Professional Documents
Culture Documents
William Bailey, Benoit Couet, Ashish Bhandari, Soussan Faiz, Sunaram Srinivasan and
Helen Weeds. Unlocking the Value of Real Options. (Oilfield Review, 2004).
Understand and apply a net present value (NPV) calculation to make investment
decisions in a gas/oil field.
Explain how a binomial lattice is used in the valuation of an asset or option and be
able to calculate the value of an up or down move.
2.
Summarize the process used to determine price differentials and identify factors that
influence the price differential including the equivalence to the buyer principle.
Describe the role of price reporting agencies (PRAs) in price identification; summarize
the methodologies used by PRAs to assess commodity prices, and identify criticisms
of PRA price assessment.
Understand the mechanics and specifications of the 21-day BFOE (Forward Brent), the
Brent Futures, the Exchange for Physical (EFP) and the Dated Brent/BFOE contracts.
Define Contracts for Differences (CFDs) and understand its application when hedging
basis risk associated with Forward Brent contracts or deriving forward prices from a
combination of Dated Brent prices and CFDs.
Compare and contrast the Brent, WTI, and Dubai-Oman crude oil benchmarks in terms
of liquidity, price transparency, and available financial products.
Summarize the mechanics of WTI futures contracts including related delivery requirements, and compare WTI Posting-Plus (P-Plus) pricing to NYMEX CMA pricing.
Understand the logistical challenges that can impact the effectiveness of WTI as a
global crude oil benchmark.
Explain how the Dubai benchmark price can be calculated using swaps.
2014
Describe Renewable Identification Numbers (RINs) and explain how RINs are
produced and traded.
Summarize the mechanics of the Renewable Fuels Standard (RFS) and describe the
classes and categories of RINs which can be used to meet biofuel requirements as
part of the RFS.
International Gas Union. Wholesale Gas Price FormationA Global View of Price Drivers
and Regional Trends. (Sections 1 to 5 and 8 to 10 only)
Understand and apply the following natural gas pricing terms: wellhead price,
border/beach price, hub price, citygate price, end user price and netback price.
Describe potential short, medium and long-term supply-side and demand-side drivers
Summarize the eight key mechanisms for pricing gas and identify the geographic
Describe the relationship between a local gas pricing mechanism, the observed
market price and the hypothetical market-clearing price.
Identify the factors that influence the volatility of natural gas prices, including
oil-linked prices.
Assess the relationship between price volatility and natural gas supply across various
hypothetical price levels.
2014
Richard Morse and Gang He, The World's Greatest Coal Arbitrage: China's Coal Import
Behavior and Implications for the Global Coal Market. (PESD Stanford, August 2010).
Understand the basic economics of the Chinese coal market; compare domestic
reserves to domestic demand.
Explain how Chinese arbitrage has impacted prices on the global coal market.
William Bailey
Benot Cout
Ridgeeld, Connecticut, USA
Ashish Bhandari
El Paso Corporation
Houston, Texas, USA
Soussan Faiz
Strategic Management Consultant
Walton on Thames, Surrey, England
Sundaram Srinivasan
Sugar Land, Texas
Helen Weeds
University of Essex
Colchester, England
Oileld Review
Oil price
Time
Winter 2003/2004
Time
Cash flow
Discount
factor
Present value
of cash flow
Present
5000
1.0000
5000
One year
+4500
0.9091
+4091
Two years
+3000
0.8264
+2479
+2500
+1570
A Synthetic-Reservoir Example
Period
Time, years
0.6
50.0
1.2
75.0
1.8
107.5
2.4
150.0
3.0
177.5
Well 1 Well 5
Well 2
Well 4
Well 6
Well 3
Oil saturation
0.0
0.2
0.4
0.6
0.8
> Reservoir model of the synthetic Charon eld. This ECLIPSE reservoir model provided input for
obtaining production predictions, using a large number of simulations with geostatistically
derived porosity and permeability values.
Oileld Review
Oil/water
contact
depth, ft
Field-wide average,
ratio of net-togross thickness
Average
porosity
Original oil
in place,
million BOE
Initial oil
production,
B/D
Low
9625
0.65
12.5%
138.6
25,384
Median
9650
0.75
14.4%
228.2
27,930
High
9675
0.85
16.3%
350.4
28,225
> Results of model realizations. Three models represent the low (5%), median
(50%) and high (95%) production predictions in Charon eld.
30,000
25,000
20,000
15,000
High
10,000
Median
Low
5,000
0
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
5500
6000
Time, days
250
236.3
200
150
100
Median case
50
0
50
100
150
200
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
5500
6000
Time, days
> Calculation of Charon net present value (NPV). Production begins in the third year of the project
and declines (top). The low (5%), median (50%) and high (95%) probability model predictions are
shown. The projects cumulative cash ow for the median case shows the expenditures in the
rst three years followed by income over the remainder of the project (bottom). The median-case
NPV of the project is $236.3 million.
Winter 2003/2004
Call optionthe right, but not the obligation, to buy shares at the exercise price within
a given time period.
Put optionthe right, but not the obligation, to sell shares at the exercise price within a
given time period.
Widgets, Inc., has a moderately volatile share price with a current price of $100. For a
small fee, an investor may buy a call option with an exercise price of $110. If the share
price subsequently rises to $120, the option holder would exercise the option to buy the
shares for the agreed exercise price of $110 and sell them on the open market for $120,
making a profit of $10 per share less the option-purchase fee.
Alternatively, if the investor has a put option with an exercise price of $90 and shares of
Widgets, Inc. fall below $90, it will benefit the option holder to buy shares from the open
market at the lower price and exercise the option to sell them at $90. Both examples
ignore the usual transaction fees paid to a broker.
Oileld Review
Winter 2003/2004
The extraction option is contingent on exercising the development option, which is contingent
on exercising the exploration option. At each
stage, a company obtains information to determine whether the project should be taken to the
next stage.
Comparing Financial- and
Real-Option Parameters
The variables used to value a nancial option can
be compared with their analogs in real options.
An option to develop oil reserves, for example, is
similar to a nancial call option (below).
The NPV of the developed hydrocarbon
reserveswhat they would be worth at todays
pricesis similar to the price of the underlying
stock, S, in a financial option. The NPV of the
expenditure needed to develop the reserves is
like a financial options exercise price, X. The
time left on an exploration and production
(E&P) lease is equivalent to the time to expiration of a nancial option, T. The risk-free rate of
return, rf the rate of return on a guaranteed
asset, such as a government bondis identical
for both nancial and real options. The volatility
of cash flows from an E&P project, including
hydrocarbon price uncertainty, is analogous to
the volatility of stock prices, . Finally, profits
foregone because production has been delayed
are like the lost dividends in the financial
option, . As long as management holds an unexercised option to invest in a project, it foregoes
the money that would have flowed from it had
the project been producing revenue.
The analogies between real and financial
options are not exact. Trying to force real
options into a conventional financial-options
framework may result in misleading outcomes.
One key difference in the two options types is
that the exercise price of a financial option is
Variable
Stock price
Exercise price
Time to expiration
rf
Dividends foregone
> Comparison of nancial and real options. The variables of a nancial call option can be
related to similar variables for a real option to develop oil reserves.
10
Oileld Review
5
Probability distribution of
future assets
S0u5
S0u4
S0u3
S0u3d1
S0u2
S0u2d1
S0u2d2
S0u1d1
S0d
S0u3d2
1 2
S0u d
S0d2
S0u2d3
S0u1d3
S0d3
u = exp( T )
d= 1
u
Large Lattice
S0u1d4
Price
S0u1
S0
S0u4d1
S0d4
S0d5
Probability
> Construction of a lattice of the underlying asset. The deterministic value for the asset today, such as a stock price, goes into the
left-most lattice node (left). In the rst time step, this value can increase by a multiplicative factor, u, which is based on the volatility,
, and length of the time step, T, or it can decrease by the inverse of that factor, d. Each node in subsequent time steps can similarly
increase or decrease, resulting in an expanding lattice. Results from a ve-step lattice are coarse. As the number of steps increases,
T becomes smaller and the resolution increases as the lattice becomes larger. A probability distribution of future assets (green
curve) can be obtained from the values in the right-hand column of a lattice with thousands of steps (right). The assumptions
governing the denition of u and d factors always give rise to a log-normal distribution of asset value at expirationthis is a basic
assumption of the Black-Scholes model.
Winter 2003/2004
d 2 = d 1 - * T,
and where N(d) = cumulative normal distribution function, ln is the natural logarithm and other terms are
dened in the text.
17. Rogers J: Strategy, Value and RiskThe Real Options
Approach. Basingstoke, England: Palgrave (2002): 61.
18. In a European option, uncertainty is assumed to be fully
resolved at expiration. However, valuation of Americantype options can be more complex and caution is
required. An American option can be exercised at any
time prior to expiration, but that does not mean that all
uncertainty has been resolved at the time the decision is
made. New information about project uncertainties is
likely to be streaming in all the time, so the decision is
based on incomplete information. Unless all pertinent
uncertainty has been resolved, it might be prudent to
wait until the last moment to decide on the option.
19. Some ROV specialists argue that it is better to keep
technical and market uncertainties separate, especially
when managerial decision-making is tied to the resolution
of technical uncertainty.
11
5
Maximum (SX, 0)
$80
$67.66
C
Excercise cost
X = $100
Stock price, S,
at expiration
A
$50
$150
$20
$120
$0
$90
$0
$60
$0
$30
Option
Value
p=
exp(rf*T )-d
u-d
> Construction of a valuation lattice. Nodes in a valuation lattice are constructed from right to left.
The asset value, such as a stock price, S, at expiration is taken from the lattice of the underlying
asset. The exercise cost, X, is known in advance. The nodes in Column 5 contain the difference
between stock and exercise price, unless that difference is negative, in which case the node contains
zero. The value in the node labelled C comes from the two adjacent Column 5 nodes, A and B, and
uses the risk-neutral probability, p, as shown in the formula (bottom left). Remaining nodes and columns
are constructed similarly, from right to left. The single node on the left contains the value of the option.
Salvage an Investment
$180
C = [ p*A+(1-p)*B]*exp (-rf*T )
inductionto obtain an option value at the farthest left node of the lattice. Backward
induction relies on a factor p, the risk-neutral
probability of a movement in the price of the
underlying asset. This is the probability that
would prevail in a world in which investors were
indifferent to risk. Applying this to each pair of
vertically adjacent nodes in the lattice provides
the real-option value at the farthest left node of
the lattice.
Binomial lattices are often referred to as
binomial trees. However, the two methods
operate differently. Trees require an analyst to
specify probabilities and appropriate discount
rates at each node, which can be highly subjective. ROV, embodying ideas such as risk-neutral
probability for financial uncertainty and riskfree rate of interest, is less subjective.20
12
Oileld Review
Input Parameters
= 66.41%
T = 0.6
u = exp(T )
= exp(0.6641*0.6)
= 1.67265
2
1
4
1849.4
1105.7
661.0
395.2
236.3
= 0.59785
3
1105.7
661.0
395.2
1
d= 1 =
u
1.67265
141.3
395.2
236.3
236.3
141.3
exp(rf *T)-d
u-d
= exp(0.05*0.6)-0.59785
1.67265-0.59785
p=
84.4
141.3
84.4
50.5
50.5
30.2
= 0.40250
18.0
Salvage Value
3093.3
Period
Years
0.6
Value,
million $
50.0
1.2
75.0
1.8
107.5
2.4
150.0
3.0
177.5
3
1105.7
2
668.1
1
420.6
4
1849.4
1105.7
retain
661.0
continue
407.4
continue
285.5
start
395.2
retain
257.3
275.3
continue
continue
209.1
continue
retain
continue
continue
continue
continue
5
3093.3
200.4
177.5
salvage
continue
175.2
continue
172.3
continue
167.2
continue
177.5
salvage
172.3
continue
177.5
salvage
Work from right to left
Salvage if < 177.5
Example: backward-induction
calculation
[
=
3093.3
1849.4
* p+
1105.7
3093.3
Retain ownership
3093.3
1105.7
Retain ownership
1105.7
395.2
Retain ownership
395.2
177.5
Salvage
141.3
177.5
Salvage
50.5
177.5
Salvage
18.0
1849.4
* (1-p)] * exp(- r f* T )
661.0
257.3
172.3
172.3
Winter 2003/2004
Lattice of underlying
13
14
Option to switch to another planA switching option can provide a hedge against the
likelihood that another technology or project
will be more economic sometime in the future
(see Switch Option, page 16).
Sequential or compound optionsReal
options may lead to additional investment
opportunities when exercised. The process of
exploration, development and production
described earlier in this article was a sequential option.
This list of options is not exhaustive. Many
other types of options are available. El Paso Corporation, the largest pipeline company in North
America and a leading provider of natural gas
services, used a location spread optionrelying
on a difference in a price between different locationsto evaluate a new line of business. Various
other spread options are possible, for example
based on different prices at different times or at
different stages of commodity processing.
Real Options for LNG Transport
El Paso owns a liquefied natural gas (LNG)
terminal at Elba Island, Georgia, USA, one of
only four land-based terminals in the USA. The
company investigated purchasing transport
vessels and expanding into the LNG transport
business. Each tanker, outfitted specifically for
use in LNG transport with a regasication capability for downloading at offshore buoys called
energy-bridge buoys, costs several hundred million US dollars.
The essence of the problem facing the evaluation team was how to value shipping and
diversion exibility. The company had a variety of
potential LNG sources and destinations, and the
evaluation was intended to determine how many
tanker ships El Paso should purchase.
El Paso felt DCF was decient for this analysis. The LNG market and its related shipping
were relatively new ventures for El Paso, and the
company had no history for forecasting the revenues and costs required by DCF. Even if those
forecasts had been available, the DCF technique
does not have the exibility necessary to reect
the additional value of a price difference between
delivery locations that occurs only for a brief time
period. The team tried to model the simple case
of a xed source and destination using DCF, but
the model could not correctly value inbuilt
options allowing El Paso not to sell if the LNG
delivery price did not cover variable expenses.
The base case for this ROV involves transporting LNG from a terminal in Trinidad, West
Indies, to the companys Elba Island facility. The
LNG producer in Trinidad would pay for infrastructure costs to enable this base-case trade and
would in turn receive the netback gas price,
which is the gas price at Elba Island less the
cost of shipping and regasication and less the
margin paid to El Paso. For example, for the
analysis presented here this margin was
assumed to be $0.20/MMBtu [$0.19/million J].
The NPV of this business over 20 years was
$176.7 million.
The rst option evaluated included diversion
flexibilityadding a second destination terminal offshore New York, New York, USA. El Paso
evaluated both intrinsic and extrinsic values for
this option. The intrinsic value of this spread
option represents the difference in pricethe
basis spreadbetween the Georgia and New
York markets (next page). The extrinsic value
includes the effects of time and reflects the
probability that the basis spread will change
over the 20-year period of the analysis.
In this spread option, El Paso would buy
the LNG on the basis of the Elba Island price
and sell it at the New York price, when that
choice adds value. Otherwise, El Paso would sell
at Elba and receive no incremental value. With
an average basis spread of $0.62/MMBtu
[$0.59/million J], the total intrinsic value of this
spread option is $558.7 million. In this model,
El Paso would assume the costs for converting
the terminals and purchasing an additional ship
to effect this option. The net value of the option
after those expenses is $68.5 million. Including
the variability over time gives an additional
extrinsic value of $101.7 million.
The company then added in the value of having multiple choices of source and destination,
termed a rainbow option. The value of a rainbow
option increases with increased price volatility
at the individual locations, and it also increases
when the cross-correlations between prices are
low. With two additional destination options, offshore New York and Cove Point, Maryland, USA,
there is an additional value of $14.8 million,
even though the price correlations between
these pairs of locations are high. The rainbow
option value increases when there is more exibility in source and destination locations. In
certain scenarios with additional sources in the
Oileld Review
Middle East and Africa and additional destinations in Europe and North America, El Paso
found the rainbow option added more than $100
million to the value of each ship.
The evaluation team provided a caveat to the
company. Spread options tend to overestimate
available exibility, because contractual obligations would have to be maintained. In addition,
the effects of price shifts caused by any reduction in supply were not included in the analysis.
Although real-options analysis indicated a
positive value to a business model based on LNG
imports to the USA and to diversion exibility as
a value-maximizing technique for shipping, El
Paso made a strategic business decision not to
enter this market.
Land-based terminal
Offshore transfer buoy
Base-case trade route
Spread-option trade routes
USA
TRINIDAD
Winter 2003/2004
0
0
1000
500
2000
1000
1500
3000 km
2000 miles
> Liquied natural gas (LNG) transport routes in a spread option. El Paso
Corporation evaluated the LNG transport business using a base case between
Trinidad, West Indies, and its terminal at Elba Island, Georgia, USA. The
company considered purchasing transport ships with regasication
capabilities to give it the capability to transport gas to other locations, such
as Cove Point, Maryland, USA, or New York, New York, USA. This rainbow
option increased the value of the business opportunity.
tives may often be evaluated using simpler methods that automatically average out the
possibilities. For example, a forward contract,
which obligates the holder of the contract to buy
or sell an asset for a predetermined price at a
predetermined time in the future, may be valued
simply, without the volatility assumption that is
required in the Black-Scholes or lattice valuation of a European option.
Within the class of options there are many distinctions. One is the distinction between nancial
and real options that has been discussed. Another
is that between options that are purely internalresiding solely within the company
itselfand ones in which an outside party provides flexibility for some agreed up-front
payment. Many real options possess just an internal character, while nancial derivatives normally
exist in the presence of a contracted external
party. Fortunately, all these option types can be
computed using the same techniques, most
notably standard lattice-type methodologies.
Real-World Complications
A real-options methodology attempts to model
behaviors of real properties. However, the many
possibilities created by human ingenuity limit
any such modeling. Actual situations typically
have many embedded options, making any analysis complicated. The few examples discussed in
this section illustrate a few types of complications that may have to be dealt with in using
real options.
The holder of a financial option is guaranteed that the option may be held until the
expiration date and, apart from general market
movement, its value cannot be undermined by
the actions of other individuals. In most real
options, there is no such guarantee.
Two oil companies might hold identical
leases on adjoining blocks. In effect, they would
both have identical options to spend money on
exploration and receive undeveloped resources
in return.
(continued on page 18)
15
Switch Option
16
Case 50
Case 60
50
60
Volatility
66.41%
71.28%
NPV, million $
236.27
228.99
> Comparison of volatility and net present value (NPV) for two surface
separator cases in the synthetic Charon example.
4.0
3.5
3.0
2.5
Switching option value for
200-step lattice is 1.653
2.0
200-step
lattice
1.5
Switching option value for
five-step lattice is 1.305
1.0
0.5
0.0
17.72
10
15
Switching cost, million $
Five-step
lattice
20
25
> Effect of lattice size on option valuation. The coarse, ve-step lattice has
been used for illustrative purposes only, yielding a less accurate option value
than the more rened 200-step lattice. At the $17.72 million switching cost,
that ner lattice indicates the option value is $1.653 million.
the last column of the Case 60 lattice, comparing this to the last column of the Case 50
lattice, and selecting the larger value at each
node. This reects Oberons right to choose
the better of the two cases in any eventuality.
The option value is then computed by backward induction using the Case 50 risk-neutral
probabilities, p.
Oileld Review
0
(Now)
1
(0.2 year)
2
(0.4 year)
3
(0.6 year)
4
(0.8 year)
5
(1.0 year)
1,043.10
775.07
575.91
575.91
427.93
317.97
236.27
427.93
317.97
175.56
317.97
236.27
236.27
175.56
130.45
175.56
130.45
96.93
96.93
72.02
53.52
1,127.26
819.57
595.86
595.86
433.21
314.96
228.99
433.21
314.96
166.48
314.96
228.99
228.99
166.48
121.04
166.48
< Lattices for synthetic Charon surface separator cases. Case 50 and
Case 60 have different lattices of
the underlying asset, but the lattice
structure is the same, allowing a
node-by-node comparison between
them (top). Nodes in Case 60 are
greyed out, except for the nal column, to indicate that no decision is
made until the end of one year. The
last column of the valuation lattice
is constructed by comparing the
value of Case 50 to equivalent node
value of Case 60 minus the implementation cost of $17.72 million
(bottom). This also provides the
decision to keep Case 50 or switch
to Case 60. The other nodes of the
valuation lattice are constructed by
back-regression, using the risk-neutral probabilities from Case 50, the
base case. The value of the project
with the switch option is
$237.57 million.
121.04
88.00
88.00
63.98
46.52
0
(Now)
1
(0.2 year)
2
(0.4 year)
3
(0.6 year)
4
(0.8 year)
5
(1.0 year)
MAX(Case 50,
Case 60-17.72)
1,109.64
589.77
continue
434.24
320.85
237.57
Start
continue
175.64
continue
continue
318.40
805.44
Switch to Case 60
428.91
Switch to Case 60
236.27
Keep Case 50
130.45
continue
Keep Case 50
72.02
Keep Case 50
continue
continue
317.97
continue
236.46
continue
130.45
continue
175.56
continue
96.93
continue
175.56
96.93
continue
continue
Winter 2003/2004
578.14
53.52
Keep Case 50
accelerated cash and throughput from reservoir production, if conditions warrant. Proteus
gets an up-front premium of $1.653 million
and a locked-in payment of $17.72 million if
Oberon chooses to upgrade the facility. Proteus has a cash incentive to explore more
cost-effective and efcient design solutions
for the upgrade.
17
Actions of one company can affect the business results of the other. Most governments now
insist on unitization, an arrangement that
requires parties on both sides to jointly develop
reserves located on more than one lease or auctioned tract. Each party pays a share of the costs
and receives a proportionate amount of the
revenues. In a sense, when governments do
this, they are ensuring the purity of the real
options involved.
It is possible to conceive of circumstances
such as constructing a pipeline to an area where
only one is neededin which if Company A took
up the option to invest, it would preempt Company B from doing so, rendering Bs option
worthless, or certainly worth less. The realoptions approach attaches a positive value to
delay, but in instances such as this one, delay
can undermine value.23
Finally, the parameters used in real-options
calculations can be difcult to determine. There
is no simple road map to computing volatility
and there is still debate over the correct
approach to nding this value. Obtaining an estimate often entails performing a Monte Carlo
simulation on the existing DCF model and examining the standard deviation of the natural
logarithm of the cash-flow returns. The cost of
deferment requires knowledge of the profits
foregone during the period prior to exercising an
option, but the value of the missed cash flows
may be poorly known.
Financial-option pricing relies on an assumption that the underlying asset can be traded,
meaning there is a large liquid market for the
asset. This is often not true for real options. Factors affecting the prices of nancial options are
also easier to determinethat is, more transparentthan those for real options.
The simplified discussion in this article is
intended only to introduce the concept of real
options. It uses simple examples that correlate
with nancial options. Use for an actual case is
typically more complicated, with an expanding
array of possible options available, as demonstrated in the El Paso LNG case. Ultimately, real
options are not financial options. The techniques of financial options provide a basis for
23. For a discussion of such investment behavior: Weeds H:
Strategic Delay in a Real Options Model of R&D Competition, Review of Economic Studies 69 (2002): 729747.
24. Trigeorgis, reference 5: 3.
18
Synthetic-Reservoir Intervention
Oileld Review
0
(Now)
1
(0.2 year)
2
(0.4 year)
3
(0.6 year)
4
(0.8 year)
5
(1.0 year)
3,130,796
2,617,977
2,189,157
1,830,577
1,530,731
1,280,000
1,530,731
1,530,731
1,280,000
1,280,000
1,070,338
2,189,157
1,830,577
1,070,338
895,019
1,070,338
895,019
748,416
748,416
625,827
523,317
Max(S-X, 0)
2,380,796
1,875,439
1,439,157
1,454,008
1,102,742
810,248
574,776
1,088,039
795,582
365,671
780,731
537,463
552,378
334,604
198,119
320,338
153,291
73,335
0
0
S-X
2,380,796
1,875,439
1,454,008
1,102,742
810,139
566,578
795,582
780,731
537,463
552,166
349,746
1,439,157
1,088,039
335,189
167,184
320,338
152,481
13,267
1,584
116,711
226,683
> Comparison of an option and an alternative. The lattice method can be used to
value an alternative. Both use the same lattice of the underlying (top). For an
option, the right-hand column is the maximum value of zero and the difference
between the underlying value and the $750,000 implementation cost (middle).
The values for an alternative can be negative, because the function is simply the
difference between the value and the implementation cost (bottom). This leads to
an $8,198 difference in value between the option, valued at $574,776, and the
alternative, valued at $566,578.
Winter 2003/2004
19
Bassam Fattouh1
WPM 40
January 2011
Bassam Fattouh is the Director of the Oil and Middle East Programme at the Oxford Institute for Energy Studies;
Research Fellow at St Antonys College, Oxford University; and Professor of Finance and Management at the
School of Oriental and African Studies, University of London. I would like to express my gratitude to Argus for
supplying me with much of the data that underlie this research. I would also like to thank Platts for providing me
with the data for Figure 21 and CME Group for providing me with the data for Figure 13. The paper has benefited
greatly from the helpful comments of Robert Mabro and Christopher Allsopp and many commentators who
preferred to remain anonymous but whose comments provided a major source of information for this study. The
paper also benefited from the comments received in seminars at the Department of Energy and Climate Change, UK,
ENI, Milan and Oxford Institute for Energy Studies, Oxford. Finally, I would like to thank those individuals who
have given their time for face-to-face and/or phone interviews and have been willing to share their views and
expertise. Any remaining errors are my own.
The contents of this paper are the authors sole responsibility. They do not
necessarily represent the views of the Oxford Institute for Energy Studies or any of its
members.
Copyright 2011
Oxford Institute for Energy Studies
(Registered Charity, No. 286084)
This publication may be reproduced in part for educational or non-profit purposes without special
permission from the copyright holder, provided acknowledgment of the source is made. No use of this
publication may be made for resale or for any other commercial purpose whatsoever without prior
permission in writing from the Oxford Institute for Energy Studies.
ISBN
978-1-907555-20-6
Contents
Summary Report ........................................................................................................................................... 6
1.
Introduction ......................................................................................................................................... 11
2.
3.
4.
Oil Price Reporting Agencies and the Price Discovery Process ......................................................... 30
5.
6.
7.
9.
Conclusions ......................................................................................................................................... 78
References ................................................................................................................................................... 81
List of Figures
Figure 1: Price Differentials of Various Types of Saudi Arabias Crude Oil to Asia in $/Barrel .............. 21
Figure 2: Differentials of Term Prices between Saudi Arabia Light and Iran Light Destined to Asia (FOB)
(In US cents) ............................................................................................................................................... 23
Figure 3: Difference in Term Prices for Various Crude Oil Grades to the US Gulf (Delivered) and Asia
(FOB) .......................................................................................................................................................... 24
Figure 4: Price Differential between Dated Brent and BWAVE ($/Barrel) ............................................... 26
Figure 5: Price Differential between WTI and ASCI ($/Barrel) (ASCI Price=0) ....................................... 26
Figure 6: Brent Production by Company (cargoes per year), 2007 ............................................................ 37
Figure 7: Falling output of BFO ................................................................................................................. 38
Figure 8: Trading Volume and Number of Participants in the 21-Day BFOE Market ............................... 42
Figure 9: Average Daily Volume and Open Interest of ICE Brent Futures Contract ................................. 44
Figure 10: Pricing basis of Dated Brent Deals (1986-1991); Percentage of Total Deals ........................... 45
Figure 11: Reported Trade on North Sea CFDs (b/d) ................................................................................. 46
Figure 12: US PADDS ................................................................................................................................ 52
Figure 13: Monthly averages of volumes traded of the Light Sweet Crude Oil Futures Contract ............. 55
Figure 14:Liquidity at Different Segments of the Futures Curve (October 19, 2010) ................................ 56
Figure 15: Spot Market Traded Volumes (b/d) (April 2009 Trade Month) ................................................ 57
Figure 16: Spread between WTI 12-weeks Ahead and prompt WTI ($/Barrel) ......................................... 59
Figure 17: WTI-BRENT Price Differential ($/Barrel)................................................................................ 60
Figure 18: Dubai and Oman Crude Production Estimates (thousand barrels per day) ............................... 62
Figure 19: Spread Deals as a Percentage of Total Number of Dubai Deals ............................................... 63
Figure 20: Oman-Dubai Spread ($/Barrel) ................................................................................................. 64
Figure 21: Dubai Partials Jan 2008 - Nov 2010 .......................................................................................... 65
Figure 22: daily Volume of Traded DME Oman Crude Oil Futures Contract ........................................... 67
Figure 23: Volume and Open Interest of the October 2010 Futures Contracts (Traded During Month of
August)........................................................................................................................................................ 68
4
Summary Report
The view that crude oil has acquired the characteristics of financial assets such as stocks or bonds has
gained wide acceptance among many observers. However, the nature of financialisation and its
implications are not yet clear. Discussions and analyses of financialisation of oil markets have partly
been subsumed within analyses of the relation between finance and commodity indices which include
crude oil. The elements that have attracted most attention have been outcomes: correlations between
levels, returns, and volatility of commodity and financial indices. However, a full understanding of the
degree of interaction between oil and finance requires, in addition, an analysis of interactions, causations
and processes such as the investment and trading strategies of distinct types of financial participants; the
financing mechanisms and the degree of leverage supporting those strategies; the structure of oil
derivatives markets; and most importantly the mechanisms that link the financial and physical layers of
the oil market.
Unlike a pure financial asset, the crude oil market also has a physical dimension that should anchor
prices in oil market fundamentals: crude oil is consumed, stored and widely traded with millions of
barrels being bought and sold every day at prices agreed by transacting parties. Thus, in principle, prices
in the futures market through the process of arbitrage should eventually converge to the so-called spot
prices in the physical markets. The argument then goes that since physical trades are transacted at spot
prices, these prices should reflect existing supply-demand conditions.
In the oil market, however, the story is more complex. The current market fundamentals are never
known with certainty. The flow of data about oil market fundamentals is not instantaneous and is often
subject to major revisions which make the most recent available data highly unreliable. Furthermore,
though many oil prices are observed on screens and reported through a variety of channels, it is important
to explain what these different prices refer to. Thus, although the futures price often converges to a spot
price, one should aim to analyse the process of convergence and understand what the spot price in the
context of the oil market really means.
Unfortunately, little attention has been devoted to such issues and the processes of price discovery in oil
markets and the drivers of oil prices in the short-run remain under-researched. While this topic is linked to
the current debate on the role of speculation versus fundamentals in the determination of the oil price, it
goes beyond the existing debates which have recently dominated policy agendas. This report offers a
fresh and deeper perspective on the current debate by identifying the various layers relevant to the price
formation process and by examining and analysing the links between the financial and physical layers in
the oil market, which lie at the heart of the current international oil pricing system.
The adoption of the market-related pricing system by many oil exporters in 1986-1988 opened a new
chapter in the history of oil price formation. It represented a shift from a system in which prices were first
administered by the large multinational oil companies in the 1950s and 1960s and then by OPEC for the
period 1973-1988 to a system in which prices are set by markets. First adopted by the Mexican national
oil company PEMEX in 1986, the market-related pricing system received wide acceptance among most
oil-exporting countries. By 1988, it became and still is the main method for pricing crude oil in
international trade after a short experimentation with a products-related pricing system in the shape of the
netback pricing regime in the period 1986-1987. The oil market was ready for such a transition. The end
of the concession system and the waves of nationalisation which disrupted oil supplies to multinational oil
companies established the basis of arms-length deals and exchange outside the vertically and
horizontally integrated multinational companies. The emergence of many suppliers outside OPEC and
many buyers further increased the prevalence of such arms-length deals. This led to the development of a
complex structure of interlinked oil markets which consist of spot and also physical forwards, futures,
options and other derivative markets referred to as paper markets. Technological innovations which made
electronic trading possible revolutionised these markets by allowing 24-hour trading from any place in the
6
world. It also opened access to a wider set of market participants and allowed the development of a large
number of trading instruments both on regulated exchanges and over the counter.
Physical delivery of crude oil is organised either through the spot (cash) market or through long-term
contracts. The spot market is used by transacting parties to buy and sell crude oil not covered by long
term contractual arrangements and applies often to one-off transactions. Given the logistics of
transporting oil, spot cargoes for immediate delivery are rare. Instead, there is an important element of
forwardness in spot transactions. The parties can either agree on the price at the time of agreement, in
which case the sport transaction becomes closer to a forward contract. More often though, transacting
parties link the pricing of an oil cargo to the time of loading.
Long-term contracts are negotiated bilaterally between buyers and sellers for the delivery of a series of oil
shipments over a specified period of time, usually one or two years. They specify among other things, the
volumes of crude oil to be delivered, the delivery schedule, the actions to be taken in case of default, and
above all the method that should be used in calculating the price of an oil shipment. Price agreements are
usually concluded on the method of formula pricing which links the price of a cargo in long-term
contracts to a market (spot) price. Formula pricing has become the basis of the oil pricing system.
