 26 21 22 19 15 26 14 11 13 13 6 17 -1 0 2 5 -3 9 -22 -13 -4 -1 -22 2 -12 -10 -9 -5 -13 -5 -27 -21 -19 -13 -27 -12 -21 -13 -6 -6 -21 -4
Swap spreads 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 2Y 48 32 22 16 13 52 5Y 40 29 25 19 16 44 10Y 16 16 8 8 2 22 30Y -30 -27 -31 -23 -38 -20 2s/5s -9 -3 3 3 -7 9 2s/10s -32 -15 -14 -8 5 33 10s/30s -46 -43 -39 -31 27 54
227
Recent sovereign cash spread movements

Spread to Germany; bp
18-Nov-11 1W ago Austria 111 94 Belgium 344 273 Finland 39 23 France 119 105 Netherlands 29 19 Greece 9131 8414 Ireland 773 765 Italy 572 543 Portugal 1507 1556 Spain 484 421 Wtd. peri. spread* 1418 1321 2W ago 1M ago 3M ago 45 47 30 219 212 132 13 17 12 63 69 33 15 14 11 8379 6499 3321 801 732 774 522 408 283 1900 1603 1099 370 322 264 1315 1042 632 2Y 1Y min 6 32 -4 4 1 1092 361 83 299 102 229 5Y 1Y min 30 78 6 21 5 937 495 102 292 155 241 10Y 1Y min 31 72 24 30 19 747 501 125 306 170 229 30Y 1Y min 28 75 32 6 458 153 254 185 205 1Y max 130 357 51 140 48 9131 2164 684 1901 492 1418 1Y avg 27 104 10 28 9 3209 808 225 962 225 574 1Y SD 1Y z-score 19 4.4 62 3.9 7 4.0 24 3.7 6 3.3 2330 2.5 314 -0.1 136 2.6 500 1.1 83 3.1 336 2.5
18-Nov-11 1W ago 2W ago 1M ago 3M ago Austria 177 157 110 100 54 Belgium 330 294 283 281 196 Finland 82 63 52 52 33 France 151 135 106 104 47 Netherlands 66 50 41 51 35 Greece 4003 3771 3573 2939 1734 Ireland 706 707 704 614 780 Italy 533 548 526 435 304 Portugal 1322 1351 1459 1375 1085 Spain 453 409 388 344 302 Wtd. peri. spread* 892 871 840 709 503
1Y max 202 351 91 187 83 4059 1509 675 1527 474 951
1Y avg 55 158 26 49 25 1714 759 250 905 266 451
1Y SD 1Y z-score 30 4.1 63 2.7 19 3.0 30 3.4 17 2.5 763 3.0 211 -0.3 132 2.1 392 1.1 72 2.6 179 2.5
18-Nov-11 1W ago 2W ago 1M ago 3M ago Austria 147 149 113 105 60 Belgium 282 256 253 244 176 Finland 67 55 51 51 43 France 147 148 120 110 64 Netherlands 56 47 43 45 38 Greece 2425 2357 2407 2266 1352 Ireland 625 611 631 637 761 Italy 494 492 471 399 285 Portugal 910 907 986 961 930 Spain 469 395 374 338 291 Wtd. peri. spread* 699 673 666 603 444
1Y max 189 309 75 188 65 2471 1142 575 1161 486 739
1Y avg 58 141 37 58 33 1342 679 244 722 267 400
1Y SD 1Y z-score 28 3.1 56 2.5 11 2.8 30 3.0 10 2.3 491 2.2 128 -0.4 113 2.2 269 0.7 65 3.1 135 2.2
18-Nov-11 1W ago 2W ago 1M ago 3M ago Austria 121 124 95 82 41 Belgium 238 238 219 210 147 Finland France 152 158 123 109 57 Netherlands 11 6 7 9 11 Greece 1523 1487 1448 1303 761 Ireland Italy 450 444 426 392 287 Portugal 648 652 639 628 544 Spain 431 389 365 343 270 Wtd. peri. spread* 563 545 525 484 345
1Y max 158 280 199 15 1544 515 682 448 595
1Y avg 45 129 58 10 781 254 443 267 320
1Y SD 1Y z-score 22 3.5 43 2.5 30 3.2 2 0.6 305 2.4 92 2.1 146 1.4 53 3.1 104 2.3
*Weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy and Spain (weighted by the size of their outstanding bond market). 30Y does not contain Ireland.
228
Recent sovereign CDS spread movements