Formula pricing has two main advantages. Crude oil is not a homogenous commodity. There are various
types of internationally traded crude oil with different qualities and characteristics which have a bearing
on refining yields. Thus, different crudes fetch different prices. Given the large variety of crude oils, the
price of a particular type is usually set at a discount or at a premium to marker or reference prices, often
referred to as benchmarks. The differentials are adjusted periodically to reflect differences in the quality
of crudes as well as the relative demand and supply of the various types of crudes. Another advantage of
formula pricing is that it increases pricing flexibility. When there is a lag between the date at which a
cargo is bought and the date of arrival at its destination, there is a price risk. Transacting parties usually
share this risk through the pricing formula. Agreements are often made for the date of pricing to occur
around the delivery date.
At the heart of formulae pricing is the identification of the price of key physical benchmarks, such as
West Texas Intermediate (WTI), Dated Brent and Dubai-Oman. The benchmark crudes are a central
feature of the oil pricing system and are used by oil companies and traders to price cargoes under longterm contracts or in spot market transactions; by futures exchanges for the settlement of their financial
contracts; by banks and companies for the settlement of derivative instruments such as swap contracts;
and by governments for taxation purposes.
Few features of these physical benchmarks stand out. Markets with relatively low volumes of production
such as WTI, Brent, and Dubai set the price for markets with higher volumes of production elsewhere in
the world. Despite the high level of volumes of production in the Gulf, these markets remain illiquid:
there is limited spot trading in these markets, no forwards or swaps (apart from Dubai), and no liquid
futures market since crude export contracts include destination and resale restrictions which limit trading
options. While the volume of production is not a sufficient condition for the emergence of a benchmark, it
is a necessary condition for a benchmarks success. As markets become thinner and thinner, the price
discovery process becomes more difficult. Oil price reporting agencies cannot observe enough genuine
arms-length deals. Furthermore, in thin markets, the danger of squeezes and distortions increases and as a
result prices could then become less informative and more volatile thereby distorting consumption and
production decisions. So far the low and continuous decline in the physical base of existing benchmarks
has been counteracted by including additional crude streams in an assessed benchmark. This had the
effect of reducing the chance of squeezes as these alternative crudes could be used for delivery against the
contract. Although such short-term solutions have been successful in alleviating the problem of squeezes,
observers should not be distracted from some key questions: What are the conditions necessary for the
emergence of successful benchmarks in the most physically liquid market? Would a shift to assessing
price in these markets improve the price discovery process? Such key questions remain heavily underresearched in the energy literature and do not feature in the consumer-producer dialogue.
The emergence of the non-OECD as the main source of growth in global oil demand will only increase
the importance of such questions. One of the most important shifts in oil market dynamics in recent years
has been the shift in oil trade flows to Asia: this may have long-term implications on pricing benchmarks.
Questions are already being raised whether Dubai still constitutes an appropriate benchmark for pricing
crude oil exports to Asia given its thin physical base or whether new benchmarks are needed to reflect
more accurately the recent shift in trade flows and the rise in prominence of the Asian consumer.
Unlike the futures market where prices are observable in real time, the reported prices of physical
benchmarks are identified or assessed prices. Assessments are needed in opaque markets such as crude
oil where physical transactions concluded between parties cannot be directly observed by outsiders.
Assessments are also needed in illiquid markets where there are not enough representative deals or where
no transactions are concluded. These assessments are carried out by oil pricing reporting agencies
(PRAs), the two most important of which are Platts and Argus. While PRAs have been an integral part of
the oil pricing system, especially since the shift to the market-related pricing system in 1986, their role
has recently been attracting considerable attention. In the G20 summit in Korea in November 2010, the
G20 leaders called for a more detailed analysis on how the oil spot market prices are assessed by oil
price reporting agencies and how this affects the transparency and functioning of oil markets. In its latest
report in November 2010, IOSCO points that the core concern with respect to price reporting agencies is
the extent to which the reported data accurately reflects the cash market in question. PRAs do not only
act as a mirror to the trade. In their attempt to identify the price that reflects accurately the market value
of an oil barrel, PRAs enter into the decision-making territory which can influence market structure.
What they choose to do is influenced by market participants and market structure while they in turn
influence the trading strategies of the various participants. New markets and contracts may emerge to
hedge the risks arising from some PRAs decisions. To evaluate the role of PRAs in the oil market, it is
important to look at three inter-related dimensions: the methodology used in indentifying the oil price; the
accuracy of price assessments; and the internal measures that PRAs implement to protect the integrity and
ensure an efficient assessment process. There is a fundamental difference in the methodology and in the
philosophy underlying the price assessment process between the various PRAs. As a result, different
agencies may produce different prices for the same benchmark. This raises the issue of which method
produces a more accurate price assessment. Given that assessed prices underlie long-term contracts, spot
transactions and derivatives instruments, even small differences in price assessments between PRAs have
important implications on exporters revenues and financial flows between parties in financial contracts.
In the last two decades or so, many financial layers (paper markets) have emerged around crude oil
benchmarks. They include the forward market (in Brent and Dubai), swaps, futures, and options. Some of
the instruments such as futures and options are traded on regulated exchanges such as ICE and CME
Group, while other instruments, such as swaps, options and forward contracts, are traded bilaterally over
the counter (OTC). Nevertheless, these financial layers are highly interlinked through the process of
arbitrage and the development of instruments that links the various layers together. Over the years, these
markets have grown in terms of size, liquidity, sophistication and have attracted a diverse set of players
both physical and financial. These markets have become central for market participants wishing to hedge
their risk and to bet on oil price movements. Equally important, these financial layers have become
central to the oil price identification process.
At the early stages of the current pricing system, linking prices to benchmarks in formulae pricing
provided producers and consumers with a sense of comfort that the price is grounded in the physical
dimension of the market. This implicitly assumes that the process of identifying the price of benchmarks
can be isolated from financial layers. However, this is far from reality. The analysis in this report shows
that the different layers of the oil market form a complex web of links, all of which play a role in the price
discovery process. The information derived from financial layers is essential for identifying the price
8
level of the benchmark. In the Brent market, the oil price in the forward market is sometimes priced as a
differential to the price of the Brent futures contract using the Exchange for Physicals (EFP) market. The
price of Dated Brent or North Sea Dated in turn is priced as a differential to the forward market through
the market of Contract for Differences (CFDs), another swaps market. Given the limited number of
physical transactions and hence the limited amount of deals that can be observed by oil reporting
agencies, the value of Dubai, the main benchmark used for pricing crude oil exports to East Asia, is often
assessed by using the value of differentials in the very liquid OTC Dubai/Brent swaps market. Thus, one
could argue that without these financial layers it would not be possible to discover or identify oil
prices in the current oil pricing system. In effect, crude oil prices are jointly or co-determined in both
layers, depending on differences in timing, location and quality of crude oil.
Since physical benchmarks constitute the pricing basis of the large majority of physical transactions,
some observers claim that derivatives instruments such as futures, forwards, options and swaps derive
their value from the price of these physical benchmarks, i.e., the prices of these physical benchmarks
drive the prices in paper markets. However, this is a gross over-simplification and does not accurately
reflect the process of crude oil price formation. The issue of whether the paper market drives the physical
or the other way around is difficult to construct theoretically and test empirically and requires further
research.
The report also calls for broadening the empirical research to include the trading strategies of physical
players. In recent years, the futures markets have attracted a wide range of financial players including
swap dealers, pension funds, hedge funds, index investors, technical traders, and high net worth
individuals. There are concerns that these financial players and their trading strategies could move the oil
price away from the true underlying fundamentals. The fact remains however that the participants in
many of the OTC markets such as forward markets and CFDs which are central to the price discovery
process are mainly physical and include entities such as refineries, oil companies, downstream
consumers, physical traders, and market makers. Financial players such as pension funds and index
investors have limited presence in many of these markets. Thus, any analysis limited to non-commercial
participants in the futures market and their role in the oil price formation process is incomplete and also
potentially misleading.
The report also makes the distinction between trade in price differentials and trade in price levels. It
shows that trades in the levels of the oil price rarely take place in the layers surrounding the physical
benchmarks. We postulate that the price level of the main crude oil benchmarks is set in the futures
markets; the financial layers such as swaps and forwards set the price differentials depending on quality,
location and timing. These differentials are then used by oil reporting agencies to identify the price level
of a physical benchmark. If the price in the futures market becomes detached from the underlying
benchmark, the differentials adjust to correct for this divergence through a web of highly interlinked and
efficient markets. Thus, our analysis reveals that the level of the crude oil price, which consumers,
producers and their governments are most concerned with, is not the most relevant feature in the current
pricing system. Instead, the identification of price differentials and the adjustments in these differentials
in the various layers underlie the basis of the current crude oil pricing system. By trading differentials,
market participants limit their exposure to the risks of time, location grade and volume. Unfortunately,
this fact has received little attention and the issue of whether price differentials between different markets
showed strong signs of adjustment in the 2008-2009 price cycle has not yet received due attention in the
empirical literature.
But this leaves us with a fundamental question: what determines the price level of a certain benchmark in
the first place? The pricing system reflects how the oil market functions: if price levels are set in the
futures market and if market participants in these markets attach more weight to future fundamentals
rather than current fundamentals and/or if market participants expect limited feedbacks from both the
9
supply and demand side in response to oil price changes, these expectations will be reflected in the
different layers and will ultimately be reflected in the assessed spot price of a certain benchmark.
The current oil pricing system has survived for almost a quarter of a century, longer than the OPEC
administered system. While some of the details have changed, such as Saudi Arabias decision to replace
Dated Brent with Brent futures in pricing its exports to Europe and the more recent move to replace WTI
with Argus Sour Crude Index (ASCI) in pricing its exports to the US, these changes are rather cosmetic.
The fundamentals of the current pricing system have remained the same since the mid 1980s: the price of
oil is set by the market with PRAs making use of various methodologies to reflect the market price in
their assessments and making use of information in the financial layers surrounding the global
benchmarks. In the light of the 2008-2009 price swings, the oil pricing system has received some
criticism reflecting the unease that some observers feel with the current system. Although alternative
pricing systems could be devised such as bringing back the administered pricing system or calling for
producers to assume a greater responsibility in the method of price formation by removing destination
restrictions on their exports, or allowing their crudes to be auctioned, the reality remains that the main
market players such as oil companies, refineries, oil exporting countries, physical traders and financial
players have no interest in rocking the boat. Market players and governments get very concerned about oil
price behaviour and its global and local impacts, but so far have showed much less interest in the pricing
system and market structure that signalled such price behaviour in the first place.
10
Although spot transactions contain an element of forwardness, they are considered as commercial agreements
under US law and are not subject to the regulation of the Commodity Exchange Act.
20
premium to a marker or reference price. These references prices are often referred to as benchmarks. The
formula used in pricing oil in long-term contracts is straightforward. Specifically, for crude oil of variety
x, the formula pricing can be written as
Px = PR D
where Px is the price of crude x; PR is the benchmark crude price; and D is the value of the price
differential. The differential is often agreed at the time when the deal is concluded and could be set by an
oil exporting country or assessed by price reporting agencies.20 It is important to note that formula pricing
may apply to all types of contractual arrangements, be they spot, forward or long term. For instance, a
spot transaction in the crude oil market, is - pricing wise - an agreement on a spot value of the differential
between the physical oil traded and the price of an agreed oil benchmark, which fixes the absolute price
level for such trade, normally around the time of delivery or the loading date.
Differences in crude oil quality are not the only determinant of crude oil price differentials however. The
movements in differentials also reflect movements in the Gross Products Worth (GPW) obtained from
refining the reference crude R and the crude x.21 Thus, price differentials between the different varieties of
crude oil are not constant and change continuously according to the relative demand and supply of the
various crudes which in turn depend on the relative prices of petroleum products. Figure 1 plots the
differential that Saudi Arabia applied to its crude exports to Asia for its different types of crude oil
relative to the Oman/Dubai benchmark during the period 2000-2010 (January). As seen from this figure,
the discounts and premiums applied are highly variable. For instance, at the beginning of 2008, the
differentials between Arab Super Light and Arab Heavy widened sharply to reach more than $15 a barrel;
fuel oil, a product of heavy crude, was in surplus while the demand for diesel, a product of lighter crudes,
was high. In the first months of 2009, the price differential between heavy and light crude oil narrowed to
very low levels as the implementation of OPEC cuts reduced the supply of heavy crude and increased the
relative value of heavy-sour crudes.
Figure 1: Price Differentials of Various Types of Saudi Arabias Crude Oil to Asia in $/Barrel
Jan-00
May-00
Sep-00
Jan-01
May-01
Sep-01
Jan-02
May-02
Sep-02
Jan-03
May-03
Sep-03
Jan-04
May-04
Sep-04
Jan-05
May-05
Sep-05
Jan-06
May-06
Sep-06
Jan-07
May-07
Sep-07
Jan-08
May-08
Sep-08
Jan-09
May-09
Sep-09
Jan-10
+10.00
+8.00
+6.00
+4.00
+2.00
0.00
-2.00
-4.00
-6.00
-8.00
-10.00
Arab Heavy-27
Arab Light-33
Official formula pricing refers to the process of setting the differential in relation to a benchmark with the
resultant price known as official formula prices. This should be distinguished from official selling prices in which
the government sets the price on an outright basis.
21
Individual crudes have a particular yield of products with a gross product worth (GPW). GPW depends both on
on the refining process and the prices at which these products are sold.
21
The differential to a benchmark is independently set by each of the oil-producing countries. For many
countries, it is usually set in the month preceding the loading month and is adjusted monthly or quarterly.
For instance, for the month of May, the differential is announced in the month before, i.e. April based on
information and data about GPW available in the month of March.22 Since the process of setting price
differentials involves long time lags and is based on old information and data, the value of the price
differential does not often reflect the market conditions at the time of loading and much less so by the
time the cargo reaches its final destination. In the case of multiple transactions under a long-term contract,
buyers can be compensated by sellers by adjusting downwards the differential in the next rounds if the
price proves to be higher than what is warranted by market conditions at the time of loading or at
delivery. This continuous process of adjusting differentials is inevitable given that setting the differential
is based on lagged data and if oil exporters wish to maintain the competitiveness of their crudes.
In other countries such as Abu Dhabi and Qatar, the governments do not announce price differentials, but
rather an outright price known as the official selling price (OSP). These are, however, strongly linked to
Dubai-Oman benchmark and thus, one can assume that outright prices contain an implicit price
differential and hence are close to formula prices (see Horsnell and Mabro, 1993; Argus, 2010).23
In setting the differential, an oil-exporting country will not only consider the differential between its crude
and the reference crude, but has also to consider how its closest competitors are pricing their crude in
relation to the reference crude. This implies that the timing of setting the differential matters, especially in
a slack market. Oil-exporting countries that announce their differentials first are at the competitive
disadvantage of being undercut by their closest competitors. This can induce them to delay announcement
of the differential or, in the case of multiple transactions, compensate the buyers by adjusting the
differential downward in the next rounds. Competition between various exporters implies that crude oils
of similar quality and destined for the same region tend to trade at very narrow differentials. Figure 2
below shows the price differential between Saudi Arabia Light (33.0 API) and Iranian Light (33.4 API)
destined to Asia. As seen from this graph, the differentials are narrow not exceeding 30 cents most of the
time although on some occasions, the differentials tend to widen. Such large differentials do not tend to
persist as adjustments are made to keep the crude oil competitive. In the mid 1990s, Saudi Arabia Light
was trading at a premium to the Iranian Light, but this premium turned into discount in the slack market
conditions of 1998. In the period 2002 to 2004, the two types of crude oil were trading almost at par, but
since 2007, Saudi Arabian Light has been trading at a discount, making its light crude more competitive
compared to the Iranian Light, perhaps in an attempt by Saudi Arabia to maximise its export volume to
Asia or due to mispricing on the part of the Iranian National Oil Company.
The above discussion focused only on the pricing mechanism implemented by an oil exporting country
via its national oil company. The value of the differential does not need not to be set by an oil producing
country and can be assessed by price reporting agencies.
22
22
Figure 2: Differentials of Term Prices between Saudi Arabia Light and Iran Light Destined to Asia
(FOB) (In US cents)
0.50
0.40
0.30
0.20
0.10
0.00
-0.10
-0.20
-0.30
-0.40
Jan 95
Jun 95
Nov 95
Apr 96
Sep 96
Feb 97
Jul 97
Dec 97
May 98
Oct 98
Mar 99
Aug 99
Jan 00
Jun 00
Nov 00
Apr 01
Sep 01
Feb 02
Jul 02
Dec 02
May 03
Oct 03
Mar 04
Aug 04
Jan 05
Jun 05
Nov 05
Apr 06
Sep 06
Feb 07
Jul 07
Dec 07
May 08
Oct 08
Mar 09
Aug 09
Jan 10
-0.50
The equivalence to the buyer principle, which means that in practice prices of crudes have equivalent
prices at destination, adds another dimension to the pricing formulae. The location in which prices should
be compared is not the point of origin but must be closer to the destination where the buyer receives the
cargo. Since the freight costs vary depending on the export destination, some formulae also take into
account the relative freight costs between destinations. Specifically, they allow for the difference between
the freight costs involved in moving the reference crude from its location to a certain destination (e.g.
Brent from Sollum Voe to Rotterdam) and the costs involved in moving crude x from the oil countrys
terminal to that certain destination (e.g. Arabian Light from Ras Tanura to Rotterdam). In such cases, the
sale contract is close to a cost, insurance and freight (CIF) contract. This is in contrast to a free on board
(FOB) contract which refers to a situation in which the seller fulfils his obligations to deliver when the
goods have passed over the ships rail. The buyer bears all the risks of loss of or damage to the goods
from that point as well as all other costs such as freight and insurance.
A major advantage of formula pricing is that the price of an oil shipment can be linked to the price at the
time of delivery which reflects the market conditions prevailing. When there is a lag between the date at
which a cargo is bought and the date of arrival at its destination, there is a big price risk. Transacting
parties usually share this risk through the pricing formula. Agreements are often made for the date of
pricing to occur around the delivery date. For instance, in the case of Saudi Arabias exports to the United
States up to December 2009, the date of pricing varied between 40 to 50 days after the loading date. The
price used in contracts could be linked to the price of benchmark averaged over 10 days around the
delivery date, which rendered the point of sale closer to destination than the origin. In 2010, Saudi Arabia
shifted to Argus Sour Crude Index (ASCI) and it currently uses the trade month (20 day minimum)
average of ASCI prices for the trade month applying to the time of delivery.
Oil exporters may have different pricing policies for different regions. For instance, for Saudi exports to
the US, the price that matters most is the cost of shipment at the delivery point. For its exports to Asia, the
pricing point is free on board and hence the price that matters most is the price at the loading terminal.
Figure 3 below shows the price difference between crude delivered to the US Gulf Coast and the price
sold at FOB to Asia for different variety of crude oils. As seen from this graph, the price differential is
highly variable depending on the relative demand and supply conditions between these two markets and
the degree of competition from alternative sources of supply. While in the US, Saudi Arabia faces tough
competition from many suppliers including domestic ones and hence its crude has to be competitive at
23
destination, the strong growth in Asian demand and the limited degree of competition in Asia give rise to
an Asian premium. Hence, in some occasions the price of a cargo delivered to the US is less than the
FOB price to Asia despite the fact that it takes longer for a cargo to reach the US.
Figure 3: Difference in Term Prices for Various Crude Oil Grades to the US Gulf (Delivered) and
Asia (FOB)
15.00
10.00
5.00
0.00
-5.00
Saudi Arabia
Light-33.0
Saudi Arabia
Arab Medium-3.5
Jan 10
Sep 09
Jan 09
May 09
Sep 08
May 08
Jan 08
Sep 07
Jan 07
May 07
Sep 06
Jan 06
May 06
Sep 05
Jan 05
May 05
Sep 04
Jan 04
May 04
Sep 03
May 03
Jan 03
Sep 02
Jan 02
May 02
Sep 01
Jan 01
May 01
Sep 00
Jan 00
May 00
-10.00
Saudi Arabia
Arab Heavy-27.6
24
Platts continues to call the physical market Dated Brent or Dated North Sea Light while Argus calls it North Sea
Dated. As shall be discussed later, the continued use of the term Dated Brent by Platts and much of the industry is
not an arcane point, because the price of physical Dated Brent cargoes will be different from its Dated Brent price.
The prices of physical Brent, Forties and Oseberg all differ from the (Argus) North Sea Dated/(Platts) Dated Brent
value.
25
Some governments (Oman, Qatar, Abu Dhabi, Malaysia, and Indonesia) do not use benchmarks at all and instead
set their own official selling prices (OSPs) on a monthly basis. These can be set retroactively or retrospectively.
24
Europe
Saudi Arabia
Iran
Kuwait
Dated Brent
ASCI from
December 2009 ;
Previously WTI
ASCI from April
2010, Previously
WTI Second
Month
Dated Brent
Brent
Iraq (Basrah
Blend)
Nigeria
Mexico (Maya
Blend)
US
ASCI from
Jan.2010, WTI
until Dec.'09
WTS x0.4
+ 3%HSFO x0.4
+ LLS x0.1
+ Dtd.Brent x0.1
The pricing may be based on physical benchmarks such as Dated Brent or on the financial layers
surrounding these physical benchmarks such as the Brent Weighted Average (BWAVE), which is an
index calculated on the basis of prices obtained in the Brent futures market. Specifically, the BWAVE is
the weighted average of all futures price quotations that arise for a given contract of the futures exchange
during a trading day, with the weights being the shares of the relevant volume of transactions on that day.
Major oil exporters such as Saudi Arabia, Kuwait and Iran use BWAVE as the basis of pricing crude
exports to Europe. As seen from Figure 4 below, the price differential between Dated Brent and BWAVE
is quite variable with the differential in some occasions exceeding plus or minus three dollars per barrel.
This is expected as BWAVE is considerably less prompt than Dated Brent and thus variability between
the two should consider this time basis issue.26 Therefore, the choice of benchmark has serious
implications on government revenues. This is perhaps most illustrated in the recent shift from WTI to
ASCI by some Gulf exporters. Figure 5 plots the price differential between the two US benchmarks WTI
and ASCI. WTI traded at a premium to ASCI through most of this time but occasionally (four significant
times) WTI moved to a discount when WTI collapsed versus other world benchmarks, with the WTI
discount to ASCI reaching close to $8/barrel on 12 February 2009. The January/ February events
prompted Saudi Arabia to consider alternatives to Platts WTI cash assessment.
26
Furthermore, as volatility is strongly backwardated itself along its own forward curve for most markets, this is
also a relevant factor.
25
-4.00
-5.00
Figure 5: Price Differential between WTI and ASCI ($/Barrel) (ASCI Price=0)
8.00
6.00
4.00
2.00
0.00
-2.00
-4.00
-6.00
-8.00
-10.00
ASCI
WTI vs ASCI
Source: Argus
Given the central role that benchmarking plays in the current oil pricing system, it is important to
highlight some of the main features of the most widely used benchmarks. First, unlike the futures market
where prices are observable in real time, the reported prices of physical benchmarks are identified or
assessed prices. These assessments are carried out by oil pricing reporting agencies, the two most
important of which are Platts and Argus.27 Assessments are needed in opaque markets such as oil where
physical transactions concluded between parties cannot be directly observed by market participants. After
all, parties are under no obligation to report their deals. Assessments are also needed in illiquid markets
27
There are other PRAs but these are often more specialised such as OMR (focus on Germany, Austria, and
Switzerland), APPI (focus on Asia), RIM (focus on Asia), ICIS-LOR (focus on petrochemicals) and OPIS (focus on
US). In December 2010, Platts announced an agreement to acquire OPIS. The acquisition is expected to be
completed in the first half of 2011, subject to regulatory approval.
26
where not enough representative deals or where no transactions take place. Oil reporting agencies assess
their prices based on information on concluded deals which they observe, or bids and offers, and failing
that on market talk, other private and public information gathered by reporters, and information from
financial markets. It is important to note that PRAs do not use in all markets a hierarchy of information
cascading down from deals to bids and offers, which would imply that deals are the best price discovery
and bids/offers are a poorer alternative. The methodology may vary from market to market in accordance
with the published methodology for that market. In some markets, bid/offer information takes precedence
over deals in identifying the published price e.g. if the deal is either not representative of the market as
defined in the methodology, or was done earlier or later in the day to the prevailing depth of market. In
other markets, price identification relies on observed deals. For instance, Argus main benchmark ASCI is
entirely deal based. Most however accept that a done deal does represent the highest form of proof of
value, unless there is a supervening issue with the trades conduct. If assessments are intended to
represent an end-of day price, analogous to a futures settlement however, a fully evidenced bid/offer
spread at a later point when markets have clearly moved in value is an acceptable proxy in the absence of
a trade .
Sometimes a distinction is made between prices identified through observed deals or transactions using a
direct mathematical formula such as volume-weighted average (referred to as an index) and prices
identified through a process of interpretation based on bids and offers, market surveys, and other
information gathered by reporters (referred to as price assessment) (see Argus, 2010). The choice of the
method varies across markets and depends on the structure of market, particularly on the degree of market
opaqueness and liquidity. While an index is suitable for markets with high trading liquidity and
transparency, assessments are more suitable in opaque and illiquid markets. In this paper, we do not make
this distinction and refer to both categories as price assessment. However, regardless of the method used,
there is an important element of subjectivity involved as the methodology has to be decided by managers
and editors. The choice of methodology (the time window in which the price is assessed, the grade
specification, location) in an index based system is just as subjective as price assessment. In that respect,
one approach (index or assessment) is no more subjective than the other.
Second, these agencies do not always produce the same price for the same benchmark as these pursue
different methodologies in their price assessments. Even if price quotations are based on a mechanical
methodology of deals done, two price reporting services could publish different prices for the same crude
because their price identification process and the deals they include in the assessment could be different.
For example, one PRA might use a volume weighted average of transactions between 9.00am and 5.00pm
while another PRA might use last trade or open bid/offer at specified period of time. Or one PRA might
include transactions within a 10-21 day price range and another includes transactions in a 10-15 day price
range. Or one PRA might only include fixed-price transactions and another include fixed-price and
formula-related transactions.
Third, the nature of these benchmarks tends to evolve over time. Although the general principle of
benchmarking has remained more or less the same over the last twenty-five years, the details of these
benchmarks in terms of their liquidity and the type of crudes that are included in the assessment process
have changed dramatically over that period. The assessment of the traditional Brent benchmark now
includes the North Sea streams Forties, Oseberg and Ekofisk (BFOE) and that of Platts Dubai price
includes Oman and Upper Zakum. These streams are not of identical quality and often fetch different
prices. Thus, the assessed price of a benchmark does not always refer to a particular physical crude
stream. It rather refers to a constructed index28 which is derived on the basis of a simple mathematical
formula which takes the lowest priced grade of the different component crudes to set the benchmark.
28
This may take the form of a matrix of closely-related prices which use the total physical liquidity by engineering
price floors and caps to reduce or eliminate the possibility of price distortion or skews.
27
Table 2 below summarises some basic statistics of the main international benchmarks: BFOE in the North
Sea, WTI and ASCI in the US, and Dubai-Oman in the Gulf. In terms of production, the underlying
physical base of the benchmark amounts to slightly more than 3 million b/d, i.e., around 3.5% of global
production. In terms of liquidity, there is wide difference across benchmarks. While in the US the number
of spot trades per calendar month is close to 600, the number of spot trades does not exceed three per
month in the case of Dubai. The divergence in liquidity across benchmarks reflects the low underlying
physical base and the different nature of benchmarks where US crudes are pipeline crudes with small
trading lots whereas Brent and Dubai are waterborne crudes with large trading lots. Table 2 also shows
that the degree of concentration in traded volumes varies considerably across markets. From the sellers
side, Dubai, Oman and Forties exhibit a high degree of concentration in the total volume of spot trades
especially when compared to US markets. From the buyers side, Dubai and Forties exhibit a high degree
of concentration whereas Oman compares favourably with other benchmarks.
Table 2: Some Basic Features of Benchmark Crudes
First-quarter 2010 averages
by Argus
ASCI
WTI CMA
+ WTI PPlus
Forties
BFOE
Dubai
Oman
Production (MBPD)
736
300-400
562
1,220
70-80
710
579
939
514
635
86
246
260
330
18
98
3.5
10
13
16
<1
<1
<1
26
27
10
24
36
43%
38%
63%
72%
100%
50%
38%
51%
76%
56%
100%
80%
Source: Argus
Notes: Daily statistics are per trade day, except production which is per calendar day; Forties: The physical grade
usually sets North Sea Dated/Dated Brent; BFOE: Forward cash contracts deliverable as physical BFOE cargoes,
used in setting the flat price against which North Sea Dated is calculated; Oman: Excludes physical deliveries
through DME. Estimated deliveries on DME contacts are 300,000-400,000 barrels per day; WTI: Includes cash
market trade for WTI Calendar Month Average and WTI P-Plus. Cash market at Cushing no longer trades except at
last three days of trade month as spread for 2 nd month. Roll trades are not included here. Also does not include any
volumes on CME Nymex futures.
Finally, in the last two decades or so, many financial layers (paper markets) have emerged around these
benchmarks. These include the forward market (in Brent), swaps, futures, and options. Some of the
28
instruments such as futures and options are traded on regulated exchanges such as ICE and CME Group,
while other instruments, such as swaps and forward contracts, are traded bilaterally over the counter
(OTC). Nevertheless, these financial layers are highly interlinked through the process of arbitrage and the
development of instruments that link the various markets together such as the Exchange of Futures for
Swaps (EFS) which allow traders to roll positions from futures to swaps and vice versa. Over the years,
these markets have grown in terms of size, liquidity, sophistication and have attracted a diverse set of
players, both physical and financial. These markets have become central for market participants wishing
to hedge their risk and to bet (or speculate) on oil price movements. Equally important, these financial
layers have become central to the oil price identification process. In Sections 5, 6 and 7, we discuss the
main benchmarks used in the current oil pricing system and the financial layers surrounding these
benchmarks.
29
Argus Response to the Report of the Working Group on the Volatility of Oil Prices chaired by Professor JeanMarie Chevalier, p.5.
30
PRAs assessment were already widely used in the price formation process for refined products prior to 1986.
31
G-20 Seoul Summit 2010, THE SEOUL SUMMIT DOCUMENT, Paragraph 61.
32
IOSCO (2010), Task Force on Commodity Futures Markets: Report to the G20, November 2010, p. 17.
33
Though other attributes such as representativeness and usefulness could also be included.
30
because their mechanical price identification process could be different. This raises the issue of which of
the methods generates a more accurate price assessment. Given that assessed prices underlie long-term
contracts, spot transactions and derivatives instruments, even small differences in price assessments
between PRAs have serious implications for exporters revenues and financial flows between parties in
financial contracts.
PRAs use a wide variety of methods to identify the oil price which may include the volume weighted
average system, low and high deals done, and market-on-close (MOC). In January 2001, Platts stopped
using the volume-weighted average system and replaced it with the MOC methodology.34 In this system,
Platts sets a time window, known as the Platts window, and only deals transacted within this time window
are used to assess the oil price.35 The price is assessed on the basis of concluded deals, or failing that, on
bids and offers. Assessment will also make use of information from financial layers about spreads and
derivative to help triangulate value.36 Thus, the MOC can be thought of a structured system for
gathering information on the basis of which Platts assesses the daily price of key physical benchmarks. In
a way, it is similar to a futures exchange where traders make bids and offers, but with two major
differences: the parties behind the bids and offers are known, and Platts decides on the information to be
considered in the assessment, i.e., the information passes through the Platts filter. These price assessments
are then transmitted back to the market through a variety of channels. The reason for the shift to MOC is a
concern that an averaging system for price determination could result in assessments that lag actual
market levels as deals done early in an assessment period at a level that is not repeatable, could
mathematically drag prices down or up (Platts, 2010a:7).37 Thus, Platts emphasises the time sensitivity of
its assessed prices which are clearly time-stamped on a daily basis.38 Time stamping not only allows for
an accurate reflection of price levels at particular point in time, but also for accurate assessment of time
spreads and inter-crude spreads.
Both the volume-weighted average method and the MOC have received their share of criticism. While the
volume-weighted average method allows the inclusion of a large number of deals and hence is more
representative, the method has been criticised as it
34
In the US, Platts used a volume weighted average for domestic crude. But for products, it has always used a low
and high of deals done. In the WTI crude market prior to 2001 Platts used a volume weighted average of a 30minute window. In Asia, Platts used the window or page 190, its first market on close, also before 2001. The
market on close went global for Platts in 2001.