CDS spread*; bp
18-Nov-11 1W ago 2W ago 1M ago 3M ago US 29 29 29 50 50 UK 51 50 49 52 36 Germany 46 45 43 48 30 Austria 149 127 109 112 48 Belgium 306 279 260 275 202 Finland 34 32 30 38 30 France 158 138 121 131 98 Netherlands 55 49 48 41 26 Greece*** 55 52 54 57 37 Ireland 911 930 917 968 1051 Italy 573 551 463 428 336 Portugal 1507 1500 1471 1527 1196 Spain 451 405 360 362 351 Wtd. peri. spread** 616 591 521 506 429 2Y 1Y min 16 21 10 20 62 12 29 10 15 531 71 392 142 148 5Y 1Y min 37 47 36 50 115 24 63 28 25 510 125 390 197 189 1Y max 50 58 61 159 320 45 167 60 59 1606 627 1757 464 657 1Y avg 33 34 24 54 156 22 68 26 31 858 229 934 274 316 1Y SD 1Y z-score 13 -0.3 9 1.8 12 1.8 33 2.9 68 2.2 9 1.3 38 2.4 12 2.5 14 1.7 223 0.2 150 2.3 440 1.3 82 2.2 145 2.1
18-Nov-11 1W ago 2W ago 1M ago 3M ago US 40 40 40 65 65 UK 90 89 87 92 79 Germany 95 92 87 95 82 Austria 207 185 159 163 120 Belgium 332 305 286 301 235 Finland 68 63 60 76 60 France 221 202 178 191 148 Netherlands 113 100 98 98 67 Greece*** 60 58 59 61 45 Ireland 715 730 720 760 785 Italy 533 530 490 453 360 Portugal 1060 1054 1034 1119 869 Spain 465 423 383 385 370 Wtd. peri. spread** 557 545 507 492 415
1Y max 65 104 120 222 346 90 234 122 63 1199 583 1235 479 595
1Y avg 52 70 61 98 201 45 116 60 41 694 269 750 310 331
1Y SD 1Y z-score 9 -1.3 13 1.6 21 1.7 39 2.8 58 2.3 18 1.3 45 2.3 25 2.2 12 1.7 121 0.2 130 2.0 258 1.2 68 2.3 115 2.0
* 25bp running coupon used for Finland, France, Germany, Netherlands and US. 100bp running coupon used for UK, Austria, Belgium, Greece, Ireland, Italy, Portugal and Spain. Spreads for all the countries except US are in $ CDS and for US it is in CDS. ** Weighted peripheral spread computed as CDS spread of Ireland, Portugal, Italy and Spain, weighted by the size of their outstanding bond market. *** For 2Y and 5Y Greece CDS, we show upfront premium instead of flat CDS spread for 100bp running coupon.
229
Euro area marketable bond* and bank debt** redemptions