35
It is important to note that the window opens all day and Platts will accept trades, bids and offers at any time of
the day. But only deals transacted within a specified period of time (for instance from 4:00 to 4:30 for European
crudes) are considered for assessing the price for that day. Some argue that this may encourage traders to present
their bids/offers to Platts during this time window in order to maximize their impact on prices.
36
Thorne, S. (2010), A User guide to Platts Assessment Processes, Presentation at the Platts Crude Oil
Methodology Forum 2010, London, May.
37
Platts (2010a), Methodology and Specifications Guide: Crude Oil, The McGraw Hill Companies, October.
38
Some commentators consider that through its window, Platts is able to establish the marginal price of oil, which in
principle should set price for the rest of the market. It is not clear what is meant by the marginal price, but in terms
of theory, the closest one can think of the Platts window is in terms of the Walrasian auctioneer. The Walrasian
auctioneer is a fictitious construct who aggregates traders demand and supplies to find a market clearing price,
through a series of auctions. While Platts window resembles the Walrasian auctioneer, it differs fundamentally in
many respects such as the existence of transactions costs, barriers to entry and the fact that the auctioneer does not
perform a passive role in the market. It decides who enters the market and when to the set the price. It has also been
long realised that trading has a timing dimension. While over time, the number of buyers and sellers may be equal,
at any particular the time, this is not guaranteed in which case it is not possible to find a market clearing price
(Demsetz, 1968). This could be overcome by participants paying an immediacy premium in which case the
equilibrium will be characterised by two demand and supply curves and two prices. Furthermore, the literature
shows that market structure such as the number of players, their size, the timing of entry matters and could affect the
trading price. Therefore, the actual mechanism used to set the price is not simply a channel, but is an input into the
price and as such cannot be ignored (see O'Hara, 1997).
31
may result in an index that is out of step and not reflective of the actual market price prevailing at the
close of the day. This would especially be the case on days with high volatility. Trade- weighted
averages may also be distorted by the pattern of trading liquidity over the day. A key weakness in
all trade-weighted average assessments is that they will lag the market price. They always reflect a
price that was rather than the price that is. (Platts, 2010b:6).39
The main criticism of the MOC methodology is that the Platts window often lacks sufficient liquidity and
may be dominated by few players which may hamper the price discovery process. For instance, Argus,
Platts main competitor, argues that in US crude markets
MOC methodology would work if the industry poured liquidity into the window. Without this
liquidity, the methodology is left to assess the value at the close based on bids, offers and other
related factors. This means that the price derived from an MOC assessment can diverge widely
from a weighted average of all deals done in the trading day.40
This divergence is expected given that the average price is different from the stamped price and the
convergence of the two is just a statistical accident if it ever happens.
Argus conducted a study on the US crude oil market in 2007 which compares the spot market traded
volume inside the window with the volume traded during the entire day. The study finds that the volume
traded within the Platts window constitutes only a very small fraction of daily traded volumes, as seen in
Table 3 below. This applies to a wide variety of US crudes. Argus argues that such low liquidity and
complete lack of participant breadth raise serious questions about the efficiency of price discovery in
the US oil market.41
Table 3: Spot Market Traded Volumes in May 2007 (May traded during May Trade Month)
Window
LLS
Window % of Total
446,920
0%
26,425
378,445
7%
Mars
5,418
185,252
3%
WTS
1,000
154,706
1%
WTI Midland
3,000
138,470
2%
HLS
1,000
100,032
1%
WTI P-Plus
1,000
88,802
1%
Eugene Island
40,044
0%
Poseidon
73,857
0%
SGC
22,100
0%
0
37,843
9,140
1,637,768
0%
2.31%
Bonito
Source: Argus (2007)
One response to such a criticism is that if some market participants think that prices in the window are not
reflecting accurately the price of an oil barrel at the margin, then those participants should enter the
39
Platts (2010 b), Platts Oil Pricing and MOC Methodology Explained, The McGraw Hill Companies, June.
Argus Global Markets (2007), Liquidity and Diversity Prevail, 24 September, p. 15.
41
There are other markets, such as Asian products which would show in contrast very high % figures for Platts
window trades. Ultimately market participants decide upon which and whose pricing system and by implication,
methodology, they wish to use. However, once a critical mass of players is using one in a market or series of
markets, it is difficult and expensive to make a switch.
40
32
window and exert their influence on the price. However, in some markets, there might be barriers to entry
preventing such an adjustment mechanism from taking place. For instance, in the context of Dubai, Binks
(2005) argues that participation (in the window) requires knowledgeable and experienced trading staff.
And many of the national oil companies that represent end-users in Asia are not allowed to participate in
speculative trading. For the same reason, Middle East producers will not participate in the partials market.
Even independent commercial buyers without these restraints in Asia feel reluctant to participate in the
partials trade out of concern that doing so could threaten their relations with Middle Eastern producers.42
It is important to note that while some barriers such as having experienced and professional staff and
qualified companies with the necessary logistics to execute physical trades can be considered as natural
barriers, others barriers arise due to policy and strategic choices which limit the trading activity in the
window to a small group of what so called professionals.43
Market participants are under no legal or regulatory obligation to report their deals to PRAs or any other
body for that matter. Whether participants decide to share information depend on their willingness, their
reporting policies, and their interest in doing so. In the US, the system is voluntary, but one potential
interpretation of the Sarbanes-Oxley legislation is that companies must report all or nothing, and cannot
selectively disclose information.44 Many companies have reporting policies that only bind them to report
deals that take place at a certain time of day, or in certain regional markets. In some markets such as the
US, confidentiality concerns dictate that some PRAs do not publish the names of the counterparties to a
deal. To ensure enough reporting takes place, PRAs such as Argus sign confidentiality agreements to
facilitate deal reporting in the US though companies may have the incentive to report prices without such
agreements. Since market participants have different interests and different positions, some traders may
have the incentive to manipulate prices by feeding false information to reporters though there have been
regulatory efforts to limit such behaviour. In the US, the Commodity Futures Trading Commission
(CFTC)45, the Federal Energy Regulatory Commission (FERC), and the Federal Trade Commission
(FTC)46 have passed regulations that prohibit false reporting. In the EU, the Market Abuse Directive is
42
Some interviewees also pointed to the high subscription cost involved in the entry of E-window, by which Platts is
assessing larger number of markets.
43
One interviewee considers this aspect as necessary otherwise enlarging the base of participants may create
logistical and serious performance issues, including safety issues.
44
Initially a law/regulation was passed in 2000 by the SECURITIES AND EXCHANGE COMMISSION (SEC)
known as Regulation FD (Fair Disclosure). This came out of and expands upon the Insider Trading law framework
and pertains to equities reporting. Sarbanes-Oxley Act expanded on this regulation. The Act deals with voluntary
reporting areas. The obligation is stated that should you volunteer to report information, the obligation is to report
that information fully. However, companies are not required to report trades to the PRAs. Lobo and Zhou (2006)
investigated the change in managerial discretion over financial reporting following the Sarbanes-Oxley Act and find
an increase in conservatism in financial reporting.
45
On November 3, 2010, the CFTC and the Securities and Exchange Commission (SEC) proposed rules under the
new anti-manipulation and anti-fraud provisions of the Dodd-Frank Wall Street Reform and Consumer Protection
Act. One of the proposed rules states that, It shall be unlawful for any person, directly or indirectly, in connection
with any swap, or contract of sale of any commodity in interstate commerce to intentionally or recklessly:..
make, or attempt to make, any untrue or misleading statement of a material fact or to omit to state a material fact
necessary in order to make the statements made not untrue or misleading. deliver or cause to be delivered.by
any means of communication whatsoever, a false or misleading or inaccurate report concerning crop or market
information or conditions that affect or tend to affect the price of any commodity in interstate commerce, knowing,
or acting in reckless disregard of the fact that such report is false, misleading or inaccurate. Source:
http://www.cftc.gov/ucm/groups/public/@lrfederalregister/documents/file/2010-27541a.pdf
46
The Energy Independence and Security Act (Energy Act) signed into law December 19, 2007, gives the Federal
Trade Commission (FTC) new authority to police market manipulation and false reporting in the petroleum
industry joining the Federal Energy Regulatory Commission (FERC) and the U.S. Commodity Futures Trading
Commission (CFTC) in this role. In section 812, the FTC is given the authority to act against false reporting in the
petroleum industry. FTCs authority however is limited to the false reporting of wholesale transactions and those to
33
also meant to perform a similar role, though its impact on price reporting is not yet clear. As discussed
above, Platts relies on a more structured system for gathering information. However traders can undertake
some anomalous deals in the Platts window by accepting high offers or underselling by delivering into
low bids in an attempt to influence the assessed price. The losses made by such transactions can be more
than compensated by entering into other contracts such as swaps. Thus, PRAs must ensure that the
information received is correct and accurate and that deals done in the window are genuine, otherwise the
whole price discovery process will be undermined. For instance, Platts will not knowingly publish any bid
or offer that is not within the market range. In addition, when offers are lifted or bids are hit, there is a
secondary process to ensure that there is no gapping and if such gapping is detected to ensure that price
assessment process is not affected by it. There are also other mechanisms to avoid the influence of nonrepeatable deals.
In a liquid market, false reporting can be less of a problem as reporters could observe concluded deals and
confirm the information they obtain from both parties. At the same time, reporters will make use of the
regular flow of information originating from the futures and OTC markets. In contrast, in illiquid markets,
a small number of reported deals or a few bids and offers can heavily influence the price assessment
process. In days when reporters cannot observe active buyers, sellers or transactions to determine the
price or simply when such deals do not exist, 47 PRAs rely on a variety of sources of information sources
or market talk to make intelligent assessments.48 In such circumstances, the reporter will look at bids
and offers from other markets, draw comparisons with similar crudes but with higher trading activity,
analyse forward curves, survey market participants opinions, and assess spread across markets to reach a
price assessment. In fact, in some instances, as in illiquid markets, the price assessment could be more
accurate in the absence of transactions, if these transactions were intended to manipulate the oil price.
In some instances, a PRA can retrospectively correct previously unidentified assessment errors. There are
some instances in which traders may dispute the assessed price reached by a PRA. There is no evidence to
suggest that this problem is widespread, but from time to time these disputes filter into media reports. For
instance, in 29 April 2010, Platts assessed the value of the June and July cash BFOE spread at minus
$0.68 a barrel. Some brokers in the market claimed that Platts assessment of the differential is inaccurate.
Based on information from the futures market and the EFP, these brokers claimed that the value of the
differential should have been minus $0.94 a barrel.49 Regardless of which value is more accurate, what is
important to note that if such disputes over price assessments ever arise there is no supervisory or
regulatory authority which would look into these claims and counter-claims.
In order to safeguard the price assessment process, PRA seek to verify the accuracy of the information
they receive and when they are unable to do so they retain the right to exclude data and information. In
this way, they guard against false data distorting their assessments. They also undertake many procedures,
both within their own organisations as well as in relation to outside participants. For instance, Platts has
control on the parties that can participate in the window. The companies behind every bid and offer must
be clearly identified with a track record of operational and financial performance and be recognisable in
the market. Trading is closely monitored and those participants that fail to meet editorial standards and/or
a Federal Department or agency. It remain unclear if the Energy Act encompasses the reporting of false or
misleading information publicly into the market or to private organizations or PRAs.
47
It should be noted that when this is the case, companies who sign contracts linked to PRA prices tend not to use
pricing centres that are illiquid. They know that no matter how well the PRA does their job the price may be volatile
or unresponsive. In many cases, the PRA chooses not to assess a crude or product because the market is too illiquid,
or there are insufficient parameters available to make an assessment based on correlative data points.
48
Intelligent assessment refers to the process of assessing prices in illiquid markets where transactions are not
observable to reporters.
49
Paddy Gourlay Dated Brent Assessment Sparks Calls For Methodology Change, Dow Jones Newswires, 30
April 2010
34
make spurious offers and bids are expelled from the window.50 Concluded transactions between parties
are sometimes subject to verification by the various price reporting agencies; spurious deals are excluded
from the assessment process. PRAs may request documentation for concluded deals such as contract
documentation or other supporting materials such as loading and inspection documents.51
Another important dimension is compliance procedures within PRAs. The accuracy of the price
assessment will depend primarily on the policies, procedures and training put in place by the PRA. Such
procedures are needed to ensure both internal and external independence and to ascertain that reporters
are following the same rules, reporting procedures and methodology as set out by the RPA. All the
regulations and compliance procedures are designed and enforced internally without being subject to
governments regulations or supervisory oversight. However, in theory, the incentive to self-regulate is
very strong. Any reputational damage due to error of design, fraud, use of insider information, or a market
perception that PRAs are herded by one party would imply a loss of confidence and would eventually
lead to their demise. If PRAs produce regularly inaccurate prices, they will cease to exist because their
subscribers will shift to another service.52
50
Nevertheless, concerns still arise that such procedures will not stop companies from using the Platts window as a
way of executing a wash trade, or trading only to set the index on index-related deals done earlier in the day. Platts
cannot track every deal down to the contract level and ask for documentary bona fides.
51
It is highly unlikely however that a PRA requesting this information would always receive it, and certainly not in
a timely enough manner to have any impact on price assessments on a given day.
52
One anonymous interviewee noted that in theory this may be true in a competitive environment but not in the case
of oil PRAs where the market is characterised by almost a duopoly.
35
53
54
See Argus (2010), Argus Guide to Crude and Oil Products Markets, January.
BP Website: https://www.icmmed0ty.com/fps/content/brochure/brochure.asp?sectionid=1
36
31
1
1
23
38
53
2
3
2
10
35
5
1
AGIP (ENI)
Amerada HESS
BGGROUP
BPPLC
CHALLENGER
CHEVRONTEXACO
CNR
DANA
DYAS
EXXONMOBIL
ITOCHU
LUNDIN
MARUBENI
MITSUBISHI
PALACE E&P
SHELL
STATOIL ASA
TFE
Source: Bossley, L. (2007), Brent: A Users Guide to the Future of the World Price Marker, London: CEAG, p.83.
Brent System
Ninian System
Total Blend
1986
885
346
885
1991
450
324
773
Notes:
(a) January 1 to July 31 1990 before comingling
(b) August 1 to December 31 1990 after co-mingling
Source: Horsnell and Mabro (2003)
55
See for Instance, Liz Bossley (2003), Battling Benchmark Distortions, Petroleum Economist, April.
37
1992
547
309
856
In July 2002, Platts broadened its definition of the benchmark Dated Brent to include Forties (UK North
Sea) and Oseberg (Norway) for assessment purposes and as deliverable grades in the Brent Forward
contract. Forties is a mixture of oil produced from separate fields and collected by pipeline to the terminal
in Hound Point in the UK. Oseberg is a mixture of oil produced from various Norwegian fields and
collected to the Sture terminal in Norway. The new benchmark was known as Brent-Forties-Oseberg
(BFO). The inclusion of these two grades increased the production volume of the benchmark. It also
resulted in the distribution of cargoes over a wider range of companies with none having a dominant
position. However, as seen from the graph below, the production of BFO started its decline, falling from
63 cargoes a month in August 2004 to around 48 cargoes in the first months of 2007. In early 2007, BFO
production amounted to less than 30 million barrels a month, distributed over more than 55 companies.
Figure 7: Falling output of BFO
Source: Joel Hanley, Assessing the Benchmarks, Platts Presentation, January 31, 2008.
In 2007, a new grade, Ekofisk, was added to the complex which led to the creation of the current
benchmark known as BFOE, though it is still commonly referred to as Brent or North Sea. Ekofisk is a
mixture of crude oil produced from different North Sea fields and is transported to the Teesside terminal
in the UK. The bulk of BFOE output is traded on the spot market or transferred within integrated oil
companies where only about one out of seven BFOE cargoes is sold on long-term basis.56 This feature
combined with the highly diversified ownership gave rise to an active trading activity around BFOE. The
inclusion of this new stream increased the physical base of the benchmark to around 45 million barrels a
month in early 2007 but since then it has been in gradual decline. Production of BFOE is expected to
decline to less than 1 million b/d by 2012. As noted by Platts (2010a:3), further changes to the
benchmarks cant be ruled out, especially if production of the key grades is deemed too low or if their
qualities were to deviate significantly from the norm. In fact such a change might occur sooner rather
56
Argus (2010), Argus Guide to Crude and Oil Products Markers, January.
38
than later. A recent article warns that unless the contract is enlarged, it faces the risk of serial squeezes
and distortions.57
Given that these various grades are not of similar quality as shown in Table 5 below, the widening of the
definition of the North Sea benchmarks has implications on the price assessment process. In particular,
the start-up of the Buzzard field in 2007 increased the viscosity and the sulfur content of Forties Blend
making Forties the least valuable among the various crudes in the BFOE benchmark. Since any of the
four varieties can be delivered against a BFOE contract, sellers often tend to deliver the cheapest grade
and hence it is Forties that sets the price for the BFOE benchmark.58 This problem becomes more acute
during periods when other fields in the Forties system are shut down for maintenance. As a result of
including the Buzzard stream, Platts had to introduce a quality de-escalator in July 2007 which applies
for deliveries above the base standard of 0.60% sulfur: the higher the sulfur content, the bigger the
discount that the seller should give. Currently, a de-escalator of 60 cents/barrel applies for every 0.10 per
cent of sulfur specified above the base standard. Prior to this innovation, the market was not sure on
how to deal with the sulfur issue and in some periods in 2007 there were no trades in the Platts window.59
This episode almost brought the physical market to a standstill with traders complaining that Platts
changes to its pricing assessment process had paralysed the market.60
Table 5: API and Sulfur Content of BFOE Crudes
Forties Before Buzzard
Buzzard
API
44.1
32.6
Sulfur Content 0.19
1.44
wt.
Brent
Oseberg
Ekofisk
38.1
0.42
37.7
0.23
37.5
0.23
Source: Bossley, L. (2007), Brent: A Users guide to the Future of the World Price Marker, London: CEAG, Table
5.
Kemp, J. (2011), Falling Output Imperils Brent Benchmark, Reuters, 19 January 2011.
For instance on May 25 2010, Forties was assessed at $67.57-67.59, Oseberg at $68.49-68.52, Ekofisk at $68.2968.32, and Brent at $68.02-68.05 by Platts. The BFOE or North Sea Light was assessed at 67.57-67.59, the same as
the assessment of the value of Forties.
59
FT.com/Alphaville (2010), Brents Got Its Problems Too, September 2010.
60
Reuters (1997), Platts to modify new oil price system after turmoil, 19 June.
58
39
pricing reporting agencies. This has a number of important implications. First, since there is OTC trade
that goes unreported, the volume of market activity reported by Argus is likely to be a fraction of the total
volume of trade conducted in the various Brent layers. Nevertheless, it is representative of the market
activity and hence any proportions based on this sample such as the relative sizes of OTC markets and
the shares held by different companies are likely to represent fairly accurately the structure of the market.
Second, when analysing trends over a period of time, changes in statistics related to liquidity or to the
number of reported deals may reflect changes in coverage by the price reporting agency rather than
underlying changes in the statistic. Third, other problems arise when making comparisons across the
different Brent layers. For instance, in the futures markets, every deal is reported and the size of the
contract is 1000 barrels. In some layers such as Dated Brent and 21 Day BFOE, players can end up with a
ship full of crude which limits the attractiveness of these markets to a large number of participants.
Hence, one should be careful when comparing across markets as although these are all part of the Brent
complex, they differ in nature and function. Furthermore, the nature of trading can be different across
markets. For instance, in Dated Brent and 21 Day BFOE, trade in outright differentials or spreads is the
norm though 21 Day BFOE can also trade on a fixed price basis. In the futures and options, trade in
differentials also constitutes an important component of trade between months. This involves buying a
contract in one month (say a June contract) and selling a contract in another month (say a July contract).
In terms of reporting, each of the two legs of the transaction is reported as an outright deal. Thus, any
comparisons across markets should adjust for the volume of such trade in spreads.
The Forward Brent
The Forward Brent is one of the first layers to emerge in the Brent complex. The forward Brent is also
referred to as 21-day Brent, 21-day BFOE or simply as paper Brent. Forward Brent is a forward contract
that specifies the delivery month but not the particular date at which the cargo will be loaded. Forward
Brent price is often quoted for three months ahead. For instance, on 25th May, the Forward Brent is
reported for the months of June, July and August. These price quotations represent the value of a cargo of
physical delivery in the month specified by the contract.
In order to understand the nature of the Forward Brent market, it is important to look at the precursor of
the 21-day Brent, the 15-day Brent market. The incentive for oil companies to engage in tax spinning
through the forward market was the main factor responsible for the emergence of the forward 15-Day
Brent market (Mabro et al. 1986; Horsnell and Mabro, 1993; Bacon, 1986). The valuation of oil for UK
fiscal purposes was based on market prices. In an arms-length transaction, market prices were obtained
from the realised prices on the deal.61 If oil was merely transferred within a vertically integrated system,
then the fiscal authorities would assign an assessed price to the transaction based on the prices of
contemporary and comparable arms-length deals. Until 1984, these followed the official British
National Oil Corporation (BNOC) price. Because of the differential rates of taxation between upstream
and downstream with the tax rate being lower in the latter, the impact of the fiscal regime was not neutral
and affected a vertically integrated oil companys decision to sell or retain crude oil.62 When the spot
price was lower than the official BNOC price, integrated oil companies had the incentive to sell their own
crude arms-length and buy the crude needed for their own refineries from the market. When the spot
price was higher than the assessed price, oil companies had the incentive to keep the oil for use in their
own refineries. In doing so, the oil companies would achieve higher after-tax profits. After the abolition
of BNOC, the assessment process of transactions within the firm became more complex. The market
value of non-arms-length transactions was based on the average price of contracts (spot and forward)
61
The fiscal authorities specified a number of conditions before a contract could qualify as arms length including
the condition that the deal is not made back to back.
62
Tax spinning refers to this situation in which for fiscal reasons oil companies would resort to buying and selling
crude oil in the market though it would have been more convenient and cheaper to internalize the transaction
(Horsnell and Mabro, 2003:63).
40
preceding the deal. This encouraged oil companies whether vertically integrated or not to engage in tax
spinning through the forward market.63
Although tax spinning continued to provide a motive for trading in these markets, its importance has
declined as tighter regulations, introduced later in 1987, made it more difficult and much less predictable.
But by then, the 15-day forward market was well established and expanding fast as various market
participants including oil companies, traders, and refiners began to trade actively in this market for risk
management and speculative purposes.
The 15-day Brent market largely evolved in response to the peculiar nature of the delivery schedule of
Brent. Companies producing crude oil in the Brent system nominated their preferred date for loading at
the relevant month by the 5th of the preceding month. The loading programme was then organised and
finalised by the 15th of the preceding month. Until the schedule was completed, producers did not know
the exact date when their crude oil would be available for delivery. But these producers may have already
entered into forward contracts in which they agreed to sell their cargoes for forward delivery for a
specified price. Under the 15-day contract, sellers were required to give the buyer of the forward contract
at least 15 days notice of the first date of a three-day loading window. Under the 21-day BFOE contract,
the seller is required to provide the purchaser at least 21 days notice as to when the cargo will be loaded.
For instance, assume that on the 10th of May, the producer enters into a 21-day BFOE contract for
delivery in July. On that day the seller does not know when its crude oil will be available for delivery. In
the month prior to delivery, i.e. in June, the loading schedule is published. The seller is given a 3-day
window between the 22nd and 24th of July in which he can load the oil into tankers. The seller has to
nominate the buyer at the latest by the 1st of July which is the period required to give the buyer notice to
take delivery. Depending on the market conditions at the time of nomination, the original buyer may or
may not want actual possession of the cargo. In fact, it is likely that the original cargo purchaser has
already sold another 21-day contract (i.e. booked out his position)64, in which case he must give notice to
the new buyer to take the cargo at least 21 days in advance. In this way, the 21-day BFOE contract can
transfer hands between buyers and sellers through a daisy chain of notices until a purchaser is ready to
accept delivery or the 21-day period expires and/or the holder of the forward can no longer provide notice
for any more buyers.65 Once the notice period is expired, the oil to be loaded on a specific date is
classified and traded as Dated Brent. For instance, on the 5th of July, the cargo is traded as Dated Brent
where the delivery date is known (17-19 days ahead).
The 21-day BFOE can be either cash-settled by traders offsetting their position in the daisy chain or can
be physically settled. However, only a small percentage of forward contracts are physically settled. Figure
8 below shows the average daily traded volume on a monthly basis and the number of participants in the
21-day BFOE market. As seen from this graph, the number of players during one month is small between
four and 12 players. Furthermore, the traded volume is low not exceeding 600,000 b/d. Between
September 2007 and August 2008, liquidity in the forward market declined at a fast rate reaching the very
low level of less than 50,000 b/d in August 2008. However, liquidity recovered in 2009 and 2010 with
daily average liquidity in the first half of 2010 reaching more than 400,000 b/d. This is less than one
cargo a day compared to around 30 cargoes a day at the heyday of the 15-day Brent market during the late
1980s. Features such as the large size of the cargoes, clocking and the daisy chain games make trading in
forward Brent a risky proposition and the domain of few players. This has pushed the industry to find
63
For details on how tax spinning can be transacted through trading in the forward market, see (Horsnell and Mabro,
1993, Chapter 6 and Bacon, 1986).
64
Book out is used to describe the process whereby a daisy chain of forward transactions having been identified
(such as creating a circle in which A sells to B who sells to C who sells to A) is closed by financial settlements of
price differences rather than physical delivery.
65
In trading terms, the holder of the contract who is unable to require another purchaser to take delivery is said to
have been five-oclocked.
41
alternative ways to manage their risk without trading in the forward market, which can explain the decline
in its trading activity. The futures market has provided such an alternative. Given the central role that the
forward market assumes in the Brent complex, ensuring that there is enough liquidity in the 21-day BFOE
is crucial to the price discovery process. This is especially the case as the settlement mechanism of the
ICE futures Brent contract is based on trading activity in the forward Brent market.
Figure 8: Trading Volume and Number of Participants in the 21-Day BFOE Market
14
600,000
12
500,000
10
400,000
8
300,000
6
200,000
4
100,000
2
0
0
Jul-07
Dec-07
May-08
Oct-08
Mar-09
Participants
Aug-09
Jan-10
Jun-10
Liquidity (b/d)
Source: Argus
There are few participants in the 21-day BFOE. Unlike the futures market, the forward contract involves
trading in 600,000 barrels which is beyond the capability of many small players and hence the
composition is not as diverse as in the futures market. Table 6 below shows the various participants in the
Brent forward market and their total volume of trading during the period 2007 and 2010 (September). On
the sales side, the main players include oil companies with equity interest such as BP, Shell, Conoco
Phillips and Total and some of their trading arms such as Totals TOTSA and physical traders such as
Vitol, Phibro and Mercuria. On the purchase side, these same companies also dominate the trading
activity. For instance, in 2010, Shell was the most important seller and the third important purchaser
while Totsa was the second important seller and the second important buyer. On the purchase side, the
four top players Vitol, Mercuria, Totsa, and Shell captured more than 70% of the observed volumes by
Argus. On the sales side, these companies captured more than 60% of the trading volumes in 2010. The
degree of concentration varies across months and in certain months few players capture the bulk of traded
volumes.
42
Table 6: Participants in the 21-Day BFOE Market and their Shares in Trading Volume
Arcadia
BP
Chevron
ConocoPhillips
Glencore
Hess
Hetco
Mercuria
Morgan
Stanley
Noble
Phibro
Sempra
Shell
StatoilHydro
Total
Totsa
Trafigura
unknown
Vitol
2007
0
23,786
0
18,447
0
0
0
12,136
0
0
46,602
15,534
34,951
0
0
31,068
0
0
68,447
252,978
Sales (b/d)
2008
2009
0
0
3,005
13,699
273
274
11,749
12,329
0
274
0
9,315
0
822
12,842
64,658
0
0
19,126
18,306
62,022
273
0
16,667
0
273
12,842
159,386
274
548
25,479
13,151
125,205
0
0
53,425
0
0
48,219
369,681
2010
0
29,545
0
32,143
0
37,338
7,143
79,545
2007
485
25,243
0
6,311
0
0
0
13,107
28,896
6,494
23,377
8,766
91,883
0
649
62,987
16,234
0
56,818
483,828
0
0
36,408
18,447
46,117
0
0
61,650
0
0
43,204
252,979
Purchases (b/d)
2008
2009
0
0
273
10,959
273
0
5,464
12,329
546
548
0
10,137
0
1,096
24,863
54,247
0
0
23,770
19,672
32,787
0
0
28,962
0
273
20,492
159,383
3,014
822
36,164
13,699
73,151
0
0
108,767
0
0
42,740
369,682
2010
0
12,662
0
29,545
0
20,779
974
89,286
19,805
5,844
14,935
7,792
75,000
0
2,273
83,442
10,714
0
108,766
483,827
43
Figure 9: Average Daily Volume and Open Interest of ICE Brent Futures Contract
Source: ICE
An interesting feature of the Brent futures contract is that at expiry it cash settles against the ICE Brent
Futures Index, also known as the Brent Index which is calculated on the basis of transactions in the
forward Brent market. In other words, unlike other futures contracts whose price converges to spot price
at expiry, the Brent futures contract converges to the price of forward Brent. Specifically, the Brent index
is calculated on the basis of weighted average of first-month and second-month cargo trades in the 21-day
BFOE plus or minus average of the spread trades between first and second months as reported by oil price
reporting agencies. At expiry, the Brent futures contract relies on the forward market for cash settlement.
Thus, the effectiveness of the futures market in the role of price discovery relies on the liquidity of the
forward market which as discussed previously is quite variable and concentrated in the hand of few
players. This feature of the Brent futures contract is the result of historical events where the development
of the forward market preceded that of the futures market plus the fact that no producer in the North Sea
would back a physically delivered contract. This meant that for any Brent futures contract to succeed, it
has to be strongly linked to the forward market.
The Exchange for Physicals
Although the Brent futures contract is not physically settled, the Exchange for Physicals (EFPs) market
allows participants to swap a futures position (a financial position) with a physical one. Specifically, by
executing an EFP, a party can convert a futures position into Brent Forward or a 21-day BFOE cargo.66
EFPs are carried outside the exchange and at a price agreed between the parties. The way the EFP works
is straightforward. Party A with a futures position sells the futures contract and buys the physical
commodity. His counterparty B buys the futures position from A and sells the physical commodity to A.
Through this process, A is able to gain physical exposure to the underlying commodity while B has
swapped his physical exposure for a financial one. Such trades can be transacted at any prices agreed
between A and B and are often different from the price prevailing in the futures market. EFPs are often
quoted as differentials to the Brent futures price but usually do not exceed it by more than a few cents.
66
It is important to note that Brent EFPs are not qualitatively equivalent to physically delivered contracts such as
WTI. EFP is optional while for WTI contract, the trader has no choice but to close the position or make or take
delivery.
44
Parties need to notify the Exchange about their agreement so it can close As position and open Bs
position. Thus, the importance of EFP is that it provides a link between the futures market and the
physical dimension of the Brent market. As discussed below, in periods of thin trading activity in the
forward Brent market, the EFP provides the necessary link to identify the price of forward Brent.
The Dated Brent/BFOE
Dated Brent/BFOE, also known as Dated North Sea Light (Platts) or Argus North Sea Dated refers to the
sale of cargo with a specific loading slot. It is often referred to as the spot market of Brent.67 A spot
transaction is often thought of as a transaction in which oil is bought or sold at a price negotiated at the
time of agreement and for immediate delivery. However, Dated BFOE contracts contain an important
element of forwardness as traders rarely deal with cargoes bought and sold for immediate delivery.
Instead, cargoes are sold and bought for delivery for at least 10 days ahead. To reflect this fact, the price
of Dated BFOE is quoted for delivery 10 to 21 days ahead. For instance, on 25th May, the price of Dated
BFOE reflects the price of delivery for the period between the 4 th of June and the 15th of June (11 days).
On 26th May, the price of Dated BFOE rolls forward one day to cover the period between the 5th and 16th
of June (11 days), and so on. This element of forwardness in Dated BFOE also implies that there is a price
risk between the time when a Dated BFOE cargo is bought and the time when it is delivered. Formula
pricing can mitigate part of this risk by pricing the cargo of Dated BFOE on the time of delivery or by
using the average of prices around the loading date, such as three days before and after the loading date.
One interesting feature of the Dated BFOE market is that very few deals are done on an outright basis.
Instead, since 1988, actual deals for physical cargoes of BFOE, including Brent, are priced as a
differential to forward Brent or Dated Brent/ North Sea Dated. As seen from Figure 10 below, by 1991,
deals based on outright prices became negligible. Thus, while the forward Brent sets the price level, the
Dated BFOE market sets the differential to the forward market. More recently, forward Brent itself is
been priced as a differential to the Futures Brent.