bn
Austria Dec-11 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 Total Sov. Banks 0 2 2 0 1 0 0 0 10 0 0 0 1 0 15 3 2 2 2 2 1 1 0 2 1 1 2 18 Belgium Sov. Banks 0 0 0 0 4 0 0 0 0 1 12 0 0 8 25 5 2 1 0 2 1 0 1 0 3 1 0 17 Finland Sov. Banks 0 0 0 0 0 0 0 0 0 1 6 0 0 0 7 Italy Sov. Banks 0 0 36 27 28 1 2 17 12 11 20 13 30 198 2 4 5 6 2 6 3 2 2 4 6 4 10 55 0 0 0 1 0 1 0 1 0 0 0 0 4 France Sov. Banks 0 10 15 0 0 18 0 0 28 0 12 19 0 5 98 5 7 7 6 2 5 13 0 5 9 1 3 74 Germany Sov. Banks 18 3 25 0 19 16 0 19 27 0 21 16 0 17 178 28 8 13 3 2 9 10 5 11 9 5 3 108 Netherlands Sov. Banks 0 1 14 0 0 0 0 0 15 0 0 0 0 0 30 8 12 3 10 4 2 2 3 2 1 1 1 52 Core Euro-area*** Sov. 18 56 1 24 34 0 19 81 2 51 35 1 30 353 Banks 16 50 31 26 22 12 20 25 10 21 23 9 9 272
Greece Sov. Banks Dec-11 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 Total 3 0 0 14 0 9 0 0 8 0 0 0 2 36 0 2 1 2 0 0 1 1 1 2 1 1 0 11
Ireland Sov. Banks 0 0 0 6 0 0 0 0 0 0 0 0 0 6 0 1 2 2 3 0 2 0 0 0 1 3 0 14
Portugal Sov. Banks 0 0 0 0 0 0 10 1 0 0 0 0 1 13 1 4 1 1 0 1 2 2 0 0 0 0 0 14 0 0 1 1
Spain Sov. Banks 6 6 21 14 15 10 25 8 3 7 5 10 5 133
Peri. Euro-area**** Sov. 3 0 38 48 40 9 12 31 20 13 40 13 34 301 Banks 10 18 29 25 21 17 33 12 5 12 12 18 15 227
12 0 0 13 0 2 20 0 0 49
* Marketable bonds include: conventionals, linkers, floaters zero coupons and international bonds. ** Maturities in all currencies and jurisdictions and include secured, unsecured and securitised issuance, including MTNs but excluding short-term (maturity of less than one year) and self-funded deals (deals where there is only one bookrunner and it is also the issuer). The data also include any government guaranteed issuance by the banks but no direct issuance by government or government sponsored institutions. *** Austria, Belgium, Finland, France, Germany and Netherlands **** Greece, Ireland, Italy, Portugal, Spain Source: J. P. Morgan, Dealogic
230
Euro area sovereign ratings / SMP purchases / Election calendar

Sovereign ratings
S&P Austria Belgium Cy prus Finland France Germany Greece Ireland Italy Netherlands Portugal Slov akia Slov enia Spain AAA AA+ BBB AAA AAA AAA CC BBB+ A AAA BBBA+ AAAANEG NEG POS NEG NEG NEG NEG* Moody's Aaa Aa1 Baa3 Aaa Aaa Aaa Ca Ba1 A2 Aaa Ba2 A1 Aa3 A1 NEG* NEG NEG DEV NEG NEG NEG* NEG* AAA AA+ BBB AAA AAA AAA CCC BBB+ A+ AAA BB+ A+ AAAANEG NEG NEG NEG NEG
09-Nov 02-Nov 26-Oct 19-Oct 12-Oct 05-Oct 28-Sep 21-Sep 14-Sep 07-Sep 31-Aug 24-Aug 17-Aug 10-Aug 03-Aug 15-Nov 08-Nov 01-Nov 25-Oct 18-Oct 11-Oct 04-Oct 27-Sep 20-Sep 13-Sep 06-Sep 30-Aug 23-Aug 16-Aug 09-Aug 14-Nov 07-Nov 31-Oct 24-Oct 17-Oct 10-Oct 03-Oct 26-Sep 19-Sep 12-Sep 05-Sep 29-Aug 22-Aug 15-Aug 08-Aug 18-Nov 11-Nov 04-Nov 28-Oct 21-Oct 14-Oct 07-Oct 30-Sep 23-Sep 16-Sep 09-Sep 02-Sep 26-Aug 19-Aug 12-Aug 8.0 4.5 9.5 4.0 4.5 2.2 2.3 3.8 4.0 9.8 14.0 13.3 6.7 14.3 22.0 78.3
SMP purchases
Fitch NEG NEG
Trade date Settlement date Weekly (bn) Amount matured (bn) 0.1 0.6 0.0 0.0 0.2 0.0 0.0 0.0 0.1 0.0 0.0 0.0 1.3 0.0 0.0 4.3 Amount offered for sterilization (bn) 194.5 187.0 183.0 173.5 169.5 165.0 163.0 160.5 156.5 152.5 143.0 129.0 115.5 110.5 96.0 74.0 First Last First Last
* represents under watch; grey highlight: below IG; Outlook:NEG - negative outlook, POS positive outlook, DEV - developing outlook and blank represents stable outlook Notes: 1 For a countrys exclusion, Barclays Capital requires 2 out of 3 credit ratings (S&P, Moodys and Fitch) to be below IG. Citigroup requires both S&P and Moodys rating below IG. J.P. Morgan EMU Investment Grade index requires Moodys, S&P, and Fitch each to rate a country above IG to maintain inclusion. 2 Markit iBoxx uses an average rating methodology (S&P, Moodys and Fitch) for a country's exclusion. Source: Bloomberg
Cumulativ e amount till 2 Aug 2011
Note: Every Tuesday ECB sterilizes SMP purchases during the period between the Wednesday two weeks back (first trade date) and the Tuesday of the previous week (last trade date). This is equivalent to SMP purchases which settle on Monday (first settlement date) to Friday of the previous week (last settlement date). The ECB started buying Italian and Spanish bonds on Monday, 8 August 2011. Source: ECB
Election calendar
Period 2012 2013 Date 19 22 2014 2015 2016 Month February April Election in Greece France Austria Germany Italy Belgium Netherlands Finland Portugal Spain Ireland
231
Euro area fact sheet