Figure 10: Pricing basis of Dated Brent Deals (1986-1991); Percentage of Total Deals
90
80
70
60
50
40
30
20
10
0
1986
1987
1988
1989
1990
Outright Price
1991
It is important to note however that physical Brent or Brent/Ninian Blend trades at a differential to the Dated
Brent or North Sea Dated Price.
45
and Forward Brent. These CFDs can be traded in Platts window or negotiated bilaterally outside the
window or the exchanges. The high volatility in the above differential increased the risk exposure for
physical players, pushing them to hedge using CFDs. This in turn created an important niche for market
makers. Figure 11 below reports the daily volumes of traded CFDs which vary from as low as 250,000
b/d in March 2008 to as high as 1.4 million b/d in April 2010. However, these figures seem to understate
the actual volume of CFD trade with some market participants indicating that the volume of traded CFDs
is much higher.
Figure 11: Reported Trade on North Sea CFDs (b/d)
1,600,000
1,400,000
1,200,000
1,000,000
800,000
600,000
400,000
200,000
0
Source: Argus
The players in this market are quite diverse and include a large number of companies as seen in the table
below. On the sales side, the dominant players are equity producers such as BP, Chevron, Shell, Statoil;
banks such as Morgan Stanley and physical traders such as Vitol, Mercuria and Phibro. On the buying
side, these companies are also dominant. There are many companies that occasionally enter the market
and trade small volumes mainly for hedging purposes.
Table 7: Participants in the CFD Market and their Trading Volumes
Addax
Arcadia
Astra
BNP Paribas
BP
Cargill
Chevron
Chinaoil
2007
0
23,301
0
0
26,214
485
17,233
0
Sales (b/d)
2008
2009
0
411
4,918
4,658
0
0
0
548
55,601
74,085
1,913
411
26,093
70,699
0
0
46
2010
0
14,448
0
5,519
76,948
0
84,659
0
2007
0
6,553
2,427
0
43,083
485
43,811
0
Purchases (b/d)
2008
2009
0
740
10,109
6,575
1,298
0
0
2,192
37,432
24,397
4,918
274
47,541
53,863
0
1,233
2010
812
17,208
0
4,221
75,010
1,136
73,195
0
ConocoPhillips
Glencore
Gunvor
Hess
Hetco
IPC
Iplom
Itochu
JP Morgan
Koch
Lukoil
Maesfield
Marathon Oil
Masefield
Mercuria
Merrill Lynch
Mitsubishi
Mitsui
Morgan Stanley
Murphy
Natixis
Neste
Nexen
Noble
OMV
ORL
Petraco
Petrodiamond
Petroplus
Phibro
Pioneer
Plains
Preem
Sempra
Shell
Sinochem
Sinopec
Socar
Sonatrach
Standard Bank
Statoil
StatoilHydro
Totsa
Trafigura
Unipec
Valero
Veba
Vitol
485
1,456
0
971
0
0
0
0
9,223
33,010
971
0
0
0
34,345
1,942
0
0
20,388
0
0
971
1,942
0
1,485
0
0
0
5,583
20,146
0
0
0
971
47,694
0
0
0
0
0
6,796
14,563
19,782
971
8,738
1,456
0
36,044
339173
10,410
1,940
7,240
273
0
273
0
546
11,380
36,284
13,798
0
0
273
46,809
4,781
0
273
24,317
0
0
4,372
4,577
0
0
1,093
820
0
3,825
48,656
0
2,732
0
7,978
131,929
0
0
0
0
0
2,186
77,945
23,087
16,940
7,377
546
0
58,579
641772
23,041
14,219
13,151
2,192
0
2,055
548
7,671
9,153
23,556
28,559
1,644
0
1,233
59,726
1,918
0
0
57,882
0
42,033
2,740
4,003
822
14,562
0
1,644
0
1,918
68,923
0
0
685
9,644
132,079
0
1,932
0
274
932
8,630
59,233
45,260
29,315
4,521
1,096
0
132,060
961675
Source: Argus
47
33,766
24,968
3,571
22,240
3,571
325
0
7,253
7,792
3,247
24,513
1,136
0
3,247
79,471
1,299
0
0
100,487
0
19,968
0
6,951
14,286
28,545
0
974
1,948
0
82,867
0
0
0
2,273
149,221
1,136
2,597
25,000
974
5,575
108,224
0
25,974
27,955
8,955
0
0
245,692
1259585
728
485
1,942
0
0
0
0
0
1,456
11,165
485
0
11,408
0
31,311
7,646
0
1,456
20,146
0
0
0
2,427
0
0
0
0
0
1,942
36,772
0
0
0
4,854
52,699
0
0
0
0
0
4,369
6,796
14,078
3,641
0
9,951
0
15,049
339172
24,863
4,372
5,464
0
0
1,481
1,093
6,126
29,358
37,205
7,049
0
9,699
3,825
68,415
4,645
0
546
17,760
410
0
3,005
2,691
0
5,787
2,186
1,735
0
0
52,117
0
0
0
15,929
39,727
273
0
0
0
0
273
54,781
46,325
13,798
12,432
14,208
1,093
49,795
641772
28,630
9,863
3,836
2,192
1,096
3,068
548
11,041
54,973
34,849
20,411
0
548
685
99,841
0
3,014
0
51,238
0
36,849
822
5,189
548
36,995
0
2,192
822
1,644
34,400
137
2,466
3,562
15,616
83,995
603
2,800
0
7,260
548
8,630
61,863
47,397
28,877
29,170
19,726
0
112,447
961674
60,065
26,299
1,299
17,532
974
1,786
1,136
10,844
14,935
32,305
21,753
1,623
6,494
0
117,156
0
0
0
88,377
0
27,110
1,623
11,685
8,442
48,880
0
0
0
0
50,487
0
1,299
0
2,922
129,545
974
1,867
9,091
8,279
5,195
118,130
325
60,575
32,649
11,578
54,545
0
98,214
1259585
OTC Derivatives
In addition to the above layers, a whole set of financial instruments that link to the Brent complex are
currently traded over the counter (OTC). These OTC contracts are customised and until recently have
been negotiated bilaterally between parties either face-to-face or through brokers. However, as the use of
OTC became more widespread, OTC contracts became more standardised and part of the OTC activity
has shifted to electronic OTC exchanges. Furthermore, after being matched, counterparties can use the
clearing facilities of exchanges such as ICE and the CME Group. The landscape has become less benign
in a number of ways for bilateral uncleared OTC, and so there has been a shift toward clearing OTC
contracts except for those with either impeccable credit/ unimpeachable credit lines, or those who simply
cannot afford the cash flow/cash flow volatility of a cleared environment (such as airlines). IOSCO
(2010) reports that market participants conduct 55% of their trades in financial oil (crude oil and refined
products) using exchange-traded instruments and hence are subject to clearing. The remaining part of the
business is conducted through OTC. A large part of this OTC trade is now being cleared where 19% of
survey participants trades are being cleared. Only 27% of the total volume traded remains un-cleared.68
The growing similarity between more standardised OTC and exchange-traded instruments has raised the
issue of disparity in supervision and oversight between markets and is at the heart of current plans to
strengthen the regulation of commodity derivatives markets. Exchange clearing of OTC has aided their
transparency already, as they make available daily settlement figures to those clearing the instruments.
The large variety of OTC instruments and the limited information on these instruments precludes an
extensive analysis of OTC markets. ICE lists more than 30 financial contracts (for crude oil alone) that
are cleared on their exchange. These contracts are used primarily for hedging, but also for speculative
purposes and are an integral part of the Brent complex. Using these instruments one can hedge between
the various layers such as between Dated Brent and futures Brent or between further away markets such
as Dubai and futures Brent or between Dated Brent and WTI. One important and active market discussed
above is CFDs. Another active swaps market is the Brent Dated-to-Frontline (DFL) market which trades
the difference between Platts Dated Brent assessments and the ICE first month futures contract. Another
related but less liquid market has emerged which trades the difference between Dated Brent and the daily
trade-weighted Brent average reported by the ICE. Through these customised contracts, traders can
establish a series of inter-linkages not only between the different layers of the Brent market, but also
between Brent and the different benchmarks and hence are likely to influence the price formation and
price discovery processes.
BOX 1: CFD Explained with an Example
To explain the rationale behind CFDs and how it works, it would be useful to provide a simple example,
but based on real data. A refiner bought a cargo of BFOE on 19th March 19 for loading on 21st-23rd of
April. The refiner has accepted to buy the cargo at the Dated Brent price averaged over five days around
the loading date (i.e. 19th-23rd April). The refiner observes that the current value of Dated Brent is $77.88.
He is concerned that by the time of loading the price of Dated Brent could increase: he would like to
hedge his risk. In principle, he could use the April Forward contract to hedge the risk. However, this
hedge is far from perfect because there is the risk that the price of the April Forward may not follow the
movements of Dated Brent at the time of loading. This risk, referred to as the basis risk, constitutes the
main rationale for CFDs.
To hedge the basis risk, the refiner could buy (a) a second-month Forward contract (i.e. a May contract in
our example) and (b) CFDs for the week of 19th-23rd April. The price for the Forward May contract on the
68
These figures however should be treated with caution and some market participants have indicated very different
numbers. The fact remains that the size of the OTC market is not known and less so the percentage of OTC that goes
to clearance.
48
19th of March stood at $79.53 while the CFD for the week 19th-23rd April was at -$0.57. By buying the
second-month forward contract and CFDs, the refiner is able to lock the price of his cargo at $79.53.
So how does the hedge work? Somewhere between 19th-23rd April (say 22nd of April) i.e. when the cargo
is being loaded, the refiner sells the Forward May contract. On the 22nd of April, the BFOE May contract
settled at a price of $84.78. Thus, the refiner has made a profit on his forward position of $5.25: he bought
the forward contract at $79.53 and sold it at $84.78. What about the gain and losses on the CFD position?
The easiest way to think of a CFD is that it is a swap in which the refinery agrees to receive the price of
Dated Brent and agrees to pay the Forward price. Assuming that the refinery unwinds his CFD over the
week, we can calculate the net gain or loss on the CDF as illustrated in the Table below.
CFD Explained
Date
19/04/2010
20/04/2010
21/04/2010
22/04/2010
23/04/2010
Total
Loss/Gain
Dated Brent
83.19
84.74
84.47
84.64
86.49
Loss/Gain CFD
0.2(83.19-83.53)
0.2(84.74-84.86)
0.2(84.47-84.62)
0.2(84.64-84.78)
0.2(86.49-86.43)
BFOE MAY
83.53
84.86
84.62
84.78
86.43
Loss/Gain CFD
-0.068
-0.024
-0.03
-0.028
0.012
-0.138
The refinerys final position as of 23rd of April 2010 is shown in the table below. The high price paid for
the cargo in April has been compensated for by the gain in forward position. In this example, the refiner
has lost on his CFD position.69
Example of CFD (continued)
Refinerys Final Position
(23rd of April 2010)
Price Paid for the Cargo
(Average Dated Brent over the period April 19-April 23)
Gain on Forward Position
Loss on CFD
84.706
5.25
-0.138
79.594
Notice from the above example that the CFD allows us to derive in March the Forward price for Dated
Brent for the week 19th-23rd April. The Forward Dated Brent is simply the CFD plus the second month
forward i.e.
Forward Dated Brent = CFD + Second Month Forward Brent
Thus, the CFD is not the price differential between the current price of Dated Brent and the Forward
Brent Contract. It is rather the difference between the Dated Brent at some stated point in the future and
69
Notice that the refinerys position is not perfectly hedged. In the above example, the May Brent is sold in one day
and is not being closed over the five day period. The average of the BFOE May over 19th-23rd April period is
$84.884 in which case the refiner would have made a profit of $5.314 on his forward position. This will yield
$79.53, the price of the original hedge.
49
the Second Month Forward Brent.70 Since CFDs are reported for eight weeks ahead, it is possible to
derive the price of Forward Brent for eight weeks ahead. Platts refers to these forward prices as BFOE
swaps. These prices provide the vital link between the 21-Day BFOE and Dated Brent and are central for
the price discovery process in the Brent market.
An alternative way to understand the equation above is to go through the above example. By buying a Forward
contract and CFDs, the trader is able to lock today the price for Dated Brent for delivery at a certain time in the
future.
71
In essence CFDs can be traded for any week that is needed to trade, but are only reported for 8-weeks ahead.
50
assessment of these differentials through MOC process or observed deals, it is possible to calculate the
price of Dated Brent/BFOE or Dated North Sea Light (Platts) or Argus North Sea Dated (Argus) for the
day.72 Specifically, the price of Dated Brent will settle on the most competitive crude among the BFOE
combination which is usually Forties.73
The above discussion implies that during the last three decades the Brent market has evolved into a
complex structure consisting of set of interlinked markets which lie at the heart of the international oil
pricing system. The Brent market is multi-layered with the various layers being strongly interconnected
by the process of arbitrage. Thus when referring to Brent, it is important to specify what Brent is being
referred to: Dated Brent, 21-Day Brent, Brent futures, Brent CFDs or even to Brent altogether as the
continuous decline in the physical liquidity meant the Brent Blend has become less important in the North
Sea physical complex. These layers and links are central for the price discovery process as identifying the
oil price relies heavily on information derived from the financial layers. The implications of these
linkages on the oil price formation process are discussed in details in Section 8.
72
Alternatively, one can take a simple average of the four crudes which would result in Platts North Sea Basket.
As an example, on May 25, 2010, North Sea Dated Strip was priced at 68.13-68.14. This value was derived from
the Dated Brent Swap based on the average of 10- 21 window. Each of the four crudes is priced as a differential to
the forward Dated Brent. On May 25, 2010, Brent was priced at -0.11/-0.09; Forties at -0.56/-0.55, Oseberg at
0.36/0.38 and Ekofisk at 0.16/0.18. These differentials are obtained from concluded deals and failing that on bids
and offers. Since Forties is the most competitive crude, the Dated Brent/BFOE is obtained by applying the Forties
differential. Specifically Dated Brent/BFOE=North Sea Dated Strip (68.13-68.14) + Differential (-0.56/-0.55) =
67.57-67.59.
73
51
6. The US Benchmarks
West Texas Intermediate (WTI) is the main benchmark used for pricing oil imports into the US, the
worlds largest oil consumer. More crude oil is priced-off the Brent complex, but the Light Sweet Crude
Oil Futures Contract, which is based on WTI,74 is one of the most actively traded commodity futures
contract. While WTI is the most widely known US crude stream, other crude streams exist alongside
WTI. One such is the Light Louisiana Sweet (LLS) crude which has become the local benchmark for
sweet crude in the US Gulf Coast. Other important streams include the US-Gulf Coast Sour and Medium
crudes such as Mars and Poseidon (produced offshore Louisiana) and Southern Green Canyon (produced
offshore Texas). On the basis of transactions in these three crude streams, Argus derives ASCI. Platts
publishes a similar index known as Americas Crude Marker which incorporates the value of the four sour
grades: Mars, Poseidon, SGC and Thunder Horse (produced offshore Louisiana).
Source: EIA
74
The Light Sweet Crude Oil Futures contract is also referred to as the WTI futures contract.
52
2005
2006
2007
2008
2009
5,419
5,178
5,102
5,064
4,950
5,361
19
23
22
21
21
18
435
443
458
470
538
591
Kansas
93
93
98
100
108
108
North Dakota
85
98
109
123
172
218
U.S.
PADD 1
(East Coast)
PADD 2
(Midwest)
Oklahoma
PADD 3
Gulf Coast)
171
170
172
167
175
184
3,016
2,804
2,838
2,828
2,699
3,121
Louisiana
228
207
202
210
199
189
Texas
Federal
Offshore
(PADD 3)
PADD 4
(Rocky
Mountain)
1,073
1,062
1,088
1,087
1,087
1,106
1,453
1,282
1,299
1,277
1,152
1,559
309
340
357
361
357
357
141
141
145
148
145
141
1,640
1,569
1,426
1,385
1,336
1,274
Alaska
908
864
741
722
683
645
North Slope
886
845
724
707
670
638
656
631
612
594
586
567
Wyoming
PADD 5
(West Coast)
California
Source: EIA Website
2005
2006
2007
2008
2009
16,974
17,196
17,385
17,450
17,607
17,678
PADD 1
1,736
1,717
1,713
1,720
1,722
1,723
PADD 2
3,526
3,569
3,583
3,595
3,670
3,672
PADD 3
7,967
8,159
8,318
8,349
8,416
8,440
PADD 4
582
589
596
598
605
622
PADD 5
3,164
3,162
3,175
3,187
3,195
3,222
U.S.
While PADD III constitutes the major production and refining centre in the US, PADD II assumes special
importance as it is the main centre for crude oil storage and the delivery point at the expiration of the
53
Light Sweet Crude Oil Futures contract. Cushing, Oklahoma located in PADD II is a gathering hub with
large storage facilities: an estimated operable crude storage capacity of 45.9 million barrels and nameplate
storage capacity of 55 million barrels.75 PADD II itself can be divided into two sub regions: the
Midcontinent and the Midwest (Purvin and Gertz, 2010). Cushing is located in the Midcontinent. It
collects crude oil from Texas, surrounding Oklahoma and other imported crude. It links to major
refineries centres both in the Midcontinent, the Midwest (PADD II) and PADD III through a complex set
of pipelines.76 Historically, the refineries in the Midcontinent relied on domestic crude for their runs.
However, with the decline in domestic production, refineries in the Midcontinent increased their reliance
on foreign imports and Canadian crude delivered into Cushing and the broader region. A similar picture
also emerged for the Midwest where historically it has relied heavily on domestic production. However,
given the decline in production and its proximity to Canada, Canadian crude started to rise in importance
displacing domestic production and imports from outside Canada, a trend which is likely to continue. As
seen in Table 10 below, in 2009 Canadian imports accounted for 90% of total oil imports into PADD II.
In contrast, refineries in PADD III have access to a wide variety of crude oil with offshore imports from
OPEC constituting the bulk of total imports.
Table 10: Total Imports by District from OPEC and Canada (Million b/d)
PADD1 (Total)
2004
1,549
2005
1,605
2006
1,497
2007
1,495
2008
1,421
2009
1,244
OPEC
764
893
844
936
807
587
Canada
197
215
210
263
260
215
PADD 2 (Total)
1,584
1,516
1,514
1,497
1,517
1,407
OPEC
370
323
300
345
297
154
Canada
1,054
1,039
1,150
1,125
1,176
1,222
PADD 3 (Total)
5,768
5,676
5,656
5,611
5,375
5,090
OPEC
3,448
3,131
3,147
3,533
3,521
3,013
Canada
18
20
59
96
106
126
PADD 4 (Total)
260
271
278
278
264
232
Canada
260
271
278
278
264
232
PADD 5 (Total)
926
1,057
1,173
1,149
1,206
1,040
OPEC
Canada
460
87
469
88
493
105
574
126
790
151
601
148
Source: EIA
Although a wide variety of crude oils is produced in the US, WTI assumes special importance in the
global oil and financial markets since WTI underlies the Light Sweet Crude Oil futures contract, one of
the largest traded commodity futures contract. It should be noted however though that trade around
Cushing, and a forward market around that trade, existed prior to the establishment of the futures market.
75
Storage operators keep 41pc of tank space for their own use and lease 59pc to third parties. Plains and Magellan
plan to add a combined 8.25mn bl of new storage at Cushing next year. See Argus Global Markets (2010), EIA
Reveals Cushing Tank, 6 December
76
For details see Purvin and Gertz, 2010.
54
That forward market existed in parallel to the futures market through the late 80s and early 90s. However,
unlike the Brent market, as futures volumes grew, it eventually eliminated the need for the forward
market. This forward market was knows as the WTI Cash Market. Its last vestige exists now only in the
3 days between futures expiry and pipeline scheduling on the 25th of each month, discussed in details
below.
WTI is a blend of crude oil produced in the fields of Texas, New Mexico, Oklahoma and Kansas. It is a
pipeline crude and deliveries are made at the end of the pipeline system in Cushing, Oklahoma. As in the
case of Brent, the WTI market is also characterised by a large number of independent producers who sell
their crude oil to large number of gatherers. However, unlike Brent which is waterborne crude, WTI is
pipeline crude and thus is subject to problems of logistical and storage bottlenecks. Brent is exportable
which makes it more flexible and more responsive to trading conditions in the Western Hemisphere.
Furthermore, as discussed later in this section, WTI can show serious dislocations from other markets in
some occasions, reducing its attractiveness as a global benchmark or even as a US benchmark.
2000000
0
1995
1996
1997
1998
1999
2000
2001
2002
55
2003
2004
2005
2006
2007
2008
2009
2010
Unlike the Brent futures contract (where delivery is elective via the EFP mechanism), the Light Sweet
Crude Oil Futures contract is fully physically delivered for every contract left open at expiry by default.
It specifies 1,000 barrels of WTI to be delivered at Cushing, Oklahoma. The contract also allows the
delivery of domestic types of crude (Low Sweet Mix, New Mexican Sweet, North Texas Sweet,
Oklahoma Sweet, and South Texas Sweet) and foreign types of crude (Brent Blend, Nigerian Bonny
Light and Qua Iboe Norwegian Oseberg Blend and Colombian Cusiana) against the futures contract. It is
important to note though that only a small percentage of the volume traded is physically settled with most
of the physical settlement occurring through the EFP mechanism. EFP provides a more flexible way to
arrange physical delivery as it allows traders to agree on the location, grade type, and the trading partner.
Crude oil futures contracts are traded for up to nine years forward. However, liquidity tends to decline
sharply for far away contracts (see Figure 14). For instance, on October 19, 2010 the bulk of the trading
activity concentrated on the December 2010 contract. There is some liquidity up to one year ahead, but as
we move towards the back end of the futures curve, liquidity tends to decline sharply. For instance, on
October 19, 2010, the traded volume of December 2017 and December 2018 contracts stood at 33 and 4
contracts respectively.
Figure 14:Liquidity at Different Segments of the Futures Curve (October 19, 2010)
500000
460127
450000
400000
350000
300000
250000
200000
150000
100000
20991
50000
4
Nov-18
Aug-18
Feb-18
May-18
Nov-17
Aug-17
Feb-17
May-17
Nov-16
Aug-16
Feb-16
May-16
Nov-15
Aug-15
Feb-15
May-15
Nov-14
Aug-14
Feb-14
May-14
Nov-13
Aug-13
May-13
Feb-13
Nov-12
Aug-12
May-12
Feb-12
Nov-11
Aug-11
Feb-11
May-11
Nov-10
In addition to the futures and option contracts, a group of OTC financial instruments link to the WTI
complex, allowing participants to use more customised instruments than those available in the futures
market. As discussed in the case of Brent, a large fraction of OTC deals linked to WTI are using the
clearing facilities of the CME Group or ICE. The CME group lists more than 90 OTC financial contracts
for crude oil that are cleared on its exchange. Contracts such as the WTI-Brent (ICE) Calendar Swap
Futures and the WTI Calendar Swap Futures are more customised and are traded OTC but cleared
through the exchange.
in physical production and consolidation within the industry. In 2009, the combined spot-market traded
volume for twelve US domestic grades (for the month of April) stood at more than 1.8 mb/d77 (see Figure
15) which is much higher than other benchmarks including BFOE, Oman and Dubai.
Figure 15: Spot Market Traded Volumes (b/d) (April 2009 Trade Month)
450,000
400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000
0
Most of those crudes imported into the US and sold in the spot market are linked to WTI with some
exceptions such as Iraq, Kuwait and Saudi Arabias sales to the US which are linked to ASCI; some
imports from West Africa and the North Sea which are linked to Dated Brent; and some Canadian East
Coast crudes which also link to Dated Brent. While producers still use the assessed prices of WTI in
their pricing formula, those assessments are often made as a differential to the settlement price in the
futures market. In other words, it is the futures market that sets the price level while assessed prices by
oil price reporting agencies set the differentials.
The physical delivery mechanisms complicate the price assessment process. In the futures market, trading
in the current delivery month expires on the third business day prior to the twenty-fifth calendar day of
the month preceding the delivery month. For instance, the March WTI futures contract expires on the 22nd
of February. Under the terms of the futures contract, delivery should be made at any pipeline or storage
facility in Cushing, Oklahoma and must take place no earlier than the first calendar day of the delivery
month (March) and no later than the last calendar day of the delivery month (March). At expiration, three
business days are needed for pipeline scheduling to organise the physical delivery in March. The threeday window between the expiration of the monthly NYMEX WTI contract and the deadline for
completing the shipping arrangements (i.e. between the 22nd and 25th of February in our example) is
known as the roll period. During this period, the March WTI futures contract has already expired while
the spot (physical) month is still March.78 To derive the spot price of WTI March, PRAs assess the cash
roll which is the cost of rolling a contract forward into the next month without delivering on it. This
transaction can also simply be a purchase/sale of current month supply valued at an EFP to next month
futures. On the 26th of February, the physical front month becomes April which can then be linked to the
April WTI futures contract.
77
78
57
Historically, a large number of independent producers used to sell their crude oil to gatherers based on
WTI posting plus (P-Plus), which is the sum of the wellhead posted prices plus delivery costs into
Cushing. Nowadays the P-Plus market is widely used with its sister market, the differential to Nymex
Calendar Monthly Average (CMA) market. The P-Plus market used the Koch posting only as a basis up
until about 3 years ago when Koch stopped publishing that. Now companies tend to transact versus the
ConocoPhillips posting. The value that the differential to Postings (P-Plus) represents is the value for
delivery into Cushing in the current calendar month, assuming a certain cost to move the barrels to
Cushing. ConocoPhillips is known to use the Nymex settlement price, adjusted by the cost of moving the
barrel to Cushing, to set the price of the posting. This way the CMA and P-Plus markets are
mathematically connected and never too far out of synchronisation. The CMA market has been gaining
liquidity and is increasingly being used to value prompt crude oil in the US. It is the most active market in
terms of volumes of spot trade as seen from Figure 15. It is important to note that CMA is an extension to
the futures market. The CMA market does not trade price levels, but often trades at a differential to the
WTI futures contract settlement price. CMA and P-Plus have replaced the WTI Cash Window.79
Platts uses its window to assess WTI differential to CMA and other domestic crudes. While the CMA
market is quite liquid with large and diverse number of players, the percentage of transactions in the
Platts window is only a small fraction of total transactions during the day.80 In June 2007 for instance,
total window trade amounted only to 4% of entire day trade observed by Argus. For all US crudes, total
window trade amounted to 2.4% of all spot trade.81 Some crude streams such as Mars show 19 days of no
trade in June 2007 and prices were assessed based on bids and offers. 82 Furthermore, despite the diversity
of players in the market, the degree of concentration in the window is quite high with a few players
dominating the trading activity.83 Given these concerns and the fact the CMA is priced as a differential to
the price in the futures market, it is surprising that producers do not more widely use futures prices
directly into their pricing formula.84 The WTI futures contract is a physical one and the price of the
futures contract converges to the spot price at the expiration of the contract. Hence, in the case of WTI,
the use of the futures price instead of assessed prices in the pricing formulae would make little difference.
The depth and the high liquidity of the futures market surrounding WTI and the diversity of its market
participants should incentivise buyers and sellers to use the futures price in their formula pricing. In
practice, there is some evidence that the front-month WTI futures price can exhibit high volatility around
the expiry date in some instances, which may partly explain the preference of some traders to stick to
assessed WTI prices. Furthermore, both the P-Plus and CMA are means of valuing WTI that is one month
prompter than the promptest futures contract.
The WTI Cash Window, which was/is a Platts mechanism for setting the price of WTI at 3:15 EST after the close
of the Nymex at 2:30 EST, has not traded for about 3 years. It is no longer an operative index because very few
companies use it for price reference.
80
Argus Global Markets (2007), Liquidity and Diversity Prevail, 24 September.
81
Argus, State of the Market Report: US Domestic Crude, Argus White Paper.
82
Ibid.
83
Ibid.
84
It is important to note though that many companies do use the NYMEX settlement as a pricing index.
58
pipelines, especially southbound towards the US Gulf major refining centre. In some occasions, this has
led to a larger than expected build-up of crude oil inventories in Cushing. For instance, in 2007, due to
logistical bottlenecks, there was a large build-up of inventories as a result of which the WTI price
disconnected not only from the rest of the world, but also from other US regions. In 2008, the build-up of
inventories in Cushing due to a deep contango and reduction in demand induced by the credit crunch
caused a major dislocation of WTI from the rest of the world.
Given the major role that WTI plays in the pricing of US domestic crude, imported oil into the US and
global financial markets, the price effects of such logistical bottlenecks are widespread. First, dislocations
result in wide time spreads as reflected in the large differential between nearby contracts and further away
contracts as seen in Figure 16 below. For instance, in January 2009, the spread between a twelve-week
ahead contract and prompt WTI reached close to $8 with implications on inventory accumulation.
Figure 16: Spread between WTI 12-weeks Ahead and prompt WTI ($/Barrel)
+10.00
+8.00
+6.00
+4.00
+2.00
...
-2.00
-4.00
Jan 10
Sep 09
May 09
Jan 09
Sep 08
Jan 08
May 08
Sep 07
Jan 07
May 07
Sep 06
Jan 06
May 06
Sep 05
Jan 05
May 05
Sep 04
Jan 04
May 04
Sep 03
Jan 03
May 03
Sep 02
Jan 02
May 02
Sep 01
May 01
Jan 01
-6.00
Dislocations also have the effect of decoupling the price of WTI from that of Brent, as reflected in the
large price differential between the two international benchmarks (see Figure 17). For instance, in
February 2009, the differential exceeded the $8/barrel mark. Similar episodes occurred in May and June
of 2007. Such behaviour in price differentials however does not imply that the WTI market is not
reflecting fundamentals. On the contrary, price movements are efficiently reflecting the local supplydemand conditions in Cushing, Oklahoma. The main problem is that when local conditions become
dominant, the WTI price can no longer reflect the supply-demand balance in the US or in the world and
thus no longer acts as a useful international benchmark for pricing crude oil for the rest of the world. It
has also become less useful as a means of pricing crude in other US regions such as the Gulf coast.
59
Most Latin American producers85 and until recently also some Middle East producers used WTI in their
pricing formula in long-term contracts. In 2010, Saudi Arabia decided to shift to an alternative index
known as the Argus Sour Crude Index (ASCI) for its US sales. Kuwait and Iraq soon followed suit. ASCI
is calculated on the basis of trade in three U.S. Gulf of Mexico grades: Mars, Poseidon and Southern
Green Canyon. Unlike WTI and LLS which are sweet and light, the ASCI benchmark is a medium sour
index. These sour crudes also do not seem to suffer from infrastructure problems and the occasional
logistic bottlenecks that affect WTI, although disruptions could take place as they exposed to potential
hurricanes, as Hurricanes Rita and Ivan illustrated. Their physical bases have benefited from the increased
production in the Gulf of Mexico and as a result the volume of spot trade in the underlying crudes is
sizeable. It is important to note that like other local US benchmarks, ASCI is linked to WTI and currently
trades as differential to WTI. In a way, the WTI Nymex price is the fixed price basis for the index and
thus ASCI is not intended to replace WTI as fixed price but instead works in conjunction with other
markets to provide a tool for valuing sour crude at the Gulf Coast (Argus, 2010:3). This explains why
newly listed derivatives instruments such as futures, options and over the counter (OTC) around ASCI did
not gain any liquidity as most of the hedging can be done using the WTI contract.86
85
Mexicos formula for sales to the USA is much more complex. It may include the price of more than one
reference crude (WTI, ANS, West Texas Sour (WTS), Light Louisiana Sweet (LLS), Dated Brent and may be linked
to fuel prices.
86
Another potential reason as to why ASCI OTC has not gained volume is because the users of the
Saudi/Kuwaiti/Iraqi crude are also often producers of the ASCI grades and as such they are internally hedged
through their own activities.
60
governments decision not to renew the oil concession in 2007 also meant that Dubai no longer satisfies
the ownership diversification criterion. The low volumes of production and thin trading activity render the
process of price discovery on the basis of physical transactions not always feasible. In a sense, Dubai has
turned into a brand or index which represents a sour basket of mid sour grades.87
The rapid decline in Dubai output has increased the importance of Oman in pricing crude oil in the East
of Suez. Oman has some of the characteristics to enable it to play the role of a benchmark such as the
volume of physical liquidity. In 2009, Omani crude oil production reached 815,000 b/d compared to an
average of 760,000 b/d in 1990-1995. The production is not subject to OPEC quotas as Oman is not a
member of OPEC and there are no destination restrictions. On the other hand, Omani crude oil production
is almost totally controlled by PDO, an upstream operating company which is responsible to all the equity
producers for optimising production and delivery through Mina Al Fahal. PDO is owned by the Omani
government (60%), Shell (34%), Total (4%) and Partex (2%). This structure has remained stable since
1977. There is an array of foreign and private domestic oil companies operating outside PDO, but these
constitute a small share of total oil output. In 2009, PDO accounted for more than 90% of the countrys
total crude oil production.