Marketable debt outstanding (bn) Avg debt maturity Conv. bond issuance Bonds Tbills % Tbills (years, JPM) Gross (bn, JPM) Net (bn, JPM) Oct-11 Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain GDP wtd. avg. US* UK* 8246 944 1480 57 167 264 62 1098 1021 247 85 1356 261 104 469 Oct-11 1 32 7 163 63 12 0 130 40 10 81 Oct-11 1 11 10 13 6 5 0 9 13 9 15 9 15 6 Oct-11 7.6 6.7 5.2 7.3 6.2 6.7 6.3 6.9 6.2 6.1 6.7 6.7 5.3 14.5 2150 186 936 142 Tot. 2012 20 40 13 180 175 0 0 145 48 0 90 Tot. 2012 10 16 7 95 18 -30 -6 35 18 -10 49 Non-dom investors (%) 79 59 84 66 81 67 83 45 68 80 38 65 46 30 Bank assets % of bank assets in govt % loans/ % of bank assets /GDP ratio securities (incl. non-dom.) deposits ratio funded at the CB Sep-11 3.3 3.2 3.0 4.2 3.2 2.3 4.1 2.5 4.0 3.4 3.3 3.3 5.4 Sep-11 2.9 8.8 1.7 3.9 3.7 9.4 3.3 6.5 4.0 4.8 5.0 4.7 3.7 1.8 Sep-11 109 59 141 117 84 134 156 126 125 119 107 108 81 91 Latest 0.7 1.6 0.0 1.0 0.4 15.9 21.8 2.8 0.7 8.0 2.4 2.1 2.0
GDP (bn) 2012 Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Portugal Spain GDP wtd. avg. US UK 310 382 198 2,028 2,623 212 159 1,617 623 169 1,094 9,415 15,495* 1,566*
GDP growth Inflation** Budget balance^ Prim. Balance (oya, %) 2012 0.9 0.9 1.4 0.6 0.8 -2.8 1.1 0.1 0.5 -3.0 0.7 0.5 1.5 0.6 (oya, %) 2012 2.2 2.0 2.6 1.5 2.0 0.8 0.7 2.0 1.9 3.0 1.1 1.8 1.9 2.9 (% of GDP) 2012 -3.1 -4.6 -0.7 -5.3 -1.0 -7.0 -8.6 -2.3 -3.1 -4.5 -5.9 -3.4 -8.5 -7.8 (% of GDP) 2012 -0.3 -1.3 0.5 -2.5 1.3 1.0 -4.3 3.1 -1.2 0.8 -3.5 -0.2 -5.4 -4.6
Gross debt Curr. acc. bal. (% of GDP) 2012 73 99 52 89 81 198 118 121 65 111 74 91 105*** 89 (% of GDP) 2012 2.8 2.1 0.0 -3.3 4.4 -7.9 1.5 -3.0 7.0 -5.0 -3.0 0.0 -3.1 -0.9
GDPpc (EU=100) 2012 124 116 123 104 108 63 119 90 126 54 80 101 -
Unempl. rate (%) Latest 3.9 6.7 7.8 9.9 5.8 17.6 14.2 8.3 4.8 12.5 22.6 10 9.0 8.3
* Local currency ** HICP; National index if not available *** IMF Estimate ^ Net lending (+) or net borrowing (-) Source: EC European Economic Forecast, Autumn 2011, IMF, Eurostat and ILO
232
European Rates Strategy Global Fixed Income Markets 2012 Outlook November 24, 2011 Pavan WadhwaAC (44-20) 7777-3370 pavan.wadhwa@jpmorgan.com J.P. Morgan Securities Ltd
Analyst Certification: The research analyst(s) denoted by an AC on the cover of this report certifies (or, where multiple research analysts are primarily responsible for this report, the research analyst denoted by an AC on the cover or within the document individually certifies, with respect to each security or issuer that the research analyst covers in this research) that: (1) all of the views expressed in this report accurately reflect his or her personal views about any and all of the subject securities or issuers; and (2) no part of any of the research analyst's compensation was, is, or will be directly or indirectly related to the specific recommendations or views expressed by the research analyst(s) in this report.
Other Disclosures
J.P. Morgan ("JPM") is the global brand name for J.P. Morgan Securities LLC ("JPMS") and its affiliates worldwide. J.P. Morgan Cazenove is a marketing name for the U.K. investment banking businesses and EMEA cash equities and equity research businesses of JPMorgan Chase & Co. and its subsidiaries. Options related research: If the information contained herein regards options related research, such information is available only to persons who have received the proper option risk disclosure documents. For a copy of the Option Clearing Corporation's Characteristics and Risks of Standardized Options, please contact your J.P. Morgan Representative or visit the OCC's website at http://www.optionsclearing.com/publications/risks/riskstoc.pdf Legal Entities Disclosures U.S.: JPMS is a member of NYSE, FINRA, SIPC and the NFA. JPMorgan Chase Bank, N.A. is a member of FDIC and is authorized and regulated in the UK by the Financial Services Authority. U.K.: J.P. Morgan Securities Ltd. (JPMSL) is a member of the London Stock Exchange and is authorized and regulated by the Financial Services Authority. Registered in England & Wales No. 2711006. Registered Office 125 London Wall, London EC2Y 5AJ. South Africa: J.P. Morgan Equities Limited is a member of the Johannesburg Securities Exchange and is regulated by the FSB. Hong Kong: J.P. Morgan Securities (Asia Pacific) Limited (CE number AAJ321) is regulated by the Hong Kong Monetary Authority and the Securities and Futures Commission in Hong Kong. Korea: J.P. Morgan Securities (Far East) Ltd, Seoul Branch, is regulated by the Korea Financial Supervisory Service. Australia: J.P. Morgan Australia Limited (ABN 52 002 888 011/AFS Licence No: 238188) is regulated by ASIC and J.P. Morgan Securities Australia Limited (ABN 61 003 245 234/AFS Licence No: 238066) is a Market Participant with the ASX and regulated by ASIC. Taiwan: J.P.Morgan Securities (Taiwan) Limited is a participant of the Taiwan Stock Exchange (company-type) and regulated by the Taiwan Securities and Futures Bureau. India: J.P. Morgan India Private Limited, having its registered office at J.P. Morgan Tower, Off. C.S.T. Road, Kalina, Santacruz East, Mumbai - 400098, is a member of the National Stock Exchange of India Limited (SEBI Registration Number - INB 230675231/INF 230675231/INE 230675231) and Bombay Stock Exchange Limited (SEBI Registration Number INB 010675237/INF 010675237) and is regulated by Securities and Exchange Board of India. Thailand: JPMorgan Securities (Thailand) Limited is a member of the Stock Exchange of Thailand and is regulated by the Ministry of Finance and the Securities and Exchange Commission. Indonesia: PT J.P. Morgan Securities Indonesia is a member of the Indonesia Stock Exchange and is regulated by the BAPEPAM LK. Philippines: J.P. Morgan Securities Philippines Inc. is a member of the Philippine Stock Exchange and is regulated by the Securities and Exchange Commission. Brazil: Banco J.P. Morgan S.A. is regulated by the Comissao de Valores Mobiliarios (CVM) and by the Central Bank of Brazil. Mexico: J.P. Morgan Casa de Bolsa, S.A. de C.V., J.P. Morgan Grupo Financiero is a member of the Mexican Stock Exchange and authorized to act as a broker dealer by the National Banking and Securities Exchange Commission. Singapore: This material is issued and distributed in Singapore by J.P. Morgan Securities Singapore Private Limited (JPMSS) [MICA (P) 025/01/2011 and Co. Reg. No.: 199405335R] which is a member of the Singapore Exchange Securities Trading Limited and is regulated by the Monetary Authority of Singapore (MAS) and/or JPMorgan Chase Bank, N.A., Singapore branch (JPMCB Singapore) which is regulated by the MAS. Malaysia: This material is issued and distributed in Malaysia by JPMorgan Securities (Malaysia) Sdn Bhd (18146-X) which is a Participating Organization of Bursa Malaysia Berhad and a holder of Capital Markets Services License issued by the Securities Commission in Malaysia. Pakistan: J. P. Morgan Pakistan Broking (Pvt.) Ltd is a member of the Karachi Stock Exchange and regulated by the Securities and Exchange Commission of Pakistan. Saudi Arabia: J.P. Morgan Saudi Arabia Ltd. is authorized by the Capital Market Authority of the Kingdom of Saudi Arabia (CMA) to carry out dealing as an agent, arranging, advising and custody, with respect to securities business under licence number 35-07079 and its registered address is at 8th Floor, AlFaisaliyah Tower, King Fahad Road, P.O. Box 51907, Riyadh 11553, Kingdom of Saudi Arabia. Dubai: JPMorgan Chase Bank, N.A., Dubai Branch is regulated by the Dubai Financial Services Authority (DFSA) and its registered address is Dubai International Financial Centre Building 3, Level 7, PO Box 506551, Dubai, UAE. Country and Region Specific Disclosures U.K. and European Economic Area (EEA): Unless specified to the contrary, issued and