Figure 18: Dubai and Oman Crude Production Estimates (thousand barrels per day)
1000
900
800
700
600
500
400
300
200
100
0
1990-1995
1999
2004
Dubai
2005
2009
Oman
Source: Leaver, T. (2010), DME-Oman: Transparent Pricing and Effective Risk Management in a New Era,
Presentation at the Asia Oil and Gas Conference, Kuala Lumpur, June.
One observer argues that the actual production or even non-existent of Dubai crude oil is irrelevant. What is of
relevance is that by buying the Dubai brand or index one can obtain physical oil and by selling the Dubai index one
has the obligation to deliver physical oil.
62
Currently, the two most important layers surrounding the Dubai market are the Brent/Dubai Exchange of
Futures for Swaps (EFS) and the Dubai inter-month swaps markets. The Brent/Dubai EFS is similar to the
EFP discussed above but where a trader converts a Brent futures position to a forward month Dubai Swap
plus a quality premium spread. This market allows traders to convert their Dubai price exposure into a
Brent price exposure which is easier to manage given the high liquidity of the Brent futures market. As in
the case of an EFP, the EFS is reported as a differential to the price of ICE Brent. It was not possible to
obtain data on EFS volumes, but sources estimate that the volumes of Brent-Dubai EFS and Brent-Dubai
swaps in total are about 1,000-2,000 lots on an average day (i.e. about 1 million-2 million b/d) and can
easily exceed 2,000 lots on a relatively busy day. The Dubai inter-month swap reflects the price
differential between two swaps and thus is different from cash spreads. It allows traders to hedge their
position from one month to the next. Dubai inter-month swaps are actively traded in London and
Singapore and are central to the determination of the forward Dubai price. The actual volumes of intermonth Dubai is also not available, but traders reckon that about 2,000 lots of Dubai swaps (which
includes Dubai outright swaps and inter-month Dubai swaps) trade on an average day. Other sources
suggest a higher estimate with the volume of total Dubai swap (the swap leg of Brent-Dubai and
intermonth combined) in the range of 8000-10000 lots per day of which around 60% is cleared by ICE or
CME. The participants in these markets are quite diverse. Apart from some Japanese refiners, the main
players include banks (Merrill Lynch BoA, JP Morgan, Morgan Stanley, Societe Generale), refiners (BP,
Shell), trading firms (Mercuria, Vitol) and Japanese firms (Mitsui, Sumitomo).
Since 1989, spread deals in Brent-Dubai and inter-month Dubai differentials have dominated trading
activity. As seen from Figure 19, while in 1986 outright deals constituted the bulk of the deals in Dubai,
by 1989 these had declined to low levels. By 1991, spread deals constituted around 95% of the total
number of deals in Dubai with the Brent-Dubai trades playing a central role. In 1991 Brent-Dubai trades
accounted for one third of the liquidity and half of the concluded deals with the Brent market providing
the Dubai market with the bulk of its liquidity. Given the links with the Brent market, Horsnell and
Mabro (1993) argue that Dubai has become close to being little more than another Brent add-on market.
Notes: Spread deals include Dubai one-month spread, Dubai two-month spreads, and Dubai-Brent and Dubai-WTI
Spreads.
63
-0.50
Nov 09
May 09
Nov 08
Nov 07
May 08
Nov 06
May 07
May 06
Nov 05
Nov 04
May 05
Nov 03
May 04
Nov 02
May 03
Nov 01
May 02
May 01
Nov 00
Nov 99
May 00
Nov 98
May 99
May 98
Nov 97
Nov 96
May 97
Nov 95
May 96
-1.50
May 95
-1.00
The price of Dubai is assessed based on concluded deals of partials in the Platts window, failing that on
bid and offers and failing that on information from the swap markets surrounding Dubai. Thus, despite
the fact that NOCs in the Gulf have large physical liquidity which in principle allows them to set the oil
88
A market was developed in the 1980s to trade Brent partials but with the development of the Brent futures market,
the market became redundant. But trade in partials is still used by Platts to assess North Sea and Dubai crudes.
89
This is equivalent to a full 500,000-barrel cargo with an implied operational tolerance of minus 5%.
90
Settlement of cash differences that result from undeliverable partials uses the last price assessment of the trading
month.
91
The pricing of a crude off Dubai-Oman requires setting two coefficients of adjustment (one off Dubai and one off
Oman) and then taking some average between the two coefficients.
64
price, oil exporting countries have avoided assuming this role, shifting the power to set the price to few
traders that participate in the Platts window. Oil exporting countries do not participate in the window;
they simply take Platts assessment of Dubai and use it in their pricing formula. This transfer of the pricing
discovery role to Platts window achieves an important objective as oil exporters do not want to be seen
as influencing oil prices: it is the market that sets the oil price, and not oil exporters. On other hand, this
transfer of power creates some sort of mistrust in the trading activity in the Platts window.
Initially, the shift to partial trading in 2004 has produced encouraging results, increasing the volume of
trading activity and hence improving the efficiency of price discovery, reducing the bid/offer spreads, and
attracting new players to the market (Montepeque, 2005). However, in recent years, the liquidity in Platts
Dubai window has declined to a point when only few deals are concluded during a month (Figure 21). In
many days, there is no execution of partial trades. In fact, since October 2008, there has been no
execution of partial trades in 50% of trading days (Leaver, 2010). This however does not preclude Platts
from producing a value for Dubai, which can be based on bids and offers and/or information from the
value of derivatives. Only a few players such as Sietco, Vitol, Glencore, and Mercuria dominate the Platts
Dubai window at any one day. On the sell side, large Asian refineries such as Unipec and SK have been
dominant. The concentration of trading activity in the hands of few players in the Platts partials market
has raised serious concerns that some traders by investing as little as in a 25,000-barrel partial contract
can influence the pricing of millions of barrels traded everyday (Binks, 2005). However, market
participants who think that prices are being manipulated by a few players have the incentive to enter
Platts window and exert their influence on the price. Critics argue that barriers to entry can prevent such
an adjustment mechanism from taking place.
Figure 21: Dubai Partials Jan 2008 - Nov 2010
250
200
150
100
50
0
January
May
September
January
May
September
January
May
September
Source: Platts
The way that Argus derives the Dubai price sheds some light on the links between the various financial
layers surrounding Dubai. Argus approach for assessing Dubai is based on deriving information from
various OTC markets, the most important of which is the Exchange for Swaps (EFS) and the inter-month
Dubai spread contracts. The EFS price is reported as a differential to the ICE Brent futures contract. This
allows Argus to identify a fixed price for Dubai in a particular month referred to as the price of Dubai
Swap. But since Dubai is loaded two months ahead, the assessed price of Dubai say in the month of
65
December is the forward price of Dubai in February i.e. it is price for delivery of Dubai in the month of
February (call it x)92. But buyers and sellers are interested in the price of Dubai in December. To derive
the price of Dubai in December, the information from the inter-month Dubai spread market is used.
Specifically, the January-February Dubai swap price differential is subtracted from x which gives the
price of delivery of Dubai in January (call it y).93 The January-December Dubai swap price differential is
next subtracted from y to give us the price of Dubai for the month of December.94
Once the price of Dubai is identified, the derivation of the Oman price follows in a rather mechanical
way, mainly by exploiting information about Dubai-Oman spreads. If Oman partials are traded in the
window, Platts uses the price of concluded deals or bids/offers to derive the Oman price. When this is not
feasible, the Oman value will be assessed using the Oman-Dubai swaps spread95, a derivative contract
which trades the differential between Omans OSP and Dubai for the month concerned. The contract is
traded over the counter and does not involve any physical delivery. The Dubai-Oman swap price
differential will then be used in a formula which links it to the value of Dubai. Similarly, Argus assesses
the value of Oman by comparing the value of Oman with that of Dubai. Argus first calculates the
differential to Dubai swaps and then adds it or subtracts it from Dubai outright swap to get the Oman
forward price. So currently, the assessment of Oman price by PRAs is a simple extension of the Dubai
market, where the Dubai/Oman spread provides the necessary link.
The above price derivation shows clearly that the Brent futures market sets the price level while the EFS
and the inter-month Dubai spread market set the price differentials. These differentials are in turn used to
calculate a fixed price for Dubai. In a sense, the price of Dubai need not have a physical dimension. It can
be derived from the financial layers that have emerged around Dubai. This has raised some concerns as
calls to use swaps as pricing benchmarks for physicals are at best uninformed as swaps are derivatives of
the core physical instruments (Montepeque, 2005). But this neglects the fact that liquidity in Platts
Dubais window is thin. In addition, the argument against using swaps is inconsistent with Platts use of
swaps (CFDs) in identifying the price of Dated Brent. It is also inconsistent with the fact that at times
when no partials are trading, Platts has no alternative but to use the EFS to identify the Dubai price.
Another concern is that unlike the WTI-Brent differential which reflects the relative market conditions in
Europe and the USA, Horsnell and Mabro (1993) argue that the Brent-Dubai differential does not usually
reflect the trading conditions of Asian markets except on some rare occasions such as the Iraqi invasion of
Kuwait. In normal times, Dubai crude is more responsive to trading conditions in Europe and the US than
the Far East. Specifically, the authors argue that the Brent-Dubai differential reflects better the
relationship between prices of sweet and sour crudes. In support of this hypothesis, they argue that when
OPEC decides to cut production, these cuts affect the production of heavy sour crudes. As a result, the
price of these crudes will strengthen relative to sweet crudes leading to the strengthening of Dubai prices
relative to Brent. The recent growth of the Asia-Pacific market and the wide entry of Asian players may
have changed these dynamics with the Dubai-Brent spread currently responding more closely to Asias
trading conditions making Brent-related cargoes either more attractive (small Brent premium) or less
attractive (large Brent premium) to Asia-Pacific buyers, but this need further empirical investigation.
66
relatively low. In 2009, the average daily volume of traded contracts amounted to slightly more than 2000
contracts, which is very low especially when compared to the traded volume of WTI or Brent futures
contracts.
Figure 22: daily Volume of Traded DME Oman Crude Oil Futures Contract
9,000
8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
1st June 2007
2nd July 2007
1st August 2007
30th August 2007
1st October 2007
30th October 2007
29th November 2007
31st December 2007
31st January 2008
3rd March 2008
2nd April 2008
1st May 2008
2nd June 2008
1st July 2008
31st July 2008
29th August 2008
30th September 2008
29th October 2008
27th November 2008*
29thDecember 2008
29th January 2009
02nd March 2009
31st March 2009
30th April 2009
1st June 2009
30th June 2009
30th July 2009
28th August 2009
28th September 2009
27th October 2009
25th November 2009
24th December 2009
27th January 2010
25th February 2010
26th March 2010
27th April 2010
26th May 2010
25th June 2010
27th July 2010
25th August 2010
24th September 2010
DMEs Oman futures contract allows settlement against physical delivery of Oman crude. One interesting
feature of the DME futures contracts is the large number of contracts that converge for physical delivery
in any given month. Figure 23 below traces the evolution of the trading volume and open interest for the
October 2010 Futures contract during the month of August. On 31st August, 2010 the open interest
reached almost 21,000 contracts. This is equivalent to 21 million barrels a month comprising more than
80% of Omans monthly crude oil production. By any standard, these are very large volumes to be
delivered through futures contracts. For instance, physical delivery on the Light Sweet Crude Oil Futures
contract exceeded four million barrels only once in January 1995. Also in contrast with other benchmark
contracts, the open interest on the DME contract tends to increase as contract expiry approaches as shown
in Figure 21. This represents an important anomaly and implies that the DME contract is simply used as a
means to access physical Oman crude oil. This feature sets aside the DME contract from the other
successful futures contracts that have evolved around Brent and WTI.
67
Figure 23: Volume and Open Interest of the October 2010 Futures Contracts (Traded During
Month of August)
25000
20000
15000
10000
5000
Volume
31-Aug-10
30-Aug-10
29-Aug-10
28-Aug-10
27-Aug-10
26-Aug-10
25-Aug-10
24-Aug-10
23-Aug-10
22-Aug-10
21-Aug-10
20-Aug-10
19-Aug-10
18-Aug-10
17-Aug-10
16-Aug-10
15-Aug-10
14-Aug-10
13-Aug-10
12-Aug-10
11-Aug-10
10-Aug-10
09-Aug-10
08-Aug-10
07-Aug-10
06-Aug-10
05-Aug-10
04-Aug-10
03-Aug-10
02-Aug-10
Open Interest
The introduction of the DME contract has changed the pricing mechanism of Omani crude. From its
inception, it was clear that both a retroactive official selling price (OSP) and futures market-related price
undermined the market function of price discovery.96 Thus, it was a matter of time before Oman decided
to change its pricing from a retroactive pricing system to a forward pricing system based on the DME
contract. The OSP for Oman crude for physical delivery is calculated as the arithmetic average of the
daily settlement prices over the month. For instance, the OSP for Oman crude for the month of June is
calculated as the arithmetic average of the daily settlement of price over the month of June for delivery in
two months i.e. in the month of August. The Government of Dubai has also ceased the pricing of its crude
oil sales off its current mechanism and instead utilises DME Oman futures prices providing additional
boost to the contract. Dubai and Oman however have been the exceptions so far. Despite Dubais low
physical liquidity, Platts Dubai/Oman assessments are still the preferred price benchmark used in the
pricing formula for exports to Asia. This raises the question why other Middle Eastern producers have not
been enthusiastic about adopting the DME Oman Crude Oil Futures contract as the basis of pricing crude
oil.
The futures market plays two important roles: price discovery and hedging/speculation or what is termed
as risk management. Liquidity is crucial for the efficient performance of these two functions. Physical
deliverability, which the DME tends to emphasize, is less important. In other words, deliverability is not a
sufficient condition for the success of the DME Oman contract. In fact, physical deliverability can reduce
the chances of the success of a futures contract if market participants have doubts about the likely
performance of the delivery mechanism or if physical bottlenecks around delivery points result in some
serious dislocations although the extensive use of the DMEs physical delivery mechanism demonstrates
confidence in its performance. Nevertheless, inability to increase trading liquidity while physical
deliverability continues to rise may undermine the contract as the risk of physical delivery tends to rise,
especially for those players that are not interested in physical delivery in the first place. If low liquidity
persists, then the two functions of price discovery and risk management would be undermined and the
contract would fail to attract the attention of market participants.
96
In a retroactive pricing system, the OSP applied to cargoes that have already been loaded. In a forward pricing
system, the price for an oil shipment to be loaded say in May is determined two months before i.e. in March.
68
Asian interest is crucial for the long term success of the contract as the Asia-Pacific region is the main
importer of Middle Eastern sour crude oil. However, big Asian refineries havent so far shown strong
enthusiasm for the contract. As to the financial players/speculators, the DME futures contract can open
new opportunities for trading and risk management. But speculative and hedging activity will not be
attracted to a market with low liquidity. Market participants often prefer to trade only in the most liquid
markets. The recent launch by CME of DME linked swap and option contracts is geared to providing new
risk management tools in the hope of attracting more financial players and Asian refineries into the
market. While Gulf oil producers do not hedge their oil production in the futures market, they have
interest in a sour futures contract for export pricing purposes. Low liquidity however is likely to
discourage the already very cautious Gulf oil exporters from setting their crude price against the DME
futures contracts. So far, none of the big gulf producers such as Saudi Arabia, Kuwait, Qatar, and Iran
have shown much interest in the newly established sour futures contracts. However, there is the
temptation for some Gulf countries to shift part of the global oil trading activity to the region, which may
induce a change in some oil exporters attitude towards the contract. There is also strong interest in the
success of the DME contract as evidenced by the heavy involvement of the CME Group and the various
stakeholders.97 Without this strong interest and support, the contract would have perhaps failed by now.
97
The DME is a joint venture between Tatweer (a member of Dubai Holding), Oman Investment Fund and CME
Group. Global financial institutions and energy trading firms such as Goldman Sachs, J.P. Morgan, Morgan Stanley,
Shell, Vitol and Concord Energy have also taken equity stakes in the DME (Source: Dubai Mercantile Exchange
Website).
69
The fact that ANS stopped acting as a benchmark suggests that there is a level below which integrity of the
benchmark is threatened.
99
See for Instance, Liz Bossley (2003), Battling Benchmark Distortions, Petroleum Economist, April. More
recently, concerns about squeezes arose when one oil trader HETCO took control of the first eight North Sea Forties
crude oil cargoes loading in February 2011 and two Brent cargoes with market observers describing such a move as
a trading play intended to influence the spot market. Reuters (2011), Oil Trader Takes Control of 10 North Sea
Oil cargoes, January 18.
100
The challenge of the U.S. Federal trade commission to the BP Amoco-Arco merger was partly based on the fear
that by controlling the physical infrastructure, the WTI futures market can be squeezed. The Federal trade
commission notes that the restriction of pipeline or storage capacity can affect the deliverable supply of crude oil in
Cushing and consequently affect both WTI crude cash prices and NYMEX futures prices (p.7). Then it states that a
firm that controlled substantial storage in Cushing and pipeline capacity into Cushing would be able to manipulate
NYMEX futures trading markets and they enhance its own positions at the expense of producers, refiners and
70
becoming less prevalent in jurisdictions where regulators enforce the laws against abuse of market power,
and where those laws are clear. Also important is the design or the architecture of the market/contracts in
which PRAs, in consultation with market participants, play a key role in determining its main features and
structures and evolution over time. Regulators have also turned their attention to this issue where some
observers consider that the proposed spot-month position limit formula seeks to minimize the potential
for corners and squeezes by facilitating the orderly liquidation of positions as the market approaches the
end of trading and by restricting the swap positions which may be used to influence the price of
referenced contracts.101
So far the low and the rapid decline in the physical base of existing benchmarks have been counteracted
by including additional crude streams in assessed benchmark. This had the effect of reducing the chances
of squeezes as these alternative crudes could be used for delivery against the contract. Although such
short-term solutions have been successful in alleviating the problem of squeezes, they should not distract
observers from raising some key questions: What are the requisite conditions for the emergence of
successful benchmarks in the most liquid market in terms of production? Would a shift to price
assessment in such markets improve the price discovery process? Such key questions remain heavily
under-researched in the energy literature and do not feature in the producer-consumer dialogue.
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
-1000
-1500
-2000
-2500
OECD
Non-OECD
Source: BP (2010)
traders (p. 7) (United States of America Before Federal Trade Commission in the Matter of BP AMOCO P.L.C. and
Atlantic Richfield Company downloadable from http://www.ftc.gov/os/2000/08/bparco.pdf
101
Proposed Position Limits for Derivatives, Statement of Bruce Fekrat, Senior Special Council, Division of
market Oversight, CFTC, December 16, 2010.
71
The emergence of the non-OECD as the main source of growth in global oil demand has had far reaching
implications on the dynamics of oil trade flows. This is perhaps best illustrated in the shift in the direction
of oil flows from Saudi Arabia and Russia, the two biggest oil producers in the world towards the East.
As shown in Figure 25, in 2002 Saudi Arabias share of oil exports to the US and Europe amounted to
28.2% and 17.9% respectively. In 2009, these shares declined to 17.8% for the US and 10% for Europe.
In 2009 Saudi Arabia abandoned its St Eustatius storage facility in the Caribbean which was mainly used
to feed US markets and instead obtained storage facility in Japan to feed Asian markets.
Figure 25: Change in Oil Trade Flow Dynamics
15%
28%
US
10%
Europe
Europe
54%
US
Others
Others
75%
18%
So far, Russias exports have been heavily concentrated towards Europe to which in 2009 it exported
around 7 mb/d compared with 1.17 mb/d to Asia Pacific.102 These dynamics however are changing as
Russia builds new infrastructure in an attempt to shift part of its oil exports towards the Far East. The
inauguration in December 2009 of the first section of the Eastern Siberia Pacific Ocean (ESPO) pipeline
represents a marginal but nonetheless important step in that direction. The first section of ESPO is a 2,757
km long pipeline connecting Taishet in East Siberia to Skovorodino in Russias Far East, near the border
with China. It has a capacity of 600,000 b/d is expected to grow to 1 million b/d by 2012, and potentially
to as much as 1.6 million b/d in 2015. The second stage of the project involved linking Skovorodino to a
new export terminal at Kozmino on the Pacific coast in order to supply some of the rapidly growing oil
demand in Asia. China and Russia then agreed to construct an offshoot from Skovorodino to Daqing in
China with a capacity of 300,000 b/d. It was completed by the close of 2010.
Such changes in trade flow patterns are likely to accelerate as the centre of consumption growth continues
to shift from OECD to emerging economies. The EIA103 predicts that between 2007 and 2035, oil
consumption is expected to increase by around 24 mb/d from 86.1 mb/d to 110.6 mb/d with non-OECD
accounting for almost all of the increase during this period. This shift in the dynamics of trade flows
towards the East is likely to have profound implications on pricing benchmarks. Questions are already
being raised as to whether Dubai, Minas and Tapis still constitute appropriate benchmarks for pricing oil
in Asia given their low liquidity or whether new benchmarks are needed to reflect more accurately the
102
103
72
shift in trade flows. In this respect, a debate has already started on the suitability of ESPO to act as an
Asian benchmark.104 Since ESPO competes with Mideast crudes, so far ESPO has strengthened the Dubai
benchmark. Since December 2009, Platts has been assessing the value of ESPO but as a differential to
Platts Dubai. In the longer term, ESPO has some of the features that may allow it to assume the role of a
benchmark itself. The pricing point in Northern Asia is particularly attractive. ESPO is close to key
refining centres in China, Japan and South Korea where the sailing time from the loading port of
Kozmino to northeast Asia is just a few days, transforming the Asian market from a long haul to a short
haul market. Furthermore, ESPO volumes are larger than many of the existing benchmark and could
increase in the future. On the other hand, there is uncertainty about the volume that will be available for
sale in the spot market as a considerable amount of it is sold on long-term basis or used in Rosneft
refineries. There is also uncertainty about the quality of ESPO over time. Most importantly, for any
benchmark to emerge, market participants should have confidence that the benchmark is not subject to
manipulation which is yet to be proven. One must consider the legal, tax, and regulatory regime operating
around any particular benchmark. WTI has the US government overseeing it and a robust legal regimen.
Brent has also stable governmental oversight. Distrust of the Russian government is strong in many
companies and hence the reluctance so far to support an ESPO benchmark. Nevertheless, if discontent
with existing benchmarks intensifies, then ESPO could be one of the few options available for the
industry to fall back on.
Regardless of whether or not ESPO will eventually emerge as a benchmark, it is already having an impact
on pricing dynamics in Asia. In a sense, ESPO is likely to become or has become the marginal barrel in
Asia, displacing West African crudes in this role. Gulf suppliers have to monitor ESPO's performance
very closely when setting their price differential in relation to Dubai to maintain their export
competitiveness to Asia. This is likely to cause a decline in the size of the Asian premium over time.
See for instance, J.P. Morgan (2010), Will EPSO Emerge as a New Pricing Benchmark?, Presentation at the
Platts Crude Oil Methodology Forum 2010, London, May.
105
See for instance Brunetti and Bykahin (2009).
106
For example, Bykahin and Robe (2010) find that the composition of traders plays a role in explaining the joint
distribution of equity and commodity returns. Specifically, they find that a subset of hedge funds, those that are
active both in equity and commodity futures market can explain the increase in the commodity-equity correlations.
In contrast, swap dealers, index traders, and floor brokers and traders play no role in explaining cross-correlations
across markets.
73
to the trading strategies of physical players. The fact remains that the participants in many of the OTC
markets such as forward markets and CFDs which are central to the price discovery process are mainly
physical and include entities such as refineries, oil companies, downstream consumers, physical traders,
and market makers. Financial players such as pension funds, index and retail investors have limited
presence in some of these markets. Thus, any analysis limited to the role of non-commercial participants
in the futures markets in the oil price formation process is likely to be incomplete.
Platts use the word triangulate: Assessments will use spread relationships and derivative values to help
triangulate value. See Platts Crude Oil Methodology Forum 2010, May 2010 (London).
74
most participants. Most trade is on spreads of some sort one regional price against another, one product
price against another, one product price against a crude (feedstock) price, one time period price against
another time period. These arbitrages self-correct by traders actions such as buying in one region, where
theres too much oil, and transporting it to another region where there isnt enough and where the price is
higher to draw in the oil. This feature of the oil pricing system poses a legitimate question: how can
markets that actively trade price differentials set a price level for a particular benchmark? As noted by
Horsnell and Mabro (1993) in the context of forward Brent,
In spread deals the relationship between specified flat prices and market prices may not be very
tight. And since the focus is to a large extent on relatives, the search for price levels that
correspond to the relevant market conditions becomes less broadly based and less active. The
liquidity in that part of the market which concerns itself with the oil price level has become a
small proportion of the total liquidity of the forward market.
We postulate that the level of the oil price is set in the futures markets; the financial layers such as swaps
and forwards set the price differentials. By trading differentials, market participants limit their exposure to
risks of time, location grade and volume. These differentials are then used by oil reporting agencies to
identify the price level of a physical benchmark. Perhaps this is most evident in the US market. As
explained by Platts (2010b),
physical crude oil assessments are still widely used by the industry, but the flat price formation
is originated by the New York Mercantile Exchange (NYMEX). The highly liquid sweet crude
futures contract traded on NYMEX provides a visible real-time reference price for the market. In
the spot market, therefore, negotiations for physical oils will typically use NYMEX as a reference
point, with bids/offers and deals expressed as a differential to the futures price. Therefore,
while NYMEX acts as a barometer of market value, and negotiations for physical oil may
reference the futures value, Platts plays a distinct and complementary role to that of the exchange
(p.3).
To illustrate this last point, the recent strikes in France in October 2010 present a good experiment. As
seen in Figure 26 below, during the strike between the period 11th and 21st of October, the price
differential between Dated Brent and ICE futures Brent widened considerably reaching a peak of -$1.53
dollars per barrel on the 22nd of October.108 The widening of the differential reflected the fact that while
global oil supplies were not affected by the strike, French refineries could not buy more crude oil which
resulted in less overall demand. Oil companies and physical traders holding more oil than originally
planned were forced to clear the ex-ante excess supply by offering larger discounts. Thus, in this episode,
the bulk of the adjustment took place through the changes in price differentials and not the price levels,
perhaps because the market thought the effects of the strike on the oil markets were only temporary.109
108
It is important to note also that there is a good chunk of term structure between prompt Dated Brent and the oil
deliverable under the nearest Brent futures contract.
109
Some consider that such evidence is a clear indication that it is the prompt physical that sets the futures price.
Such natural experiments however dont shed light on this issue. One needs to show that these adjustments in
differentials occur in other than crisis situations and they are strong enough to drag down the price level. More
importantly, such evidence doesnt provide an answer to the question of how the level of oil price is determined in
the first place. It reinforces the point, however, that the futures markets set the price level and the physical layers set
the differentials, which reflect changes in the underlying fundamentals of the oil market.
75
Figure 26: The North Sea Dated differential to Ice Brent during the French Strike
01 Sep
0.2
0.0
11 Sep
21 Sep
01 Oct
21 Oct
Ice Brent=0
-0.2
Strike
officially
ends
-0.4
-0.6
$/bl
11 Oct
-0.8
-1.0
French refinery
strike begins
-1.2
-1.4
-1.6
-1.8
Source: Argus
Thus, the level of oil price, which consumers, producers and their governments are most concerned with,
is not the most relevant feature in the current pricing system. Instead, the identification of price
differentials and the adjustments in these differentials in the various layers underlie the basis of the
current oil pricing system. If the price in the futures market becomes detached from prompt fundamentals,
the differentials adjust to correct for this divergence through a web of highly interlinked and efficient
markets. The key question is whether the adjustments in differentials are strong and large enough to
induce adjustments in the futures price level. The issues of whether price differentials between different
crude oil markets and between crude and product markets showed strong signs of adjustment and whether
those adjustments affected the behaviour of oil price over the 2008-2009 price cycle have not yet received
their due attention in the empirical literature.110
But this leaves us with a fundamental question: what factors determine the price level of an oil
benchmark? The crude oil pricing system and its components such as the PRAs reflect how the oil market
functions: if oil price levels are set in the futures market and if participants in these markets attach more
weight to future fundamentals rather than current fundamentals and/or if market participants expect
limited feedbacks from both the supply and demand side in response to oil price changes, these
expectations will be reflected in the different layers and will ultimately be reflected in the assessed price.
The adjustments in differentials are likely to ensure that these expectations remain anchored in the
physical dimension of the market.
In fact, one explanation attributes the upward rise in the crude oil price in the first half of 2008 to the high
demand for very-low-sulfur diesel (Verleger, 2008). This increased the price differential between diesel and crude
oil, which in turn pushed the crude oil price up. Such an explanation points to the importance of integrating products
into the analysis. Due to space constraints, products markets were not discussed in this paper, but are the subject of
current research at the Oxford Institute for Energy Studies.
76
the functioning of the oil market and its price discovery function. Transparency in oil markets however
has more the one dimension. Although improving transparency in the physical dimension of the market is
key to understanding oil market dynamics and enhancing the price discovery function, our analysis shows
that transparency in the financial layers surrounding the physical benchmarks is as important. In this
regards, it is important to emphasize three dimensions to the transparency issue. First, obtaining regular
and accurate information on key markets depends largely on the willingness of PRAs to release or share
information. PRAs are under no legal obligation to report deals to a regulatory authority or to make the
information at their disposal publicly available. Thus, some basic but key information and data on market
structure, trade volumes, liquidity, the players and their nature, and the degree of concentration in a
trading day are not always available to the public, but they are sold to market participants at a price which
makes it worthwhile for PRAs to collect such data.
Second, the degree of transparency varies considerably within the different layers in the Brent, WTI and
Dubai-Oman complexes as well as across benchmarks. Within the Brent complex, the degree of
transparency between the various layers such as the Forward Brent, CFD and Dated Brent and futures
market is different. Similarly, in the Dubai complex, basic data on the Dubai/Brent Swaps market or the
inter-month Dubai swaps are not publicly available though the volumes and open interest of Dubai swaps
cleared through the exchanges are published. Transparency in the futures markets at least when it comes
to prices, open interest and traded volumes is relatively well established. The futures market generates a
set of prices throughout the day which are instantaneously transmitted through a variety of channels
increasing price transparency. On the other hand, a detailed description of the participants in the futures
market and the identity of counterparties to a futures contract are not made publicly available although the
exchange and regulators via the exchange do have detailed data for futures markets on these areas. This is
in contrast to the OTC market where the identities of counterparties to a transaction are known. Some
market players place a high premium on such information and thus prefer to conduct their operations over
the counter.
The third dimension of transparency relates to the extent to which assessed prices are accurate and are
reached through a transparent and efficient process. There are two aspects to this issue. The first relates to
the structural features of the oil market trading which impose certain constraints on these agencies efforts
to report deals and identify the oil price. As mentioned before, traders are under no obligation to report
prices; it is not always feasible to verify reported deals; in opaque and unregulated markets, PRAs may
need to rely on their evaluation of market conditions of specific crudes to reach an intelligent price
assessment. Thus, an important element of price transparency is the ability of PRAs to collect reliable
information in imperfect and often illiquid markets and analyse the information in an efficient and
objective manner. The second aspect is linked to the internal operations of PRAs. As discussed above, the
methodologies used to assess the oil price differ considerably across agencies. Their access to information
and the type of data used in their assessment process vary across PRAs and across markets. The
procedures applied within each of the organisations to ensure an efficient price discovery process differ as
these are internally driven and are not subject to external regulation or supervision. Thus, the degree of
price transparency is very much interlinked to the activities of PRAs and the reporting standards and other
procedures that they internally set and enforce.
77
9. Conclusions
Based on the above analysis of the current international crude oil pricing system, it is possible to draw the
following conclusions:
Markets with relatively low volumes of production such as WTI, Brent, and Dubai-Oman set the
price for markets with higher volumes of production elsewhere in the world but with fewer or
none of the commonly accepted conditions to achieve an acceptable benchmark status. So far
the low and continuous decline in the physical base of existing benchmarks has been counteracted
by including additional crude streams in an assessed benchmark. Such short-term solutions
though successful in alleviating the problem of squeezes should not distract observers from some
key questions: What are the conditions necessary for the emergence of successful benchmarks in
the most liquid market? Would a shift to assessing price to these markets improve the price
discovery process? Such key questions remain heavily under-researched in the energy literature
and do not feature in the producer-consumer dialogue. The emergence of the non-OECD as the
main source of growth in global oil demand will only increase the importance of such questions.
Doubts about the suitability of Dubai as an appropriate benchmark for pricing crude oil exports to
Asia have been raised in the past (Horsnell and Mabro, 1993). This raises the question of whether
new benchmarks are needed to reflect more accurately the recent shift in trade flows and the rise
in importance of the Asian consumer.