approved for distribution in the U.K. and the EEA by JPMSL. Investment research issued by JPMSL has been prepared in accordance with JPMSL's policies for managing conflicts of interest arising as a result of publication and distribution of investment research. Many European regulators require a firm to establish, implement and maintain such a policy. This report has been issued in the U.K. only to persons of a kind described in Article 19 (5), 38, 47 and 49 of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2005 (all such persons being referred to as "relevant persons"). This document must not be acted on or relied on by persons who are not relevant persons. Any investment or investment activity to which this document relates is only available to relevant persons and will be engaged in only with relevant persons. In other EEA countries, the report has been issued to persons regarded as professional investors (or equivalent) in their home jurisdiction. Australia: This material is issued and distributed by JPMSAL in Australia to "wholesale clients" only. JPMSAL does not issue or distribute this material to "retail clients". The recipient of this material must not distribute it to any third party or outside Australia without the prior written consent of JPMSAL. For the purposes of this paragraph the terms "wholesale client" and "retail client" have the meanings given to them in section 761G of the Corporations Act 2001. Germany: This material is
233
distributed in Germany by J.P. Morgan Securities Ltd., Frankfurt Branch and J.P.Morgan Chase Bank, N.A., Frankfurt Branch which are regulated by the Bundesanstalt fr Finanzdienstleistungsaufsicht. Hong Kong: The 1% ownership disclosure as of the previous month end satisfies the requirements under Paragraph 16.5(a) of the Hong Kong Code of Conduct for Persons Licensed by or Registered with the Securities and Futures Commission. (For research published within the first ten days of the month, the disclosure may be based on the month end data from two months prior.) J.P. Morgan Broking (Hong Kong) Limited is the liquidity provider/market maker for derivative warrants, callable bull bear contracts and stock options listed on the Stock Exchange of Hong Kong Limited. An updated list can be found on HKEx website: http://www.hkex.com.hk. Japan: There is a risk that a loss may occur due to a change in the price of the shares in the case of share trading, and that a loss may occur due to the exchange rate in the case of foreign share trading. In the case of share trading, JPMorgan Securities Japan Co., Ltd., will be receiving a brokerage fee and consumption tax (shouhizei) calculated by multiplying the executed price by the commission rate which was individually agreed between JPMorgan Securities Japan Co., Ltd., and the customer in advance. Financial Instruments Firms: JPMorgan Securities Japan Co., Ltd., Kanto Local Finance Bureau (kinsho) No. 82 Participating Association / Japan Securities Dealers Association, The Financial Futures Association of Japan, Type II Financial Instruments Firms Association and Japan Securities Investment Advisers Association. Korea: This report may have been edited or contributed to from time to time by affiliates of J.P. Morgan Securities (Far East) Ltd, Seoul Branch. Singapore: JPMSS and/or its affiliates may have a holding in any of the securities discussed in this report; for securities where the holding is 1% or greater, the specific holding is disclosed in the Important Disclosures section above. India: For private circulation only, not for sale. Pakistan: For private circulation only, not for sale. New Zealand: This material is issued and distributed by JPMSAL in New Zealand only to persons whose principal business is the investment of money or who, in the course of and for the purposes of their business, habitually invest money. JPMSAL does not issue or distribute this