PRAs play an important role in assessing the price of the key international benchmarks. These
assessed prices are central to the oil pricing system and are used by oil companies and traders to
price cargoes under long-term contracts or in spot market transactions; by futures exchanges for
the settlement of their financial contracts; by banks and companies for the settlement of derivative
instruments such as swap contracts; and by governments for taxation purposes. PRAs do not only
act as a mirror to the trade. In their attempt to identify the price, PRAs enter into the decisionmaking territory. The decisions they make are influenced by market participants and market
structure while at the same time these decisions influence the trading strategies of the various
participants. New markets and contracts may emerge to hedge the risks that emerge from some of
the decisions that PRAs make. The accuracy of price assessments heavily depends on a large
number of factors including the quality of information obtained by the RPA, the internal
procedures applied by the PRAs and the methodologies used in price assessment.
The assumption that the process of identifying the price of benchmarks in the current oil pricing
system can be isolated from financial layers is rather simplistic. The analysis in this report shows
that the different layers of the oil market are highly interconnected and form a complex web of
links, all of which play a role in the price discovery process. The information derived from
financial layers is essential for identifying the price level of the benchmark. One could argue that
without these financial layers it would not be possible to discover or identify oil prices in the
current oil pricing system. In effect, crude oil prices are jointly co-determined and identified in
both layers, depending on differences in timing, location and quality.
Since physical benchmarks constitute the basis of the large majority of physical transactions,
some observers claim that derivatives instruments such as futures, forwards, options and swaps
derive their value from the price of these physical benchmarks i.e. the prices of these physical
benchmark drive the prices in paper markets. However, this is a gross over-simplification and
does not accurately reflect the process of crude oil price formation. The issue of whether the
paper market drives the physical or the other way around is difficult to construct theoretically and
test empirically in the context of the oil market.
78
The report also calls for broadening the empirical research to include the trading strategies of
physical players. The fact remains though that the participants in many of the OTC markets such
as forward markets and CFDs which are central to the price discovery process are mainly
physical and include entities such as refineries, oil companies, downstream consumers, physical
traders, and market makers. Financial players such as pension funds and index investors have
limited presence in many of these markets. Thus, any analysis limited to non-commercial
participants in the futures market and their role in the oil price formation process is incomplete.
The analysis in this report emphasises the distinction between trade in price differentials and trade
in price levels. We postulate that the level of the price of the main benchmarks is set in the futures
markets; the financial layers such as swaps and forwards set the price differentials depending on
quality, location and timing. These differentials are then used by oil reporting agencies to identify
the price level of a physical benchmark. If the price in the futures market becomes detached from
the underlying benchmark, the differentials adjust to correct for this divergence through a web of
highly interlinked and efficient markets. Thus, our analysis reveals that the level of oil price,
which consumers, producers and their governments are most concerned with, is not the most
relevant feature in the current pricing system. Instead, the identification of price differentials and
the adjustments in these differentials in the various layers underlie the basis of the current oil
pricing system. By trading differentials, market participants limit their exposure to risks of time,
location grade and volume. Unfortunately, this fact has received little attention and the issue of
whether price differentials between different markets showed strong signs of adjustment in the
2008-2009 price cycle has not yet received its due attention in the empirical literature.
But this leaves us with a fundamental question: what factors determine the price level of an oil
benchmark in the first place? The crude oil pricing system and its components such as the PRAs
reflect how the oil market functions: if oil price levels are set in the futures market and if
participants in these markets attach more weight to future fundamentals rather than current
fundamentals and/or if market participants expect limited feedbacks from both the supply and
demand side in response to oil price changes, these expectations will be reflected in the different
layers and will ultimately be reflected in the assessed price. The adjustments in differentials are
likely to ensure that these expectations remain anchored in the physical dimension of the market.
Transparency in oil markets has more than one dimension. Although improving transparency in
the physical dimension of the market is key to understanding oil market dynamics and enhancing
the price discovery function, our analysis shows that transparency in the financial layers
surrounding the physical benchmarks is as important. In this regards, it is important to emphasize
three dimensions to the transparency issue. First, obtaining regular and accurate information on
key markets is not straightforward and depends largely on the willingness of PRAs to release or
share information. Second, the degree of transparency varies considerably within the different
layers in the Brent, WTI and Dubai-Oman complexes as well as across benchmarks. The third
dimension of transparency relates to the extent assessed prices are accurate and are reached
through a transparent and efficient process. There are two aspects to this issue. The first aspect
relates to the structural features of the oil market trading which impose certain constraints on
these agencies efforts to report deals and identify the oil price. The second aspect is linked to the
internal operations of PRAs. Thus, the degree of price transparency is very much interlinked to
the activities of PRAs and the reporting standards and other procedures that they internally set
and enforce.
The current oil pricing system has now survived for almost a quarter of a century, longer than the OPEC
administered system did. While some of the details have changed, such as Saudi Arabias decision to
replace Dated Brent with Brent futures price in pricing its exports to Europe and the more recent move to
replace WTI with Argus Sour Crude Index (ASCI) in pricing its exports to the US, these changes are
79
rather cosmetic. The fundamentals of the current pricing system have remained the same since the mid
1980s i.e. the price of oil is set by the market with PRAs using various methodologies to reflect the
market price in their assessments and making use of information generated both in the physical and
financial layers surrounding the global benchmarks. In the light of the 2008-2009 price swings, the oil
pricing system has received some criticisms reflecting the unease that some observers feel with the
current system.111 Although alternative pricing systems can be devised (at least theoretical ones) such as
bringing back the administered pricing system or calling for producers to assume a greater responsibility
in the method of price formation by removing destination restrictions on their exports, or allowing their
crudes to be auctioned,112 the reality remains that the main market players such as oil companies,
refineries, oil exporting countries, physical traders and financial players have no interest in rocking the
boat. Market players and governments get very concerned about oil price behaviour and its global and
local impacts, but so far have showed much less interest in the pricing system and the market structure
that signalled such price behaviour in the first place.
111
See, for instance, Mabro (2008). Mabro argues that the issue is whether the current price regime for oil in
international trade is an appropriate one. Nobody questions it because the vested interests in maintaining it are
extremely powerful. Banks and hedge funds are wedded to it. Some of the major oil companies have trading arms
that operate in these derivative markets like financial institutions. Their trading profits are substantial. OPEC
accepted it because they thought that it would protect them from blame. It didnt. And the question always asked is:
What is the alternative? I will simply say that no alternative will ever be found if nobody is looking for one.
112
See for instance, Luciani (2010).
80
References
Amuzegar, (1999), Managing the Oil Wealth: OPEC's Windfalls and Pitfalls, London: I.B. Tauris.
Argus (2007), State of the Market Report: US Domestic Crude, Argus White Paper, Argus Media Ltd.
Argus (2009), Argus US Crude Prices Explained, 24 September, Argus Media Ltd.
Argus (2010a), Argus Guide to Crude and Oil Products Markers, London: Argus Media Ltd.
Argus (2010b), Argus Sour Crude Index, White Paper, Argus Media ltd.
Argus Global Markets (2007), Liquidity and Diversity Prevail, 24 September.
Bacon, R. (1986), The Brent Market: An Analysis of Recent Developments. WPM 10, Oxford: Oxford
Institute for Energy Studies.
Barrera-Rey, F. and A. Seymour (1996), The Brent Contract for Differences (CFD): A Study of an Oil
Trading Instrument, its Market and its Influence on the Behaviour of Oil Prices, SP5, Oxford: Oxford
Institute for Energy Studies.
Binks, A. (2005), Middle East Crude Pricing: The Dubai Debate, Middle East Economic Survey, VOL.
XLVIII, No 7, 14-February.
Bossley, L. (2004) BFO Futures to Be Affected By New Crude Streams, Warns CEAG, MEES,
XLVII(17), 26 April.
BP (2010), Statistical Review of World Energy 2010, London: BP.
Jalali-Naini, A.R. (2009), The Impact of Financial Markets on the Price of Oil and Volatility:
Developments since 2007, OPEC Secretariat, Research Division, April.
Kilian, L. and Murphy, D. (2010), The Role of Inventories and Speculative Trading in the Global Market
for Crude Oil, CEPR Discussion Paper No. DP7753.
Leaver, T. (2010), DME-Oman: Transparent Pricing and Effective Risk Management in a New Era,
Presentation at the Asia Oil and Gas Conference, Kuala Lumpur, June.
Lobo, G. and J. Zhou (2006), Did Conservatism in Financial Reporting Increase after the Sarbanes-Oxley
Act? Initial Evidence, Accounting Horizons, Vol. 20, No.1, pp. 57-73.
Luciani, G. (2010) From Price Taker to Price Maker? Saudi Arabia and the World Oil Market,
Rahmania Foundation Occasional Papers.
Mabro, R. (1984) On Oil Price Concepts, WPM3, Oxford: Oxford Institute for Energy Studies.
Mabro, R. (1986), The Netback Pricing System and the Price Collapse of 1986, WPM 10, Oxford:
Oxford Institute for Energy Studies.
Mabro, R. (1998), The Oil Price Conundrum, Oxford Energy Forum, Issue 74, August.
Mabro, R. (2000), Oil Markets and Prices, Oxford Institute for Energy Studies, Oxford Energy
Comment, August.
Mabro, R. (2005), The International Oil Price Regime: Origins, Rationale, and Assessment, The Journal
Of Energy Literature, XI(1).
Mabro, R. et al (ed.) (1986), The Market for North Sea Oil, Oxford: Oxford University Press.
Mollgaard, H.P. (1997), A Squeezer Round the Corner? Self-Regulation and Forward Markets, The
Economic Journal, 107(440):10412.
Montepeque, J. (2005), Sour Crude Pricing: A Pressing Global Issue, Middle East Economic Survey,
XLVIII(14), April.
Newbery, D.M.G. (1984), The Manipulation of Futures Markets by a Dominant Producer, Chapter 2 in
Anderson, R.W. (ed.), The Industrial Organization of Futures Markets (Lexington, MA: Lexington
Books).
O'Hara, M. (1997). Market Microstructure Theory. UK:Blackwell.
Parra, F. (2004), Oil Politics: A Modern History of Petroleum, London: IB Tauris.
Penrose, E. (1968), The Large International Firm in Developing Countries: The International Petroleum
Industry, USA: Greenwood Press.
Pirrong, S. (1996), The Economics, Law and Public Policy of Market Power Manipulation, Boston, MA:
Kluwer.
Platts (2004), Review of the Partial Dubai and Oman Price Discovery Mechanism, Platts Discussion
Paper, November. Downloadable from: http://www.platts.com/Oil/Resources/Market%20Issues/
Platts (2010 a) Methodology and Specifications Guide: Crude Oil.
October.
Platts (2010 b), Platts Oil Pricing and MOC Methodology Explained. The McGraw Hill Companies,
June.
Purvin & Gertz (2010), The Role of WTI as a Crude Oil Benchmark, January.
82
Rehaag, K. (1999), Disappearing Benchmarks: The Demise of the Market Indexed Pricing?, IEA
Energy
Prices
and
Taxes,
4th
Quarter,
downloadable
from:
http://data.iea.org/ieastore/assets/products/eptnotes/feature/4Q1999.pdf
Seymour, A. (1990), The Oil Price and Non-OPEC Supply, Oxford Institute for Energy Studies.
Silvennoinen, A., and S. Thorp (2010), Financialization, Crisis and Commodity Correlation Dynamics,
Research Paper Series 267, Quantitative Finance Research Centre, University of Technology, Sydney.
Skeet, I. (1988), OPEC: Twenty Five Years of Prices and Politics, UK: Cambridge University Press.
Soligo, R. and A. Jaffe (2000), A Note on Saudi Arabia Price Discrimination, Energy Journal,
21(1):3160.
Stevens, P. (1985), A Survey of Structural Change in the International Oil Industry 1945-1984, in
David Hawdon (ed.), The Changing Structure of the World Oil Industry, USA: Croom Helm Ltd.
Tang, Ke and Wei Xiong. (2010), Index Investing and the Financialization of Commodities, NBER
Working Paper 16385, National Bureau of Economic Research, Inc.
Telser, L.G. (1992), Corners in Organized Futures Market, in Aggregation, Consumption and Trade:
Essays in Honour of H.S. Houththaker (eds. L.Phlips and L.D. Taylor), Dordrecht: Kluwer Academic
Publishers.
Terzian, P. (1985), OPEC: The Inside Story, London: Zed Books.
Thorne, S. (2010), A User guide to Platts Assessment Processes, Presentation at the Platts Crude Oil
Methodology Forum 2010, London, May.
Verleger (2008), Explaining the 2008 Crude Oil Price Rise, July, downloadable from:
http://www.pkverlegerllc.com/TIE0807.PDF
Yergin, D. (2009) It's Still the One, Foreign Policy, Sept/October.
83
Summary
The federal Renewable Fuel Standard (RFS) was established in the Energy Policy Act of 2005
(EPAct) and significantly expanded in the Energy Independence and Security Act of 2007 (EISA).
The RFS requires the use of renewable biofuels in transportation fuel. For 2012, the RFS requires
the use of 15.2 billion gallons of renewable fuel. Within the larger mandate, there are submandates (carve-outs) for advanced biofuels, including at least 1 billion gallons of biomassbased diesel fuel (BBD) in 2012. By 2022, the RFS requires the use of 36 billion gallons of
renewable fuels, including 21 billion gallons of advanced biofuels.
The RFS is a market-based compliance system in which obligated parties (generally refiners
and/or terminal operators) must submit credits to cover their obligations. These credits
Renewable Identification Numbers, or RINsare effectively commodities that can be bought or
sold like other commodities. For each gallon of renewable fuel in the RFS program, one RIN is
generated. Each RIN is a 38-digit number, with blocks of digits corresponding to various data,
including the year the RIN was generated, the producer of the fuel, and the type of fuel. RINs are
valid for use in the year they are generated and the following year.
From the beginning of the RFS program, there have been concerns with RIN generation and the
RIN market. As the RINs are essentially numbers in a computerized account, there have been
errors and opportunities for fraud. Because of concerns over transposed digits, invalid characters,
allegations of double-counting (intentional or unintentional) and other errors and inaccuracies,
when EPA finalized rules for the RFS as expanded by EISA (the RFS2), EPA also established a
new in-house trading system in an effort to address these concerns. All RIN transactions must be
cleared through this in-house system, called the EPA Moderated Transaction System (EMTS).
From the beginning of the RFS2 EPA has maintained that all due diligence remains the duty of
obligated parties. Under this buyer beware system those purchasing or receiving RINs must
certify their validity on their own, and they are responsible for any fraudulent RINs they pass on
to other buyers or submit to EPA for compliance.
In late 2011 and early 2012, EPA issued Notices of Violations (NOVs) to three companies that the
agency alleges fraudulently generated a combined 140 million biodiesel RINs in 2010 and 2011.
Because of these RIN fraud cases, EPA is looking at establishing a system whereby RINs can be
certified by third parties registered with EPA. (EPA may be considering other options but this is
the only one the agency has publicly discussed.) EPA is considering whether such certification
would provide obligated parties with an affirmative defense if RINs are later found to
fraudulentthat is, obligated parties would not be liable for penalties under the Clean Air Act for
the use of such RINs. Key questions include whether such an affirmative defense would also
eliminate the requirement to purchase make-up RINs. EPA expects to issue a proposed rule in late
2012 or early 2013, with a final rule some time in mid-2013.
In addition to agency action, at least one bill has been introduced that would amend the
compliance system. H.R. 6444 would require EPA to establish a RIN certification system and
would preclude the agency from later invalidating any certified RINs. Thus, under the bill, any
RIN found subsequently to be fraudulent would still count toward an obligated partys
compliance, without penalties.
Contents
Introduction...................................................................................................................................... 1
Current RFS Requirements .............................................................................................................. 1
The Role of RINs ............................................................................................................................. 3
RINs........................................................................................................................................... 4
EPA Moderated Transaction System (EMTS) ........................................................................... 4
The Market for RINs........................................................................................................................ 6
RIN Prices ................................................................................................................................. 6
RIN Volumes ............................................................................................................................. 8
Fraudulent RINs............................................................................................................................. 11
Effects on Obligated Parties .................................................................................................... 11
Quality Assurance Program ..................................................................................................... 11
Policy Options ......................................................................................................................... 12
Additional Questions ............................................................................................................... 13
What Other Types of RIN Fraud Are Possible? ................................................................ 14
How Likely Is RIN Fraud in the Future? .......................................................................... 14
How Do Various Players Benefit from the Different Policy Options? .............................. 15
Conclusion ..................................................................................................................................... 15
Figures
Figure 1. Nested RFS Mandates for 2012........................................................................................ 3
Figure 2. Simplified Schematic of RIN Trading System ................................................................. 5
Figure 3. Spot Renewable Fuel (Corn Ethanol) RIN Prices ............................................................ 7
Figure 4. Spot BBD RIN Prices ....................................................................................................... 7
Figure 5. Spot Advanced Biofuel RIN Prices .................................................................................. 8
Figure 6. Total RINs Registered 2011.............................................................................................. 9
Figure 7. Total RINs Registered 2012 YTD .................................................................................... 9
Figure 8. Estimated Aggregate RIN Value for 2011 ...................................................................... 10
Figure 9. Estimated Aggregate RIN Value for 2012 YTD ............................................................. 10
Contacts
Author Contact Information........................................................................................................... 16
Introduction
The Energy Policy Act of 2005 (EPAct, P.L. 109-58) established a renewable fuel standard (RFS),
requiring the use of biofuels (such as ethanol) in the nations fuel supply. The Energy
Independence and Security Act of 2007 (EISA, P.L. 110-140) significantly expanded this
mandate.1 The RFS mandate has been a major impetus to the development of U.S. biofuels
industries, especially the ethanol and biodiesel industries. In 2005, the United States produced 3.9
billion gallons of ethanol and 0.1 billion gallons of biodiesel. In 2011, production had increased
to roughly 14 billion gallons of ethanol and 1 billion gallons of biodiesel.
Covered parties meet their obligations under the RFS by surrendering renewable fuel credits to
EPA equal to the number of gallons in their annual obligation. These credits, known as
Renewable Identification Numbers (RINs), are generated when a batch of biofuel is produced,
and separated from the fuel by obligated parties (generally gasoline and diesel fuel refiners or
blenders). Once separated, these RINs may be traded like other commodities. Recent civil and
criminal action against parties accused of registering and selling fraudulent RINs has raised
questions about the integrity of the RIN market and EPAs oversight of the market.
This report outlines the RFS and the current RIN system, discusses the current market for various
RINs, and outlines policy considerations to address RIN fraud going forward.
For more information on the RFS, see CRS Report R40155, Renewable Fuel Standard (RFS): Overview and Issues,
by Randy Schnepf and Brent D. Yacobucci.
2
See CRS Report R41106, Meeting the Renewable Fuel Standard (RFS) Mandate for Cellulosic Biofuels: Questions
and Answers, by Kelsi Bracmort.
the vast majority of this unspecified portion has beenand is expected to besupplied by cornbased ethanol largely produced in the Midwest. At the beginning of each year, EPA determines a
percentage standard that all suppliers must meet, which is based on expected total U.S. gasoline
and diesel demand for the prior year. For example, for 2012, the overall biofuel standard is
9.23%, the advanced biofuel standard is 1.21%, the BBD standard is 0.91%, and the cellulosic
biofuel standard is 0.006%.3
The sub-mandates for advanced biofuels are nested together (Figure 1). As noted by the arrows in
the figure, fuel qualifying as one type of biofuel in the RFS qualifies for all levels above it. For
example, a gallon of cellulosic biofuel may be used to meet the cellulosic mandate, the advanced
biofuel mandate, and the overall RFS. A gallon of other advanced biofuel (e.g., sugarcane
ethanol) may be used to meet the advanced biofuel mandate and the overall mandate, but may not
be used to meet the cellulosic or BBD mandates. Corn starch ethanolthe most widely used
biofuel in the United Statesmay only be used to meet the overall RFS.4
The limitation on corn starch ethanol is roughly 8% on gasoline and diesel fuel combined.
Thus, the effective cap on corn-based ethanol is 13.2 billion gallons in 2012, based on the difference between the
overall mandate (15.2 billion gallons) and the advanced biofuel mandate (2.0 billion gallons).
4
Cellulosic Biofuel
(8.6 Mgal)
Source: CRS.
Notes: As noted by the arrows, fuel qualifying as one type of biofuel in the RFS qualifies for all levels above it.
For example, cellulosic biofuel may also be used to meet the advanced biofuel mandate and the overall RFS
mandate. However, non-cellulosic advanced biofuel (e.g., sugarcane ethanol) may not be used to meet the
cellulosic or BBD mandates. Likewise, corn starch ethanol may only be used to meet the total RFS mandate (and
not the advanced, cellulosic, or BBD mandates).
RINs
A RIN is a unique 38-character number that is issued (in accordance with EPA guidelines) by the
biofuel producer or importer at the point of biofuel production or the port of importation.5 Each
qualifying gallon of renewable fuel has its own unique RIN. RINs are generally assigned by
batches of renewable fuel production. (See box at right.)
RIN Codes
Under the RFS2 RIN formulation, Code D
identifies which of the four RFS categories
RIN=KYYYYCCCCFFFFFBBBBBRRDSSSSSSSSEEEEEEEE
total, advanced, cellulosic, or biodieselthe
Where
biofuel satisfies. Together, SSSSSSSS and
K
= code distinguishing RINs still assigned to a
EEEEEEEE identify the RIN block which
gallon from RINs already separated
demarcates the number of gallons of
YYYY
= the calendar year of production or import
renewable fuel that the batch represents in the
context of compliance with the RFSthat is,
CCCC = the company ID
RIN gallons. The RIN-gallon total equals the
FFFFF
= the company plant or facility ID
product of the liquid volume of renewable fuel
BBBBB = the batch number
times its energy equivalence value (relative to
a gallon of ethanol). For example, because
RR
= the biofuel energy equivalence value
biodiesel has an equivalence value (EV) of 1.5 D
= the renewable fuel category
when being used as an advanced biofuel,
SSSSSSSS = the start number for this batch of biofuel
1,000 gallons of biodiesel would equal 1,500
6
EEEEEEEE= the end number for this batch of biofuel
RIN gallons of advanced biofuels. If the RIN
block start for that batch was 1 (i.e.,
SSSSSSSS = 00000001), then the end value (EEEEEEEE) would be 00001000, and the RR code
would be RR = 15.
Any party that owns RINs at any point during the year (including domestic and foreign
producers; refiners and blenders; exporters and importers of renewable fuels; and RIN traders)
must register with the EPA and follow RIN record-keeping and reporting guidelines. RINs can
only be generated if it can be established that the feedstock from which the fuel was made meets
EISAs definitions of renewable biomass, including land-use restrictions. The feedstock
affirmation and record-keeping requirements apply to RINs generated by both domestic
renewable fuel producers and RIN-generating foreign renewable fuel producers or importers.
For more discussion on RINs see Robert Wisner, Renewable Identification Numbers (RINs) and Government
Biofuels Blending Mandates, AgMRC Renewable Energy Newsletter, Agricultural Marketing Research Center, Iowa
State University, April 2009, available at http://www.agmrc.org/renewable_energy/
agmrc_renewable_energy_newsletter.cfm; or Wyatt Thompson, Seth Meyer, and Pat Westhoff, Renewable
Identification Numbers are the Tracking Instrument and Bellwether of U.S. Biofuel Mandates, EuroChoices 8(3),
2009, pp. 43-50.
6
Unlike the other biofuel categories, the BBD mandate is a requirement on actual gallons. Thus, the 1.0 billion (actual)
gallons required for the 2012 BBD mandate will generate 1.0 billion BBD RINs, but 1.5 billion advanced
biofuel/renewable fuel RINs.
blender), any attached RINs are also transferred.7 The Code K status of the RIN is changed at
separation (generally when the fuel is sold from a biofuel producer to an obligated party). (See
Figure 2.)
As noted by EPA in the rule establishing the RFS2 and the EMTS, EPA views the EMTS solely as
a screening system, and all due diligence remains the duty of obligated parties.8 Under this
buyer beware9 system those purchasing or receiving RINs must certify their validity on their
own, and they are responsible for any fraudulent RINs they pass on to other buyers or submit to
EPA for compliance.
Figure 2. Simplified Schematic of RIN Trading System
Biofuel Producer /
Importer
Attached
RINs
EMTS
Renewable
Fuel
Separated
RINs
Secondary
Market
Retirement
Source: CRS, based on Rakesh Radhakrishnan, Market ConsiderationsRECs and RINs Overlap, Thompson
Reuters, September 25, 2012, p. 8, http://www.renewableenergymarkets.com/docs/presentations/2012/
Radhakrishnan.pdf.
In many cases, the RINs are detached from the actual fuel at the point of initial sale or transfer, and thus RINs may be
detached for fuel that has not yet been blended into motor fuel or sold as motor fuel.
8
EPA, Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program; Final Rule, 75
Federal Register 14732, March 26, 2010.
9
Ibid., p. 14733.
Notes: Black lines indicate RINs attached to actual biofuel gallons. Solid blue lines indicate separated RINs that
may be traded among all market participants. Dashed blue line indicates end-of-year submission of RINs by
obligated parties to EPA to meet RFS mandates. Green lines indicate actual biofuel gallons separated from RINs.
Orange lines indicate that all RIN transactions must be cleared through EMTS.
$0.04
$0.03
$0.02
$0.01
$0.00
1-Jan-11
11-Apr-11
20-Jul-11
28-Oct-11
2012
5-Feb-12
15-May-12
23-Aug-12
2011
Source: Ethanol and Gasoline Component Spot Market Prices, OPIS Ethanol & Biodiesel Information Service,
various editions (January 10, 2011-October 29, 2012).
Notes: Weekly average prices for weeks when data are available. The years in the legend refer to the year the
RINs were originally generated. For example, a 2011 RIN was generated some time in calendar year 2011, and
generally may only be used for compliance with the 2011 or 2012 standards. Most biofuels are sold under
contract, and thus spot prices may not reflect the value of all RINs traded at any given time.
$2.50
$2.00
$1.50
$1.00
$0.50
$0.00
1-Jan-11
11-Apr-11
20-Jul-11
28-Oct-11
2012
5-Feb-12
15-May-12
23-Aug-12
2011
Source: Ethanol and Gasoline Component Spot Market Prices, OPIS Ethanol & Biodiesel Information Service,
various editions (January 10, 2011-October 29, 2012).
Notes: Weekly average prices for weeks when data are available. The years in the legend refer to the year the
RINs were originally generated. For example, a 2011 RIN was generated some time in calendar year 2011, and
generally may only be used for compliance with the 2011 or 2012 standards. Most biofuels are sold under
contract, and thus spot prices may not reflect the value of all RINs traded at any given time.
$1.00
$0.80
$0.60
$0.40
$0.20
$0.00
1-Jan-11
11-Apr-11
20-Jul-11
28-Oct-11
2012
5-Feb-12
15-May-12
23-Aug-12
2011
Source: Ethanol and Gasoline Component Spot Market Prices, OPIS Ethanol & Biodiesel Information Service,
various editions (January 10, 2011-October 29, 2012).
Notes: Weekly average prices for weeks when data are available. The years in the legend refer to the year the
RINs were originally generated. For example, a 2011 RIN was generated some time in calendar year 2011, and
generally may only be used for compliance with the 2011 or 2012 standards. Unlike ethanol and BBD RINs, OPIS
only tracks prices for current-year advanced biofuel RINs. Most biofuels are sold under contract, and thus spot
prices may not reflect the value of all RINs traded at any given time.
RIN Volumes
The market for RINs is potentially very large, although the amount of RIN trading that occurs is
unclear. Although EPA reports total RINs registered by month, and the EMTS tracks trades and
RIN prices, EPA does not report these data. Likewise, publicly available data from other sources
are similarly limited.10 Figure 6 and Figure 7 show total RINs registered for 2011 and 2012
(through September). As noted above, by volume the RFS is dominated by ethanol produced from
corn starch. However, based on RIN value (multiplying yearly RIN volumes by average RIN
prices for the year), BBD RINs represented over 70% of the market in 2011 and 2012 (Figure 8
and Figure 9).
10
For example, OPIS reports daily RIN spot prices for four types of RINs, but does not report trading volume.
Advanced
Biofuel
0.23 billion
2%
Ethanol
13.6 billion
91%
CRS-9
2011
Biomass-Based
Diesel
0.90 billion
8%
Advanced
Biofuel
0.32 billion
3%
Jan.-Sept. 2012
Ethanol
9.8 billion
89%
2011
Jan.-Sept. 2012
Advanced
Biofuel
$0.16 billion
8%
Ethanol
$0.37 billion
19%
Biomass-Based
Diesel
$1.38 billion
73%
Advanced
Biofuel
$0.22 billion
14%
Ethanol
$0.21 billion
13%
Biomass-Based
Diesel
$1.18 billion
73%
Notes: Aggregate value based on total RIN volume for 2011, multiplied by the
mean of weekly average RIN prices reported by OPIS for 2011 ($0.027 for
ethanol RINs, $1.26 for BBD RINs, and $0.71 for advanced biofuel RINs).
Notes: Aggregate value based on total RIN volume for 2012 YTD, multiplied by
the mean of weekly average RIN prices reported by OPIS for January 2012
through September 2012 ($0.021 for ethanol RINs, $1.32 for BBD RINs, and
$0.70 for advanced biofuel RINs).
CRS-10
Fraudulent RINs
As noted above, in late 2011 and early 2012, EPA issued Notices of Violations (NOVs) to three
companies (Clean Green Fuels, LLC, Absolute Fuels, LLC, and Green Diesel, LLC) that the
agency alleges fraudulently generated a combined 140 million biodiesel RINs in 2010 and 2011.11
Subsequently, individuals representing two of these companies have also faced criminal
prosecution.12 Because these investigations involve potentially criminal actions, EPA has limited
the amount of information available to traders and obligated parties who may have purchased
fraudulent RINs. Thus, it is unclear whether any other NOVs will be issued in the future.
The 140 million fraudulent RINs from the three NOVs represent roughly 11% of the biodiesel
RINs generated between mid-2010 and the end of 2011, but less than 1% of the total RINs
generatedas noted above, ethanol produced from corn starch currently dominates the RFS.
However, as noted above, biodiesel RINs trade at considerably higher prices than ethanol RINs.
Thus, the fraudulent RINs represent roughly 8% of the aggregate market value for that time.13
EPA, Civil Enforcement of the Renewable Fuel Standard Program, October 18, 2012, http://www.epa.gov/
enforcement/air/renewable-fuels/fuel-novs.html.
12
On June 25, 2012, Rodney R. Hailey of Clean Green Fuels, LLC was found guilty of 8 counts of wire fraud, 32
counts of money laundering, and 2 counts of violating the Clean Air Act. On August 8, 2012, a Federal Grand Jury
indicted David Gunselman of Absolute Fuels, LLC on wire fraud, money laundering, and violating the Clean Air Act.
13
140 million RINs represent 11% of the BBD market. 11% x 73% (aggregate market share for BBD RINs) = 8% .
14
40 C.F.R. 80.1431(b)(2).
11
certification would provide obligated parties with an affirmative defense if RINs are later found
to fraudulentthat is, obligated parties would not be liable for civil penalties under the Clean Air
Act for the use of such RINs. As noted by Gina McCarthy, EPA Assistant Administrator for Air
and Radiation, the affirmative defense would ensure that refiners and other program participants
who meet the conditions of the affirmative defense will not face civil penalties.15
A key component of a Quality Assurance Program would be the establishment of a quality
assurance plan (QAP). The QAP would serve as the basis for audits by third party verifiers
certified by EPA. However, while EPA has issued draft requirements of what might be included in
a QAP, EPA has not issued proposed rules.16 According to various comments by the agency, EPA
expects to issue a proposal by the end of 2012, with a final rule some time in 2013.
Key questions include whether such an affirmative defense would also eliminate the requirement
to purchase make-up RINs. While refiners and others would prefer to not pay twice for RINs,
in general biofuel producers argue that not making up the RINs would undermine the legitimate
RIN market. Simply put, if obligated parties are not required to replace invalid RINs with valid
RINs, the size of the legitimate renewable fuel market is reduced.
Policy Options
There are various policy options to address the issues of RIN fraud. EPA could undertake some of
these under existing Clean Air Act authority, while others would require congressional action. In
general terms, there are at least four options:
1. Do nothing, and let market participants determine the credibility of actors they
trade with;
2. Establish a Quality Assurance Program or some other certification to provide
greater credibility, but do not tie it to EPAs determination on RIN validity;
3. Establish a certification procedure with an affirmative defense such that
purchasers of invalid RINs are not liable for civil penalties; and
4. Establish a system where all certified RINs are valid for RFS compliance
regardless of subsequent determination that they are fraudulent or otherwise
deficient.