material to members of "the public" as determined in accordance with section 3 of the Securities Act 1978. The recipient of this material must not distribute it to any third party or outside New Zealand without the prior written consent of JPMSAL. Canada: The information contained herein is not, and under no circumstances is to be construed as, a prospectus, an advertisement, a public offering, an offer to sell securities described herein, or solicitation of an offer to buy securities described herein, in Canada or any province or territory thereof. Any offer or sale of the securities described herein in Canada will be made only under an exemption from the requirements to file a prospectus with the relevant Canadian securities regulators and only by a dealer properly registered under applicable securities laws or, alternatively, pursuant to an exemption from the dealer registration requirement in the relevant province or territory of Canada in which such offer or sale is made. The information contained herein is under no circumstances to be construed as investment advice in any province or territory of Canada and is not tailored to the needs of the recipient. To the extent that the information contained herein references securities of an issuer incorporated, formed or created under the laws of Canada or a province or territory of Canada, any trades in such securities must be conducted through a dealer registered in Canada. No securities commission or similar regulatory authority in Canada has reviewed or in any way passed judgment upon these materials, the information contained herein or the merits of the securities described herein, and any representation to the contrary is an offence. Dubai: This report has been issued to persons regarded as professional clients as defined under the DFSA rules. General: Additional information is available upon request. Information has been obtained from sources believed to be reliable but JPMorgan Chase & Co. or its affiliates and/or subsidiaries (collectively J.P. Morgan) do not warrant its completeness or accuracy except with respect to any disclosures relative to JPMS and/or its affiliates and the analyst's involvement with the issuer that is the subject of the research. All pricing is as of the close of market for the securities discussed, unless otherwise stated. Opinions and estimates constitute our judgment as of the date of this material and are subject to change without notice. Past performance is not indicative of future results. This material is not intended as an offer or solicitation for the purchase or sale of any financial instrument. The opinions and recommendations herein do not take into account individual client circumstances, objectives, or needs and are not intended as recommendations of particular securities, financial instruments or strategies to particular clients. The recipient of this report must make its own independent decisions regarding any securities or financial instruments mentioned herein. JPMS distributes in the U.S. research published by non-U.S. affiliates and accepts responsibility for its contents. Periodic updates may be provided on companies/industries based on company specific developments or announcements, market conditions or any other publicly available information. Clients should contact analysts and execute transactions through a J.P. Morgan subsidiary or affiliate in their home jurisdiction unless governing law permits otherwise. "Other Disclosures" last revised September 30, 2011.
Copyright 2011 JPMorgan Chase & Co. All rights reserved. This report or any portion hereof may not be reprinted, sold or redistributed without the written consent of J.P. Morgan. #$J&098$#*P
Unless otherwise expressly noted, all data and information for charts, tables and exhibits contained in this publication have been sourced via J.P. Morgan information sources.
234
235
236