Currently, RIN market participants are acting under the first option. They are independently
determining whether to trust the validity of the RINs they purchase. In many cases, obligated
parties have decided to purchase biodiesel and biodiesel RINs only from the largest producers.17
At the same time, small producers have complained that they are unable to afford the verification
procedures that some obligated parties now require. Others have argued that the RIN fraud
prosecutions have improved the integrity of the market. For example, one witness to a House
15
Letter from Gina McCarthy, Assistant Administrator for Air and Radiation, EPA, to The Honorable Gene Green,
Ranking Member, Ranking Member Subcommittee on Energy and Economy, Committee on Energy and Commerce,
August 14, 2012.
16
EPA, Public Release of Draft Quality Assurance Plan Requirements, EPA-420-B-12-063, Washington, DC, October
31, 2012, http://www.epa.gov/otaq/fuels/renewablefuels/documents/420b12063.pdf.
17
Testimony of various biofuel companies before the House Committee on Energy and Commerce Hearing on RIN
Fraud: EPAs Efforts to Ensure Market Integrity in the Renewable Fuels Program. July 11, 2012.
12
Energy and Commerce Committee hearing on RIN fraud stated [i]n essence, the Wild West of
buying and selling RINs from market participants you dont know has ended, the wrongdoers are
being rooted out, and everyone now knows that deals that are too good to be true are in fact too
good to be true.18
Because of substantial remaining uncertainty about the integrity of the market, the National
Biodiesel Board (NBB) and others are working to establish a more formal process for RIN
certification. With the NBB, biodiesel producers have established a RIN Integrity Network where
obligated parties can subscribe to a service where they can receive information on participating
biodiesel producers.19 This and similar efforts are in their early stages, and it is unclear how much
credence RIN purchasers will give these networks. Further, it is unclear whether they will become
the industry standard.
EPA is currently pursuing the third option, where a Quality Assurance Program is established with
third parties auditing RIN generators. In its public communications, EPA has stated that the
agency plans to offer an affirmative defense such that users of certified RINs would not be
subject to civil penalties under the Clean Air Act. It is unclear whether an obligated party would
be required to purchase additional RINs to make up for any certified RINs later found to be
fraudulent. Obligated parties generally would prefer not to pay twice for RINs, adding to their
compliance costs. However, eliminating this requirement would effectively shrink the market for
biofuels under the RFS, harming legitimate biofuel producers. As EISA establishes specific fuel
volume requirements, it is unclear whether EPA has the authority under existing statute to waive
that requirement.
In addition to agency action, at least one bill has been introduced that would amend the RIN
system. H.R. 6444 would require EPA to establish a RIN certification system by January 1, 2013.
The bill would preclude the agency from later invalidating any certified RINs. Thus, under the
bill, any RIN found subsequently to be fraudulent would still count toward an obligated partys
compliance, without penalties. As noted above, refiners and other obligated parties would likely
prefer this to other policy options, while biofuel producers are unlikely to support such blanket
protection. Further, it is unclear whether EPA could issue a final rule for the new system by the
January 1, 2013, deadline specified in the bill.
Additional Questions
The concerns raised above, and proposed policy remedies, raise additional questions about the
potential for RIN fraud in the future, as well as the effects on stakeholders from any policy
solution. These questions include:
1. Beyond the instances of fraud currently being prosecuted, what other instances
are unreported, and what other types of fraud are possible in the future?
2. How likely is fraud in the future, and what are the implications? and
3. How do various players benefit from the different policy options?
18
Joe Jobe, Chief Executive Officer, National Biodiesel Board, Testimony Before the United States House of
Representatives Committee on Energy and Commerce Hearing RIN Fraud: EPAs Efforts to Ensure Market Integrity
in the Renewable Fuels Program, Washington, DC, July 11, 2012.
19
Genscape, Genscapes RIN Integrity Network, http://info.genscape.com/RIN.
13
For example, see Jon P Fjeld-Hansen, Managing Director, Musket Corporation, Testimony Before the United States
House of Representatives Committee on Energy and Commerce Hearing RIN Fraud: EPAs Efforts to Ensure Market
Integrity in the Renewable Fuels Program, Washington, DC, July 11, 2012.
21
40 C.F.R. 1430(a).
14
involved may be less well known to market participants; and (3) limited verification procedures
exist.
Between actions taken by industry, forthcoming regulations from EPA, and potential
congressional action, the latter two reasons for fraud in the biodiesel RIN market may be fully
addressed. However, the first causea high price for some RINsmay continue in the future. By
2022, the RFS requires the use of 36 billion gallons of renewable fuels, more than double the
amount required in 2012. Thus, the absolute volume of the market will be larger in the future.
Further, of the 36 billion gallons required in 2022, 16 billion gallons are required to come from
cellulosic biofuels. Currently, there is very limited production of these fuels, and their production
costs are high. To the extent that cellulosic fuel costs remain high in the future, the aggregate
value of the cellulosic RIN market could be significantly higher than the total RIN market
today.22 The higher value of this market might be a draw to actors looking to circumvent the law.
Conclusion
The establishment of the Renewable Fuel Standard has created a market for RINs that has grown
both in volume and in value over time and is expected to continue to grow over the next decade.
However, cases of fraud in the early years of the biodiesel RIN market raise questions about the
22
For example, assuming a cellulosic RIN price of $0.50 in 2022, in aggregate cellulosic RINs would be valued at $8
billion, roughly four times the aggregate value of all RINs in 2011. A higher cellulosic RIN price would raise the
aggregate value proportionally.
23
H.R. 6444 would address this by removing EPAs authority to invalidate (for any reason) a certified RIN later found
to be deficient. Thus, even a deficient RIN could be traded or used to meet a partys RVO as long as it had been
certified by the agency.
15
integrity of RIN markets, as well as EPAs oversight of the markets. Various policies have been
proposed to address the potential for RIN fraud in the future, and the details of those policies will
affect the potential for fraud in the future as well as the relative benefits to different market
players.
16
Foreword
Following the successful 24th World Gas Conference in Buenos Aires in October 2009, we have
decided to convert some of the study reports presented at the conference into IGU publications,
including the report Gas Pricing written by Study Group B2 of the IGU Strategy Committee
(PGCB).
The IGU Strategy Committee is updating the review for the WGC in June 2012. Some interim
findings have been published in the April 2011 edition of the IGU Magazine and are available on
www.igu.org.
Historically, gas prices have not been in the news to the same extent as oil prices. This is changing.
The share of gas in global energy and fuel consumption has increased and also the share of
internationally traded gas globally is greater than before. LNG is providing intercontinental linkages
that eventually could constitute a global gas market.
Natural gas is an abundant resource, it is clean and cost-competitive, and should therefore play an
important role in the mitigation of climate change. However, the pricing of this valuable commodity
is critical to a sustainable market growth.
It is our hope that this publication can serve as one example of how vital information related to
gas pricing can be shared across borders to the benefit of the global gas industry and also to enable
new gas regions to learn more about the different pricing models that are being used.
June 2011
Torstein Indreb
Secretary General of IGU
Table of contents
1. Executive Summary.......................................................................................................................................4
2. Introduction....................................................................................................................................................7
Mandate..........................................................................................................................................................7
Why this report?.............................................................................................................................................7
Outline of report.............................................................................................................................................8
Terms and concepts........................................................................................................................................8
3. Gas price drivers...........................................................................................................................................10
Competitive markets....................................................................................................................................10
Short to medium term supply and demand drivers...................................................................................10
Long term supply and demand drivers......................................................................................................11
Current scenarios.......................................................................................................................................13
Other market organisation ...........................................................................................................................13
OECD area................................................................................................................................................13
Non-OECD area........................................................................................................................................14
4. Key gas pricing mechanisms........................................................................................................................15
5. Origins of individual pricing mechanisms...................................................................................................18
Origins of gas or oil market based pricing...................................................................................................18
North America...........................................................................................................................................18
The UK......................................................................................................................................................18
Continental Europe....................................................................................................................................19
Asia Pacific................................................................................................................................................19
Origins of regulated gas pricing...................................................................................................................20
6. Recent gas price developments....................................................................................................................23
OECD area...................................................................................................................................................23
Rest of the world..........................................................................................................................................24
7. Current extensiveness of individual pricing mechanisms............................................................................28
Introduction..................................................................................................................................................28
Price Formation Mechanisms.......................................................................................................................30
Types of Price Formation Mechanism......................................................................................................30
Results..........................................................................................................................................................30
Format of Results......................................................................................................................................30
World Results............................................................................................................................................30
Regional results.........................................................................................................................................34
Wholesale Prices.......................................................................................................................................38
Changes between 2005 and 2007..............................................................................................................40
Conclusions...............................................................................................................................................41
8. Trends in the extensiveness of individual pricing mechanisms...................................................................42
Towards a globalisation of gas pricing?.......................................................................................................45
Bumps in the road toward globalised gas pricing.....................................................................................47
9. Price volatility..............................................................................................................................................51
General.........................................................................................................................................................51
Causes of volatility.......................................................................................................................................52
Volatility associated with gas price increases...........................................................................................53
Volatility associated with gas price declines.............................................................................................53
Volatility of oil indexed prices.....................................................................................................................54
Volatility and LNG.......................................................................................................................................54
10. Towards further changes in the extensiveness of individual pricing mechanisms?.....................................55
Appendix 1 Price Formation Mechanisms 2005 Survey...........................................................................58
Format of Results......................................................................................................................................58
World Results............................................................................................................................................58
Regional Results........................................................................................................................................61
Wholesale Prices.......................................................................................................................................65
Conclusions...............................................................................................................................................66
Gas Pricing; Study group B2 of IGU Programme Committee B (PGC B). The work was coordinated by Runar Tjrsland, Study Group leader
with contributions of Meg Tsuda, Mike Fulwood, Ottar Skagen and Howard Rogers.
1. Executive Summary
Historically gas prices have not been in the news to the same
extent as oil prices. This is changing. The share of gas in
global energy and fuel consumption has increased. The share
of internationally traded gas in global gas consumption has
increased. LNG is providing intercontinental linkages that
eventually could constitute a global gas market. With most
gas producing OECD countries struggling to replace reserves
and sustain production growth, the centre of gravity of gas
production and exports has shifted towards the same regions
and to some extent countries that for 40 years have dominated
oil production and exports. Finally, gas prices have increased
and become more volatile.
Gas prices are not determined but definitely influenced by
individual markets choices between available price formation
mechanisms. The two main debates in this respect is the
one that goes on in Europe and to an extent in Asia between
proponents of continued indexation of gas prices to oil prices
and proponents of gas-on-gas competition based pricing, and the
one that goes on inside a number of Non-OECD countries and
between these countries governments and entities like the EU
Commission, the IEA and the multilateral development banks
on the sustainability of the more or less heavy handed price
regulation that prevails in big parts of the Non-OECD world.
Arguably the former of these debates is the least important.
Evidence from North America where gas prices are not
contractually linked to oil prices suggests that gas prices
nonetheless tend to track oil prices in a fairly stable long term
relationship. Gas and oil prices are linked by interfuel competition
in the industrial sector. They are also influenced in the same
manner and to the same extent by the oil and gas industrys cost
cycles. Finally price deviations may be arrested, eventually, by
changes in oil and gas industry investment priorities.
countries gas demand, and the gas flows from Russia, the
Middle East and Africa to the OECD are expected to further
increase. Several gas exporting non-OECD countries are however
struggling to sustain, let alone increase, their exports in the face
of booming domestic gas demand. Domestic demand reflects
among other things domestic prices. Consequently the outlook
for domestic gas pricing in these countries is no longer of local
interest only but of global importance.
This report examines the extensiveness in different parts of the
world of the following gas pricing mechanisms:
Regulation
below cost
26 %
Gas-on-gas
competition
32 %
Regulation
social and
Regulation cost
political
of service
9%
3%
Bilateral
monopoly
8%
Netback from
final product
1%
Globally, in 2007 one third of all gas sold and purchased was
priced according to the gas-on-gas competition mechanism.
Regionally the share of gas transactions in this category varied
from 99% in North America to zero in most of the developing
world.
North America, UK
u ro pe
enta l E
Con tin
Continental Europe,
Developed Asia
Bilateral monopoly
Netback from final product
Regulation cost of service
Regulation social and political
Regulation below cost
No price
2020
Gas-on-gas competition
Oil price escalation
Bilateral monopoly
Ru
ss
ia
,C
hi
na
?
Select Non-OECD
EC
N on-O
Select
Select Non-OECD
2. Introduction
Mandate
This report is as noted not an attempt to analyse in great
detail gas price movements around the world in great detail,
nor to provide another set of gas price forecasts. The mandate
given to IGU PGC B/SG2 was:
To carry out a comprehensive analysis of gas price formation
models at regional level: price drivers, indexation, price
arbitrage, demand elasticity;
To investigate future trends and the factors which could help
to minimize price anomalies and contribute to a sustainable
market growth
The work group has on the basis of this mandate set itself
the following targets:
Identify the main gas price drivers and discuss how they
operate in the short and longer term;
Offer a categorisation of how gas is priced around the world;
Discuss how individual pricing methods or models have
arisen;
Present the results of a global pricing method mapping
exercise;
Examine select trends in the use of individual pricing methods;
Discuss the roots and consequences of gas price volatility;
Offer some views on how the popularity and prevalence of
individual methods may change in the years ahead.
The gas share of the fuel mix has increased world wide;
Gas prices have increased;
Gas prices have become more volatile;
LNG is providing intercontinental gas price linkages that
eventually could constitute a global gas market;
With most gas producing OECD countries struggling to
replace reserves and sustain production growth, the centre
of gravity of gas production and exports has shifted towards
the same regions and to some extent the same countries that
for 40 years have dominated oil production and exports.
Gas prices in North America, Europe and developed Asia are
being more closely monitored than prices in the rest of the
world. This has several reasons:
Historically the OECD area has accounted for the bulk of
world gas consumption,
The worlds leading energy research institutions are located
in the OECD area and sponsored by OECD area governments
and companies,
While prices in the OECD area are market driven and therefore
amenable to standard economic theory and models, prices
in the rest of the world are with a few notable exceptions
politically determined and therefore essentially beyond
forecasting.
The validity of the first reason is wearing thin. 2007 world gas
use was split evenly between the OECD countries and the rest
of the world, and since OECD area consumption is growing
at a slower pace than non-OECD consumption, the latter area
will soon have a lead on the former. Moreover, several nonOECD countries are already playing key roles in determining
the supply of gas to world markets, and will only become more
important in this respect in the future. Their domestic gas pricing
decisions could therefore be strongly felt in the OECD area.
Russia is a case in point. Eurasian gas balance studies typically
conclude that the call on Russian gas will increase significantly
and that Gazprom, the Russian oil companies and Russias
independent gas producers need to invest massively in the upstream
and midstream to stave off shortages. This from time to time
prompts discussions on the adequacy of budgeted investments.
However, if a gap exists it may be closed by dampening future
demand as well as by boosting future supply. The bulk of Russian
gas currently almost 70% is consumed at home. Thus if the
pace of growth of Russian domestic gas use can be contained
through for instance price increases, budgeted investments in
supply may be more than adequate.
The Middle East is another case in point. Forecasters tend to
vest high shares of the responsibility for supplying world gas
demand in the decades ahead, with this region. But the Middle
Easts current and potential gas exporters are currently struggling
to sustain or start exports in the face of stagnant production and
booming domestic demand. The latter aspect of the regions
Outline of report
Chapter 3 of this report identifies the gas price drivers at
work in different markets and offers some views on how they
may develop in the years ahead.
Chapter 4 presents and briefly explains eight gas pricing
mechanisms that together capture nearly all gas produced and
consumed in the world.
Chapter 5 discusses the origins and history of each of these
mechanisms, with an emphasis on those in use in the OECD
countries.
The current interest in gas pricing models has a context, and
this context is the gas price turbulence experienced since 2000
in big parts of the world. For this reason chapter 6 offers a brief
overview of recent price developments inside and outside the
OECD area.
Chapter 7 is the core of the report in that it presents the result of
an empirical investigation of the prevalence of individual pricing
models in individual markets in 2007, and also a comparison of
the situation in 2007 to that in 2005. A sample of IGU member
organisations were asked to estimate the shares of gas sales in
their home countries that belonged to each of the eight pricing
categories. The member organisations were selected so as to
ensure that all regions and preferably all key countries were
covered. The replies were then analysed by SGB2.
Chapter 8 addresses the tensions inherent in individual pricing
mechanisms, the consequent challenges of sustaining the current
pattern of methods, and the attempts being made by market
players, politicians and regulators to introduce new methods,
typically with a view to shifting prices to more efficient levels.
Chapter 9 addresses this issue of gas price volatility. Since the
turn of the decade, gas prices have not only fluctuated around
(until recently) rising trends, they have also gyrated more
violently than typical for the 1980s and 1990s. The reasons for
and nature of the post 2000 gas price instability, and whether
the future will bring even more, or less, volatility, are questions
on every gas market players mind.
Wellhead
price Citygate
price
Wellhead price
The value of gas at the mouth of the gas well
In general the wellhead price is considered to be the sales price
obtainable from a third party in an arms length transaction
Wellhead prices are well documented for the US, less so for
other countries with less transparency in the upstream
Border/beach price
The price of gas at a border crossing or landing point
US and European natural gas and LNG import prices are
well documented by the US Department of Energys Energy
Information Administration (DOE/EIA) and Eurostat, and
by the International Energy Agency (IEA) in its quarterly
Energy
Hub price
The price of gas at a hub, typically a pipeline junction where
a significant amount of gas sales and purchases takes place
and where sellers and buyers can also purchase storage
services
A hub does not need to be physical, it can be virtual like the
UKs National Balancing Point
Serving as marketplaces, hubs are a prerequisite for gas
pricing through gas-to-gas competition
Hub prices are well documented as they underpin the worlds
gas futures markets
The US Henry Hub is the closest thing there is to a world
gas pricing point
Hub prices are optimal wholesale price indicators
However, hubs liquid enough to convey reliable price signals
exist for the moment only in the US, in the UK and to an
extent in the Benelux area
Citygate price
The price of gas at a citygate, typically at the inlet to a
low pressure pipeline grid owned and operated by a local
distribution company
US citygate prices on a monthly state-by-state and weighted
average US basis are published by the DOE/EIA
US citygate prices on balance reflect the prices on the hubs
where the gas is sourced plus transportation costs, but may from
time to time due to local supply and demand circumstances
include substantial premiums or discounts
Citygate prices are not systematically documented anywhere
else
Netback price
Gas supply chains have multiple links, and for each point of
transfer from one link to another a so-called netback price
may be calculated by deducting from the end user price the
unit costs of bringing the gas from that point to the end user
The netback price to the upstream shows the value per unit
of gas produced left for sharing between the producer and
the state after distribution, transmission, storage and in
the event of LNG regasification, shipping and liquefaction
costs have been deducted from the end user price, and is as
such a key indicator of project feasibility
2 500
50
2 000
40
1 500
30
Trend lines
20
1 000
Recent droughts,
impact on LNG
imports
10
500
08
M
ar
ch
07
06
ar
ch
M
05
M
ar
ch
04
M
ar
ch
03
M
ar
ch
02
ar
ch
M
01
M
ar
ch
00
ar
ch
M
ar
ch
99
0
ar
ch
3 000
60
16,00
Katrina, Rita
1800000
14,00
12,00
1700000
10,00
1600000
8,00
1500000
6,00
1400000
4,00
jan.08
jan.07
jan.06
jan.05
jan.04
jan.03
jan.02
0,00
jan.01
2,00
1200000
jan.00
1300000
HH (USD/MMBtu)
1900000
jan.97
3 500
70
jan.99
80
jan.98
Competitive markets
14
NBP
Zeebrugge
TTF
USD/MMBtu
13
12
11
10
9
Supply option 11
Supply option 12
Supply option 13
Supply option114
Supply option 15
Supply option 10
Supply option 7
Supply option 8
Supply option 9
Supply option 6
Supply option 5
Supply option 4
Supply option 3
Supply option 2
Supply option 1
Volume
8
7
Fe
b
10
Ja
n
27
ec
ec
Ja
n
13
D
30
ov
ov
N
D
08
24
ct
10
O
27
O
ct
13
19
Se
p
Source: WGI
800
700
600
500
Bcm
30
100
15
19
93
07
Ja
n
Ja
n
06
Ja
n
04
03
02
01
00
05
Ja
n
Ja
n
Ja
n
Ja
n
Ja
n
n
Ja
99
19
96
19
99
20
02
20
05
20
08
20
11
20
14
20
17
20
20
20
23
20
26
20
29
10
Ja
300
200
Henry Hub
20
400
19
90
USD/MMBtu
25
Current scenarios
Will all these factors driving or dampening gas supply and
demand growth sustain prices at or close to the levels observed in early-mid 2008, or has the financial crises deflated prices on a long term basis? There are as many answers to this
question as there are market observers. However, the widely
held view from a few years back that gas as the obvious bridging fuel between the oil intensive 20th century and a cleaner
21st century could look forward to several decades of robust
supply and demand growth, is being challenged.
The International Energy Agency presents in its 2008 World
Energy Outlook a business as usual scenario where world gas
demand increases by some 1500 bcm between 2006 and 2030,
or by 1,8% a year. The IEA sees US gas consumption peak at
about 650 bcm a year in 2015 before declining to about 630
bcm a year by 2030. All in all this means a 0,1% a year growth
in demand for the entire 2008-30 period.
The clause also addresses the need for regular revisits to the
formula in response to structural changes in the marketplace.
Continental European and Developed Asian border gas prices
are thus driven by the prices of crude oil and refined products,
and indirectly by all the factors that drive these prices, rather
than by developments in Continental European and Developed
Asian gas demand or in world gas supply.
This is a simplification insofar as the price signals coming
from the spot markets around Europe, from the UK via the
Interconnector and from the US via LNG do influence Continental
European and Developed Asian contract prices. Long term
import contracts always have some offtake flexibility. If spot
prices fall significantly below contract prices, buyers will
respond by offtaking as little as they can under their contracts,
turning instead to the alternatives. This will lift spot prices but
could also lead to contract renegotiations and eventually some
realignment of contract prices with gas market realities.
The current trend is towards shorter, more flexible import
contracts, so the influence from gas supply and demand on
Continental European and Asian contract prices will likely
increase. However, as we will revert to later in this report, there
is currently little to indicate that either Continental Europe or
Developed Asia will abandon oil linked pricing any time soon.
Non-OECD area
Outside the OECD area there are many gas consuming countries
that neither allow gas supply and demand to determine
prices nor practice oil linked pricing. Instead they set prices
administratively according to principles and procedures that
are not always transparent.
Supply costs may be a consideration, but do not always receive
systematic attention. If supply costs are taken into account, they
may be defined so as to include both operating, depreciation
and financial costs and a return on investments, but they may
as well be defined so as to cover operating costs only, leaving
nothing for maintenance not to mention system expansions. The
more supply costs are ignored as a driver, i.e., the further below
full cycle supply costs prices are set, the smaller is the role that
sales revenues play in financing the countrys gas supply. The
state actor(s) involved then need to be funded directly from
the state budget.
Social and political considerations are probably the most important
regulated price drivers, with the regulators aiming to set prices so
as not to hurt industrial consumers competitiveness, overburden
residential consumers and potentially trigger political unrest.
These criteria are course ambiguous, reflecting what consumers
have grown accustomed to rather than objective thresholds. The
same gas bill as a share of a households real disposable income
may be acceptable in one country and intolerable in another.
In some countries gas prices are regulated at low levels to
stimulate substitution from other fuels to gas. This is common
Pinternational
Pdomestic
Pinternational
Pdomestic
Gas exports
becoming possible
Pinternational
Pdomestic
Pinternational
>1
<1
>1
<1
Gas price regulation that does not take costs fully into account
and involves a degree of subsidisation typically becomes harder
to sustain when international gas prices are high. This was the
situation in 2008. Importing country governments needed if
they wished to continue shielding their populations to accept
increasing budget deficits. Producer country governments that
could export the gas rather than keeping it at home had to accept
increasing growth in export and tax revenues foregone. The latter
governments were on the other hand typically also the biggest
beneficiaries of the 2008 oil price escalation and therefore able
to continue offering cheap gas to the domestic market.
In response to such pressures governments typically deregulate
prices to some market segments while retaining regulated prices
to other, more vulnerable segments.
Deregulation may be a long and cumbersome process as the
Higher cost
of new
domestic
production
Growing
dependence on
imported gas =>
Exposure to
world gas price
fluctuations
Time
Drive to shift
domestic fuel use
from oil to gas to
sustain oil exports
Price riots,
accommodating
leadership
Price
Supply =
Marginal Cost
Demand
P1
Average
Cost
V1
Volume
Supply =
Marginal Cost
Demand
P2
P1
Average
Cost
P3
V2
V1
V3
Volume
The gas price under oil price escalation will likely be above the
market-clearing price if oil prices are very high, and below if
oil prices are very low. Thus by summer 2008, when oil prices
were in the $120-130/bbl range, gas prices may have been
close to P2, while at low oil prices they could be around P3. If
oil prices are in the fuel-switching range, the oil indexed gas
prices will presumably be close to P1.
Bilateral monopoly negotiations were the dominant pricing
mechanism in interstate gas dealings in the former East Bloc
including the Former Soviet Union (FSU) and Central and
Eastern Europe. The gas price was determined for a period of
time typically one year through bilateral negotiations at
government level. There were often elements of barter with
the buyers paying for portions of their gas supply in transit
services or by participating in field development and pipeline
building projects.
The underlying valuation of the gas, the capital goods and the
services that changed hands in the intra-East Bloc gas trade was
opaque, with politics playing a major role alongside economics.
Examples of gas pricing based on bilateral negotiations may
still be found in countries where one dominant supplier, e.g.,
the national oil company, faces one or a couple of dominant
buyers, say, the state owned power company and maybe 1-2
large industrial companies. A number of immature developing
country gas markets have this structure.
Netback from final product means that the price received by the
gas supplier reflects the price received by the buyer for his final
product. For instance, the price received by the gas supplier
from the power sector may be set in relation to, and allowed to
fluctuate with, the price of electricity. Netback based pricing is
also common where the gas is used as a feedstock for chemical
production, such as ammonia or methanol, and represents the
major variable cost in producing the product.
This mechanism should not be confused with contractual
arrangements whereby the price to the producer/wholesaler
is netted back from the wholesale gas prices in countries
further downstream. A netback arrangement such as this would
be categorised depending on how the wholesale gas price in
the downstream country is determined through gas-on-gas
competition, oil price escalation, etc.
Direct gas price regulation remains widespread. It would however
be unhelpful to lump all kinds of regulation together. We need
to distinguish between the principles applied by the regulator.
Under cost of service based regulation the price is determined,
or approved, by a regulatory authority, or possibly a Ministry,
so as to cover the cost of service, including the recovery of
investment and a reasonable rate of return, in the same way as
pipeline service tariffs are regulated in the US. Normally, cost
of service based prices are published by the regulatory authority.
Pakistan provides an example of cost of service based prices,
with the wellhead price being the target.
Prices may also be regulated on an irregular social and political
basis reflecting the regulators perceptions of social needs and/or
gas supply cost developments, or possibly as a revenue raising
exercise for the government. In all probability the gas company
would be state-owned.
Many Non-OECD countries still practice below cost regulation,
meaning that the gas price is knowingly set below the sum of
production and transportation costs as a form of state subsidy to
the population. Again the gas company would be state-owned.
In some countries where a substantial proportion of indigenous
gas supply comes from oil fields with gas caps or gas-condensate
fields, the marginal cost of producing this gas may be close to
zero and as such it could be sold at a very low wholesale price
and still be profitable. However, to the extent it is sold below
the average cost of production and transportation it would still
be included in the regulation below cost category.
Supply =
Marginal Cost
Demand
P1
P5
Average
Cost
P4
V1
V5
V4 Volume
North America
US gas production has always involved a number of companies, and US gas prices have as a rule been determined competitively by supply and demand. For decades prices were
very low, reflecting producer competition for very limited
local markets. After World War 2 rapid expansion of the US
pipeline system enabled a gradual absorption of the surplus
reserves.
The Supreme Court Phillips Decision in 1954 ushered in a period
of wellhead price regulation that was to last for 24 years. The
regulation applied only to gas traded across state borders. Gas
produced and consumed in the same state was not affected by
the decision.
The wellhead price controls were of the historic E&D cost plus
type. They stimulated gas demand but not investment in the
upstream and eventually led to gas shortages in those parts of
the US that depended on other states for their gas supply. The
Natural Gas Policy Act of 1978 sought to fix the imbalance by
deregulating high cost gas prices while retaining most interstate
gas under price control and placing also intrastate gas under price
regulation so as to eliminate the particular shortage problems of
the importing states. These steps however paved the way for a
further dismantling of price controls in the years that followed.
Deregulation, and the impact of the first and second oil price
The UK
The UK gas industry was nationalised in 1948. The UK at
that time neither produced nor imported any natural gas.
However, there were more than 1000 manufactured gas companies some private, the other municipally owned that
were vested into 12 so-called area gas boards. In 1959 LNG
imports commenced on a trial basis. In 1964 the government
started to issue North Sea E&D licences. In 1965 the first
natural gas discoveries were made. In 1966 the government
decided to introduce natural gas into the UK fuel mix on a
big scale.
The 1972 Gas Act paved the way for further centralisation of
the industry with the creation of the British Gas Corporation
(BGC). This entity was until 1986 the sole buyer of UKCS
gas and the sole transmitter and distributor of this gas to UK
customers. It was also a key upstream player.
Wellhead prices were in these years set through negotiations
between BGC and the producers. BGCs legal monopsony on
UKCS gas purchases, and good grasp on upstream costs thanks
to its own UKCS interests, ensured prices that left little rent
to the producers.
The Thatcher years saw a general, ideologically driven shift
from state involvement through major public enterprises in the
economy, towards private solutions. The gas sector exemplified
this trend.
The 1982 Oil and Gas (Enterprise) Act permitted UKCS gas
producers and major industrial customers to contract directly
with each other, and ordered BGC to offer third party access
to its pipelines. These first steps towards a liberalisation of the
market failed to boost competition. The customers that producers
could now approach directly were too few, and BGCs grip on
the market remained too strong. The next steps were however
more forceful. The 1986 Gas Act returned the gas industry to
Continental Europe
The market value pricing principle that dominates in Continental Europe originated in the Netherlands. The Groningen
field discovered in 1959 and put on-stream in 1964 presented
the Dutch government with a marketing challenge. Western
European gas consumption in 1965 was about 21 bcm a year
. The Dutch themselves consumed a mere 1,8 Bcm a year2.
Continental European cross border gas trade was negligible.
Thus Groningen had to be sold into a small and immature
market area. The government did not want to sell the field
cheaply, thus giving away value. Delaying its development
seemed an equally unattractive option. There was a perception of urgency stemming from the emergence of a new source
of energy nuclear that conceivably could shorten the era
of fossil fuels.
In 1962 the then Dutch Minister of Economic Affairs suggested
to base prices not on production costs which were low for
Groningen gas and would have left the government with limited
revenues, but on the market or replacement value of the gas to
individual market segment in individual countries.
Specifically, the idea was that the price of Groningen gas to
a given customer should be based on the price of the best
alternative to Groningen gas typically heavy fuel or gas oil
for that customer.
The price of Groningen gas should not be mechanically aligned
with the price of the best alternative. On the one hand rebates
could be necessary to encourage customers that did not already
use Groningen gas to start doing so, and discourage existing
customers from switching back to competing fuels. The rebates
to attract new customers might need to be substantial if switching
would require investment in new heating systems. On the other
hand, due consideration should be paid to the convenience of
burning gas compared to oil products, potentially giving rise
to a price premium.
Since it is not possible to price discriminate at individual
2
While the market value principle placed the price risk in the
Groningen gas sales contracts with the seller, the take or pay
principle another feature of these contracts placed the
volume risk with the buyer. These provisions on risk sharing
paved the way for rapid growth in Dutch gas exports and for a
rapid maturation of European gas markets. The latter effect was
accentuated when Algeria, Russia and Norway adopted both
market value pricing and the TOP principle in their contracting
with European gas buyers. .
Asia Pacific
Japan was a 2 bcm a year gas market until 1970 when imported (Alaskan) LNG entered the fuel mix. Import growth
accelerated in the 1970s and 1980s in response to the first
and second oil price shock. South Korea and Taiwan started
to import LNG in 1986 and 1990 respectively. Australia and
New Zealand the two developed economies in the region
with indigenous gas reserves started to exploit these reserves around 1970.
The Asian countries that do not have significant domestic natural
resources and access to international pipeline networks and
underground storages like Europe and the US, have come to
rely almost 100% on imported LNG for their natural gas supply.