JPM Fixed in

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

JPM Fixed in

Uploaded by

Copyright:

Available Formats

European Rates Strategy

J.P. Morgan Securities Ltd. November 24, 2011

Global Fixed Income Markets 2012 Outlook

(44-20) 7777-3370 pavan.wadhwa@jpmorgan.com

J.P. Morgan Securities Ltd.

See page 233 for analyst certification and important disclosures.

Key 2012 Global Fixed Income Markets Trading Themes

Euro Pavan Wadhwa, Fabio Bassi, Gianluca Salford

US Terry Belton, Srini Ramaswamy

Japan Takafumi Yamawaki, Yuya Yamashita

Inflation Jorge Garayo, Francis Diamond, Kimberly Harano

Australia / New Zealand Sally Auld

The sovereign pandemic

Greek EFSF crisis concerns

Irish concerns Italian crisis and ECB buys IT/ES

PROBLEM: Overleverage Loss of capital market access Solvency crisis

/ ion t s es s rec l unre ia s oc

PROBLEM: Slow growth Recession risk Social unrest

Why is this crisis so pernicious?

Exhibit 3: Peripheral sovereign balance sheets have been severely impaired

158 112 102 120 68 112 97 85 82 88

* Official estimate as provided by Eurostat. ** Simple average. Source: Eurostat

Aaa Aa A Baa Ba B <= Caa Total

48% 26%^ 1%^^ 75%

48% 30% 9% 6% 4% 2% 1% 100%

Greece Ireland Portugal Italy Spain Belgium France Av erage*

2425 625 910 494 469 282 147 765

2471 1142 1161 575 486 309 188 905

98% 55% 78% 86% 97% 91% 79% 85%

67% 83% 80% 45% 38% 49% 66% 81%

140 120 100 80 60 40 20 0 -20 Aug

Netherlands 3 12 36 5 53 109 102 107%

Source: BIS, ECB. Data as-of 2Q11.

12 10 8 6 4 2 0 2006 2007 2008 2009 2010 2011 Spain Greece

Claims on Greece Ireland Italy Portugal Spain Total

% chg -41% -18% -4% -16% -4% -11%

Source: BIS. Data as-of 2Q11.

Current : 216bp Current : 198bp

150 100 50 0 -50 2007

Cov ered bonds

Source: J.P. Morgan Maggie Credit Index

300 250 200 150 100 50 Total = 65bn (53+12bn) 2007

Av g Total = 130bn (90+40bn)

Bank debt + CB Bank debt only

See Overview, Global Fixed Income Markets Weekly, 30 September 2011.

Total = 50bn (33+17bn) 2010 2011

estimated capital shortfall 254 48 43 53 35 40 17 18

J.P.Morgan EBA estimate 151 39 38 27 20 10 9 7

y = -70.0x + 6.7 R 2 = 31%

Analysing the risk posed by peripherals

Chg. 7.4 1.3 1.2 0.0 3.5 2.7 0.1

Greece Ireland Portugal Italy Spain Av g., peripherals* Euro area

17.6 14.2 12.5 8.3 22.6 15.0 10.2

10.2 12.9 11.3 8.3 19.1 12.4 10.1

The latter is typically more up-to-date for latest economic trends.

Exhibit 21: Fiscal sustainability in Italy and Spain is uncertain

2014 125 1.2 5.0 5.1 85 1.6 -0.6 5.2

2015 123 2.7 5.0 5.3 86 2.2 0.8 5.4

2016 121 3.0 5.0 5.4 87 2.4 2.0 5.6

...2020 114 2.8 5.0 5.9 84 3.4 3.0 6.0

...2030 100 2.8 5.0 6.3 77 3.4 3.0 6.4

Marginal borrow ing rate on debt 7 22 5 16

SWFs IMF* 40 20 60 210 110 320