The largest importers, Japanese LNG buyers, are gas and power
companies carrying out business in an integrated manner,
from procurement and imports to transmission, distribution,
downstream gas and power supply and marketing. When they
first initiated discussions on potential LNG imports, they had
to emphasize long-term security of supply to make sure that
they would be able to fulfil their supply obligation to end-users.
At the same time, since LNG projects require enormous initial
investments on the sellers side, the latter needed security of
demand, meaning long-term and stable offtake by buyers. Sellers
and buyers thus had a common interest in long-term and stable
relationships. Commercial LNG projects have been developed
based on cooperative arrangements, and this is reflected in the
history of LNG pricing as well.
In 1969 when LNG was first imported into Japan, and through the
early 1970s, the price was fixed. This suited the suppliers since
they could recover their huge initial investment with certainty.
Fixed prices also enabled them to lock in the economics of their
relied on oil thermal power plants for 70% of their power supply.
Therefore, it was a reasonable decision for them to make LNG
pricing competitive against oil. For Japanese gas companies, the
main competing fuels were oil products such as kerosene for
heating and fuel oil for industrial use. Hence indexation to oil
was to an extent acceptable to them too. JCC is used as index
since it is calculated from data in Japan Exports & Imports
Monthly published by Japan Tariff Association, and therefore
can be considered a credible, transparent and neutral index.
In the 1990s, the generally low oil price environment caused
LNG suppliers to suffer from deteriorating project economics.
In response to suppliers call for a helping hand, a new pricing
mechanism with lower slopes at very low or very high oil
prices the so-called S-curve was introduced (Chart 5.2).
Later, when the LNG industry started to suffer from the impact
of sluggish demand related to the Asian currency crisis in the
late 1990s, some buyers obtained price floors and ceilings as
an extension of the S-curve mechanism.
Chart 5.2: LNG pricing with S-curve
kinkpoints
?
high oil
price zone
Oil price
$/BBL
?
low oil
price zone
Oil price
$/BBL
$/MMBTU
$6,00
$5,00
$4,00
$3,00
$2,00
$1,00
$0,00
OPE
GOG
BIM
NET
RCS
RSP
RBC
TOT
The only firm but also rather trivial conclusion that can be
made on the relationship between gas pricing model and gas
price level, is that a shift from subsidised to unsubsidised prices
will push prices up.
The long term impact of alternative pricing mechanisms on
gas supply and demand has been a hot topic in particular in
Europe. Observers, and the gas industry itself, in 2008 noted the
incongruence between the need for gas to remain the preferred
fuel to the power sector if we were to see further growth in
overall demand, and the disincentives that oil linked gas prices
at USD 100-150/b oil represented to the dispatching of existing
and the building of new gas power plants.
On the other hand, demand destruction first became a big issue
in the US following the gas price spike in 2005, and the supply
boosting impact of the post 2000 price gas price environment
is nowhere more obvious than in the US. So again it is not so
that one pricing model necessarily represents a bigger threat to
future gas demand, and a bigger encouragement to future gas
supply, than another.
Gas price volatility is generally seen to be a problem mainly for
actors in liberalised markets with gas-on-gas competition based
pricing. And indeed, Europes and Asias oil linked prices are
less volatile, reflecting the way the indices are defined. With
gas prices set to reflect the average of oil prices over a period
of time many months prior to delivery, short term peaks and
troughs are automatically smoothened out.
However, apart from the fact that price volatility to many actors
represents opportunities rather than problems, it might not be
very difficult to shape the price clause in a contract based on
gas indexation so as to obtain the same smoothening effect.
Continental Europe
40 %
1974: Oil
share 25%
30 %
20 %
10 %
0%
1974
1980
1990
2000
Gas
Oil
Coal
2004
Source: IEA
Asia Pacific
The changes that are occurring in Asian LNG import and gas
end user pricing are changes within the paradigm of oil linked
prices. As the Asian LNG market tightened, the gas priceoil
price curve steepened towards full parity in energy equivalence
terms between LNG and crude oil import prices. Also the S
shape of the curve that Japanese buyers prefer i.e., the ceiling
offering protection to the buyer if oil prices should increase above
a preset level and the floor offering protection to the seller if
oil prices should become too low came under pressure. The
financial crisis and the current outlook for slower growth in LNG
demand in a period when much new LNG will come on the
market, have reversed these trends but not affected the oil link.
However, the globalisation of the LNG business, the growth
in LNG spot transactions as a share of total LNG sales and
purchases (Chart 8.2) and in the future the emergence of LNG
transactions across the Pacific will shape Asian buyers pricing
habits too. Kogas uses the spot market to manage seasonal
swing in Koreas gas demand. As a result of several nuclear
incidents, since 2006 also Japanese buyers have been active
in the spot market. Japan in 2007 had to compete on price for
around 20% its total LNG supply. For the moment (1st quarter
2009) Asian buyers are not very active in the LNG spot market
but demand could bounce back once the financial crisis is over.
Asian buyers will then need to reckon with Henry Hub and the
NBP i.e., indirectly with supply and demand conditions in
North America and Europe as references that sometimes kick
in as floors, other times as ceilings.
Chart 8.2: Asian LNG importers spot purchases
25
Bcm
20
15
Taiwan
Korea
Japan
10
5
-
19
95
19
96
The Japanese gas and power utilities, Kogas and Taiwans CPC
have traditionally paid more than European and North American
buyers for their LNG imports. This is mainly because of their
traditional preoccupation with supply security and ability to pass
the costs of added security on to their customers. Japanese end
user prices, to take them as an example, have been regulated by
the Ministry of Economy, Trade and Industry on a cost plus basis.
Some of these companies campaigned for lower prices in the early
2000s, in response to Indias and Chinas successes in securing
20
06
20
07
50 %
Others
Hydro
Nuclear
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
70 %
60 %
Non OECD
In countries where gas end user prices are set below supply costs
and where the government is able to ensure that gas demand
growth is accommodated by supply growth, gas subsidisation
may increase to the point of representing a serious drain on
the budget. According to IEA estimates, gas subsidisation is
an issue for Iran, Russia, Ukraine, Kazakhstan, Pakistan and
Argentina in particular (Chart 8.3).
Chart 8.3: Energy subsidies by fuel in non-OECD countries, 2007
Energy subsidies by fuel in non-OECD countries, 2007
200
150
100
50
250
China
India
Dom.Rep.
200
Puerto Rico
Portugal
Greece
Turkey
Taiwan
South Korea
Germany
Belgium
Italy
USA
150
Bcm
250
Mexico
100
50
Spain
0
6 4 68 72 76 80 84 88 92 9 6 0 0 0 4
1 9 1 9 1 9 1 9 1 9 1 9 1 9 1 9 19 20 20
Japan
France
UK
Norway
Eq. Guinea
Egypt
Oman
Trinidad
Nigeria
Qatar
Australia
Malaysia
Indonesia
Abu Dhabi
Brunei
Libya
US
Algeria
19
6
19 4
6
19 8
72
19
7
19 6
8
19 0
8
19 4
8
19 8
92
19
9
20 6
0
20 0
04
Bcm
Source: Cedigaz
LNG contract prices are linked to the oil price, spot purchases
were apparently priced on an Atlantic basin netback basis,
though they could also reflect substitute fuel prices (usually in
Japan and usually distillate prices).
There were particular reasons for the Asian countries needs
for Atlantic LNG in 2007-08 in the case of Japan TEPCOs
temporary loss of big parts of its nuclear capacity, in the case of
South Korea a fuzzy regulatory situation that prevented Kogas
from signing new long term contracts, and in both cases poor
utilisation of storage tanks to manage seasonal demand and
Indonesias problems delivering on its commitments. Some of
these drivers will weaken, and the global recession has put an
end to the sellers market conditions that characterised LNG in
2007-08. In 2009 few Atlantic cargos have ended up with Asian
buyers. On the contrary, Asia Pacific exporters have needed to
place a few cargos with Atlantic buyers. These developments
do not constitute evidence that the integration of regional gas
markets has stopped in its tracks, but serve as a reminder that
the road towards globalised gas pricing may see set-backs and
could take longer than expected.
Bcm
45
43
41
39
37
35
Jan 97 - Oct 08
Trend line
47
ja
n.
97
ja
n.
98
ja
n.
99
ja
n.
00
ja
n.
01
ja
n.
02
ja
n.
03
ja
n.
04
ja
n.
05
ja
n.
06
ja
n.
07
ja
n.
08
51
49
Jan 97 - Oct 08
Bcm
open seasons for new regas terminals, that got the globalization
debate started.
Trend line
1
0
ja
n.
97
ja
n.
98
ja
n.
99
ja
n.
00
ja
n.
01
ja
n.
02
ja
n.
03
ja
n.
04
ja
n.
05
ja
n.
06
ja
n.
07
ja
n.
08
US LNG imports
7
6
2000
2500
2005
2006
2007
2008
2009
2000
Mcm
Tcf
1500
1000
03
20
06
20
09
20
12
20
15
20
18
20
21
20
24
20
27
20
30
Trinidad
Qatar
Norway
Nigeria
Eq. Guinea
Egypt
Algeria
ja
n.
06
ap
r. 0
6
ju
l.0
6
ok
t.0
6
ja
n.
07
ap
r.0
7
ju
l.0
7
ok
t.0
7
ja
n.
08
ap
r.0
8
ju
l.0
8
ok
t.0
8
20
20
00
500
97
19
3000
LNG prices could then decouple from the US price level which
if US gas demand and/or indigenous gas supply is flexible
enough to quickly accommodate any loss of flexible LNG to
other market regions might not change at all.
If the US instead develops the dependence on LNG that observers
in the early 2000s thought they could see around the corner,
but now tend to discard, US buyers would need to compete on
price with the rest of the world for LNG supply. Then the LNG
price ceiling provided by US indigenous gas supply costs could
disintegrate but we would still in this scenario characterised by
intercontinental competitive bidding see gas market integration
and price globalisation.
The differences between recent long term US LNG import
forecasts testify to the complicated nature of this issue. US gas
demand growth will play a key role, implying that economic
growth and the current administrations energy and environmental
policies will be important drivers. The exact shape of the North
American unconventional gas supply curve, today and 5, 10
and 20 years from now considering the resource base and the
scope for further technological progress, is another key to the
outlook for LNG. Whether incremental LNG supply costs will
stay at todays level or fall back towards their 2004 level is yet
another key.
9. Price volatility
General
3,00
2,50
USD/MMBtu
2,00
1,50
1,00
16
18
14
16
12
USD/MMBtu
12
10
8
6
0,00
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
10
Return(t) = ln(Price(t)/Price(t-1)).
8
6
4
13
.0
13 1.9
. 4
13 01.
.0 95
13 1. 9
.0 6
13 1. 9
.0 7
13 1.9
.0 8
13 1.9
.0 9
13 1.
. 0 00
13 1. 0
.0 1
13 1. 0
.0 2
13 1. 0
.0 3
13 1. 0
.0 4
13 1. 0
.0 5
13 1. 0
. 6
13 01.
. 0 07
1.
08
Ja
nJa 9 7
nJa 98
nJa 99
nJa 00
nJa 0 1
nJa 02
nJa 03
nJa 04
nJa 0 5
nJa 06
nJa 07
n08
USD/MMBtu
14
0,50
40 %
30 %
20 %
10 %
18
-20 %
16
14.01.08
14.01.07
14.01.06
14.01.05
14.01.04
14.01.03
14.01.02
14.01.01
14.01.00
14.01.99
14.01.98
14.01.97
-50 %
10
14.01.96
-40 %
12
14.01.95
-30 %
14
14.01.94
USD/MMBtu
0%
-10 %
4
2
0
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
In the 1980s and 1990s oil price volatility was much debated.
Many politicians and market actors recommended producer to
consumer cooperation to dampen price fluctuations. While the
oil price increases in 1973 and 1979-80 triggered consumer
country interest in this concept, the oil price collapse in 1986
persuaded many producer countries to support it too. The 1990
mini-shock related to the Iraqi invasion of Kuwait further
boosted enthusiasm for some kind of dialogue.
Economists however cautioned against politicising markets in
this way. One study5 examined the allegations that oil price
volatility had boosted inflation and dampened economic growth by:
Boosting oil prices
Reducing oil industry investments and thereby oil supply,
Boosting transaction costs e.g., costs associated with
investments in facilities to increase flexibility for consumers
and producers
The study failed to find conclusive proof for any of them. Price
volatility as such did not seem to be the reason for any of these
three situations.
Price volatility may keep investors that pursue low risk
activities with correspondingly low returns, and look for a
stable environment, from launching new investments. As such,
volatility may be an issue from a gas supply security point
of view. However, to other investors price volatility may, by
providing arbitrage opportunities, be seen as preferable to price
stability in terms of value added. It is important to nuance the
perception of volatility as a problem for the industry. It needs
to be acknowledged that different types of stakeholders look
for different price contexts.
This difference is related to the one between long term oil
indexed gas prices and shorter term gas to gas competition
based prices on gas exchanges.
D 21
D12
D 22
P2
B
P1
Volume
D2
S2
Causes of volatility
Many explanations have been offered for the perceived increase in
gas price volatility in the 2000s. Those that are most popular with
the media are not necessarily on top in terms of explanatory power.
Blaming fingers are pointed at commodity trading techniques
resulting from time to time in waves of speculative gas sales or
purchases. The public is also occasionally fired up by reports
on downright market manipulation. However neither trading
techniques nor criminal activity are credible explanations for
a general increase in price volatility.
Basic gas supply and demand fundamentals go a long way
towards explaining this increase.
Price volatility is the consequence of supply failing to respond
immediately, smoothly and precisely to price signals caused by
5
P2
P1
S1
If the market is
at A when
supply
shifts
from S1 to S2,
then if demand
is flexible (the
D1
curve)
prices
only
increase
by
P1, whereas
if demand is
inflexible (the
D2
curve)
prices increase
by P 2.
Volume
North America, UK
u ro pe
ental E
Con tin
Continental Europe,
Developed Asia
2020
Gas-on-gas competition
Oil price escalation
-O EC
t N on
Selec
Select Non-OECD
No price
Ch
ina
Bilateral monopoly
Ru
ss
ia,
Bilateral monopoly
Select Non-OECD
USD/MMBtu
20,00
15,00
10,00
5,00
0,00
2007
2010
2015
2020
2025
2030
The possibility of oil linked gas falling out of favour with the
key power sector is particularly worrisome. Here gas needs to be
perceived as competitive with coal and in the future increasingly
with biomass, wind, solar, etc. The competition from coal is
blunted by differences in capital costs, lead times, taxation and
regulatory provisions. The competition from renewables other
than hydro is blunted by the still high costs of these options.
Extended oil driven gas price rallies could still erode gas
position as the preferred fuel.
Industrial buyers benefit from oil indexation when oil prices
are sufficiently low for sufficiently long to make oil linked
gas cheaper than spot gas. Sellers of course benefit from the
opposite situation. Oil market cycles in combination with price
renegotiation clauses in long term contracts may deliver a
balanced distribution of costs and benefits over time. Oil driven
gas price rallies like the one in 2007-08 that led to significant
industrial demand destruction are nevertheless bad for gas
image as a reliable and affordable fuel across cycles.
More gas-on-gas competition and more use of gas exchange
prices would to an extent decouple gas prices from oil prices. It
would however increase short term price volatility, and whether
it would eliminate the risk of longer term price rallies is an open
question. Basically that would depend on Continental Europes
and Developed Asias future gas supply-demand balances.
For the moment there is ample spare capacity in Europes
pipeline gas supply chains as well as in the worlds LNG supply
system. The financial crisis, the recession and the consequent
drop in gas demand nearly everywhere have forced gas suppliers
to significantly lower capacity utilisation. Sharp declines
in sales revenues and doubts about the timing and shape of
the anticipated recovery are however delaying vital up- and
midstream investments. The IEA and others are concerned that
the current global gas market downturn will only pave the way
for another rally.
Evidence from North America underlines the question mark at
the long term consequences for gas prices of switching from oil
escalation to gas-on-gas competition. Although gas prices are
not in any way linked to oil prices in US contracts, gas has over
the years across frequent, sometimes violent short-medium
term disturbances tended to track oil in a fairly stable long
term relationship. This is probably because gas and oil prices
besides being linked by interfuel competition in the industrial
sector are influenced in the same manner and to the same extent
by the oil and gas industrys cost cycles, and with deviations also
being arrested, eventually, by changes in oil and gas industry
investment priorities.
Oil indexation will in any event not disappear any time soon,
for several reasons.
350
300
bcm
250
200
150
100
140
50
120
100
bcm
2025
2024
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
80
60
40
This is not to say that there is a desire among gas sellers and
buyers to get rid of the oil link overnight even if they could.
As noted, the incumbents on both sides of the table seem for
the moment to be broadly in favour of retaining oil indexation.
The EU Commission will likely continue to push for gas-on-gas
competition based pricing, but it cannot push very hard in the
absence of trading places offering reliable price information
and the full range of trading facilities and services. Continental
Europes gas hubs will take on these characteristics and functions
but that will take time.
Japanese, South Korean and Taiwanese gas importers have on
balance been even more hesitant than their Continental European
counterparts to switch from oil indexed import prices and cost
plus based domestic prices to more competitive arrangements.
6
20
2025
2024
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
Including deals at HoA or MoU level as well as firm sales and purchase contracts
North America, UK
ro pe
enta l Eu
Con tin
Continental Europe,
Developed Asia
Bilateral monopoly
Netback from final product
Regulation cost of service
Regulation social and political
Regulation below cost
No price
2020
Gas-on-gas competition
Oil price escalation
Bilateral monopoly
Ru
ss
ia
,C
hi
na
?
Select Non-OECD
Select
EC
N on-O
Select Non-OECD
Appendix 1
Price Formation Mechanisms 2005 Survey
Format of Results
World Results
Asia Pacific
10 %
North America
27 %
Middle East
10 %
Former Soviet
Union
21 %
Latin America
5%
Europe
19 %
2,785 bcm
Asia
Africa 5 %
6%
Asia Pacific
8%
663 bcm
North America
26 %
Middle East
0,9 %
Africa
6,7 %
Asia
1,8 %
Asia Pacific
1,0 %
North America
18,8 %
Latin America
2,6 %
Middle East
11 %
Latin America
5%
Europe
11 %
Former Soviet
Union
28 %
With respect to imports by pipeline (both intra- and interregional), Europe accounts for more than half of the world total.
Both European intra-regional gas imports (Norway to various
countries) and Europes imports of gas from outside Europe
(Russia and Algeria) are very significant. In the other regions,
pipeline imports are all intra-regional.
Former Soviet
Union
44,8 %
Europe
23,4 %
Latin America
0,5 %
Europe
25,2 %
2,785 bcm
Asia
3,2 %
Asia
Africa 5 %
6%
Middle East
11 %
Asia Pacific
8%
North America
26 %
Latin America
5%
Former Soviet
Union
28 %
Europe
11 %
With respect to gas exports via pipeline, the Former Soviet Union
in 2005 accounted for some 44% of the world total. Africa,
meaning in this case Algeria, is also a significant exporter to
Europe, while any trade in the Asian and American regions is
intra-regional.
Bilateral monopoly
22,7 %
North America
1,0 %
Gas-on-gas
competition
22,4 %
Latin America
7,4 %
Middle East
23,0 %
Asia Pacific
44,5 %
Pipeline imports at 660 bcm account for some 22% of total world
consumption. Three categories account for internationally
traded pipeline gas OPE almost all in Europe; GOG in North
America with small amount in Europe into UK and BIM almost
all intra-Former Soviet Union trade.
Africa
24,1 %
1,940 bcm
No price
1,7 %
Regulation below
cost
33,6 %
Not known
0,4 %
Gas-on-gas
competition
13,4 %
Bilateral monopoly
3,7 %
Gas-on-gas
competition
36,3 %
Regulation
social/political
15,6 %
Regulation cost of
service
3,6 %
Netback
0,7 %
Bilateral monopoly
3,7 %
Regional Results
Bilateral monopoly
18,5 %
Gas-on-gas
competition
20,4 %
Total imports at 850 bcm account for some 30% of total world
consumption. 60% is OPE with Europe (pipeline mainly) and
Asia Pacific (LNG) dominating. GOG is both pipeline and
LNG imports, with BIM largely intra-Former Soviet Union
pipeline trade.
Price Formation: Total Consumption
Chart A11: World price formation 2005 total consumption
World price formation 2005: Total
consumption
2,790 bcm
Regulation below
cost
23,3 %
No price
1,2 %
Not known
0,3 %
Consumption
Production
629.8
91.4
47.6
768.8
511.8
185.9
39.2
736.9
Imports
Pipeline
LNG
104.2
17.9
10.1
10.1
124.5
17.9
Exports
Pipeline
LNG
20.3
1.8
104.2
0.0
124.5
1.8
Regulation
social/political
10,9 %
Regulation cost of
service
2,5 %
Country
USA
Canada
Mexico
Total North America
Gas-on-gas
competition
31,4 %
Netback
0,5 %
No price
1,2 %
Bilateral monopoly
8,2 %
Gas-on-gas
competition
98,8 %
Country
Consumption
Production
40.4
2.1
19.9
8.5
6.8
0.3
0.3
1.5
0.7
16.3
0.1
28.9
125.7
45.6
12.4
11.4
2.0
6.8
Imports
Pipeline
LNG
1.7
8.8
6.5
0.3
0.3
1.6
30.3
28.9
139.2
Exports
Pipeline
LNG
6.8
10.4
0.7
14.0
0.1
17.2
0.9
17.2
14.0
Regulation
social/political
59,5 %
Argentina
Bolivia
Brazil
Chile
Colombia
Dominican Republic
Ecuador
Peru
Puerto Rico
Trinidad
Uruguay
Venezuela
Total Latin America
Europe
Netback
9,5 %
Regulation cost of
service
5,4 %
Austria
Belgium & Luxembourg
Bosnia-Herzegovina
Bulgaria
Croatia
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Latvia
Lithuania
Netherlands
Norway
Poland
Portugal
Romania
Serbia & Montenegro
Slovakia
Slovenia
Spain
Sweden
Switzerland
Turkey
United Kingdom
Total Europe
Consumption
Production
10.0
16.6
0.4
3.0
2.7
8.5
5.0
1.5
4.0
45.8
86.2
2.8
13.2
3.9
78.7
1.8
3.3
39.5
4.5
13.6
4.2
17.3
2.2
6.6
1.1
32.4
0.8
3.1
26.9
95.1
534.6
1.6
1.5
0.2
10.4
1.2
15.8
3.0
0.6
12.1
62.9
85.0
4.3
12.1
0.3
0.2
0.2
0.9
87.5
299.7
Imports
Pipeline
LNG
8.7
18.0
3.0
0.4
2.9
1.2
9.5
0.7
4.2
36.2
90.7
2.3
10.8
3.1
71.0
1.2
2.9
23.0
10.2
2.6
6.3
1.9
6.4
1.1
11.6
1.0
2.8
22.2
14.7
367.4
12.8
0.5
Exports
Pipeline
LNG
4.4
5.3
9.8
2.5
46.8
79.5
1.6
21.9
4.9
0.5
47.6
9.7
155.4
0.0
124 bcm
Regulation cost of
service
1,2 %
Regulation
social/political
11,6 %
Netback
0,6 %
Bilateral monopoly
2,0 %
Gas-on-gas
competition
47,0 %
Bilateral monopoly
1,5 %
Consumption
Production
1.7
8.9
18.9
1.5
19.6
0.7
2.5
405.1
1.4
16.6
72.9
44.0
593.8
5.3
0.3
0.2
23.3
0.0
0.1
598.0
0.0
58.8
19.4
55.0
760.5
Armenia
Azerbaijan
Belarus
Georgia
Kazakhstan
Kyrgyzstan
Moldova
Russian Federation
Tajikistan
Turkmenistan
Ukraine
Uzbekistan
Total FSU
Imports
Pipeline
LNG
1.7
4.5
20.1
1.5
11.6
0.7
2.5
25.6
1.4
0.0
55.3
0.0
124.8
0.0
Exports
Pipeline
LNG
7.6
229.0
45.2
2.5
12.4
296.7
0.0
The situation for total imports (both pipeline and LNG, comprising
415 bcm or 78% of total consumption) is markedly different,
with OPE dominating at 92%. The small amount of GOG (6%)
is predominantly the UK, plus Ireland and a small amount in
the Netherlands. The BIM category (2%) is accounted for by
imports into the Baltic States (Estonia, Latvia and Lithuania)
from Russia.
No price
2,8 %
Regulation below
cost
65,1 %
Bilateral monopoly
29,1 %
Regulation
social/political
3,0 %
At 595 bcm the Former Soviet Union accounts for just over
20% of world consumption. All imported gas is priced on a
BIM basis, together with some Russia domestic production
sold to large users. The dominant price formation mechanism,
however, is RBC in Russia, Uzbekistan and Kazakhstan. Since
2005, however, this situation in Russia, at least, is likely to have
changed with increased prices to domestic consumers raising
levels above the average cost of production and transportation.
Domestic production in Ukraine is the RSP category and NP
in Turkmenistan.
Middle East
Table A5: Middle East consumption and production 2005 (BCM)
Country
Bahrain
Iran
Iraq
Israel
Jordan
Kuwait
Oman
Qatar
Saudi Arabia
Syria
United Arab Emirates
Total Middle East
Consumption
Production
10.7
102.4
2.5
0.7
1.6
12.3
9.2
18.7
71.2
6.1
41.3
276.6
10.7
100.9
2.5
0.7
0.3
12.3
19.8
45.8
71.2
5.4
47.0
316.6
Imports
Pipeline
LNG
5.8
No price
1,1 %
1.3
1.4
1.4
8.5
0.0
5.7
275 bcm
Regulation
social/political
14,8 %
Afghanistan
Bangladesh
China
China Hong Kong
India
Myanmar
Pakistan
Total Asia
Regulation below
cost
80,4 %
Consumption
Production
0.2
14.2
45.7
3.1
38.1
4.1
29.3
134.7
0.2
14.2
50.0
32.1
13.0
29.3
138.8
Imports
Pipeline
LNG
3.1
3.1
Exports
Pipeline
LNG
3.1
6.0
8.9
6.0
12.0
0.0
Algeria
Angola
Egypt
Equatorial Guinea
Ivory Coast
Libya
Nigeria
South Africa
Tunisia
Total Africa
Regulation cost of
service
32,6 %
Regulation
social/political
11,2 %
Country
Country
Netback
1,2 %
Regulation below
cost
48,2 %
7.1
43.5
9.2
27.1
No price
1,3 %
75 bcm
4.3
Not known
0,9 %
Exports
Pipeline
LNG
Consumption
Production
23.2
0.8
25.8
1.3
1.3
5.8
10.4
2.2
4.3
75.1
88.2
0.8
34.6
1.3
1.3
11.3
22.4
2.2
2.5
164.6
Imports
Pipeline
LNG
1.8
1.8
0.0
Regulation below
cost
3,0 %
Exports
Pipeline
LNG
39.1
25.7
1.1
6.9
4.5
0.9
12.0
44.7
45.5
Regulation
social/political
57,9 %
Excluding its export trade, Africa has virtually not traded gas,
with only Tunisia importing some gas from Algeria via the
pipeline to Italy.
Regulation cost of
service
21,8 %
Wholesale Prices
Asia Pacific
Country
Consumption
Production
26.8
2.4
37.5
79.0
39.3
3.5
3.0
6.6
33.7
10.7
29.9
6.9
279.3
40.3
11.5
73.8
5.1
59.9
3.8
2.9
Australia
Brunei
Indonesia
Japan
Malaysia
New Zealand
Philippines
Singapore
South Korea
Taiwan
Thailand
Vietnam
Total Asia Pacific
0.5
0.8
23.7
6.9
229.2
Imports
Pipeline
LNG
76.3
6.6
8.9
15.5
Exports
Pipeline
LNG
14.9
9.2
4.8
31.5
1.8
28.5
30.5
9.6
Asia
Africa
116.4
6.6
84.0
Latin America
Former Soviet Union
Regulation
social/political
24,8 %
Oil price
escalation
50,4 %
Regulation cost of
service
2,9 %
Bilateral monopoly
8,1 %
Gas-on-gas
competition
11,8 %
Middle East
$0,00
$1,00
$2,00
$3,00
$4,00
$5,00
$6,00
$7,00
$8,00
$9,00
$/MMBTU
2,790 bcm
Regulation below
cost
23,3 %
World Average
$1,00
$2,00
$3,00
$4,00
$5,00
$6,00
$7,00
$8,00
Regulation cost of
service
2,5 %
$9,00 $10,00
$/MMBTU
North America
Latin America
Europe
Former Soviet Union
Middle East
Africa
Asia
Asia Pacific
Total World
Bilateral monopoly
8,2 %
OPE
0.0
18.7
426.6
0.0
0.0
4.3
14.2
140.8
604.6
22%
GOG
759.4
2.0
83.4
0.0
0.0
0.0
0.0
33.0
877.9
31%
Total Consumption
BIM NET RCS RSP RBC NP
0.0
0.0
0.0
0.0
0.0
9.4
8.4 11.9 6.8 74.8 3.1
0.0
8.7
0.7
1.7 14.4 0.0
3.5
172.9 0.0
0.0 17.6 386.6 16.6
7.2
0.0
0.0 40.6 221.5 3.6
0.0
0.9 24.5 8.4 36.2 0.8
9.1
0.0 29.3 78.0 4.1
0.0
22.6 0.0
8.0 69.3 0.0
0.0
228.9 13.5 70.3 303.2 651.4 33.8
8%
0%
3% 11% 23% 1%
NK
0.0
0.0
0.0
0.0
2.5
0.0
0.0
5.6
8.1
0%
TOT
768.8
125.7
539.1
593.7
275.3
75.1
134.7
279.3
2,791.7
100%
T
TO
t
be
lo
w
co
s
lit
ic
al
Netback
0,5 %
eg
ul
at
io
n
so
c
Gas-on-gas
competition
31,4 %
The chart above illustrates the overall results at the world level,
while the table looks at the breakdown by region.
eg
ul
at
io
n
co
s
io
n
eg
ul
at
ia
l/p
o
er
vi
ce
ck
to
fs
et
ba
N
op
ol
y
Bi
la
te
ra
l
m
on
tit
io
n
as
-o
nga
co
m
pe
es
ca
la
tio
n
$/MMBTU
il
pr
ic
O
$9,00
$8,00
$7,00
$6,00
$5,00
$4,00
$3,00
$2,00
$1,00
$0,00
Not known
0,3 %
Regulation
social/political
10,9 %
$0,00
No price
1,2 %
Conclusions
In 2005 just over 70% of the worlds consumption of gas
comprised of domestic production consumed within that country,
with no trade across international borders. Some 22% was
traded through pipelines and some 6% LNG. The wholesale
price formation mechanisms are largely very different for
internationally traded gas compared to gas which is produced
purely for domestic consumption.
T
TO
t
be
lo
w
co
s
lit
ic
al
R
eg
ul
at
io
n
so
c
eg
ul
at
io
n
co
s
io
n
eg
ul
at
R
ia
l/p
o
er
vi
ce
ck
to
fs
et
ba
N
op
ol
y
m
on
tit
io
n
te
ra
l
Bi
la
co
m
pe
s
nga
as
-o
G
il
pr
ic
es
ca
la
tio
n
$/MMBTU
$9,00
$8,00
$7,00
$6,00
$5,00
$4,00
$3,00
$2,00
$1,00
$0,00
IGU
The International Gas Union is a worldwide non-profit
organisation aimed at promoting the technical and
economic progress of the gas industry. The Union has
more than 100 members worldwide on all continents.
The members of IGU are national associations and
corporations of the gas industry. IGUs working organisation covers all aspects of the gas industry, including
exploration and production, storage, LNG, distribution and natural gas utilisation in all market segments.
IGU promotes technical and economic progress of
the gas industry emphasising environmental performance worldwide. For more information, please visit
www.igu.org
Telephone:
Fax:
Email:
Website:
IGU members
110352 IGU 2011, Photo: Eiliv Leren, Harald Pettersen, Helge Hansen, yvind Hagen
+ 47 51 99 00 00
+ 47 22 53 43 40
secrigu@statoil.com
www.igu.org
Sponsored by:
PROGRAMONENERGYAND
SUSTAINABLEDEVELOPMENT
WorkingPaper
#94
August2010
T HE
W ORLD S G REATEST C OAL A RBITRAGE : C HINA S C OAL
I MPORT
B EHAVIORAND I MPLICATIONSFORTHE G LOBAL C OAL
M ARKET
RICHARDK.MORSEANDGANGHE
FREEMANSPOGLIINSTITUTEFORINTERNATIONALSTUDIES
Photocredit:Reuters
The Program is part of the Freeman Spogli Institute for International Studies at Stanford University.
PESD gratefully acknowledges substantial core funding from BP.
Richard K. Morse leads global coal market research at the Stanford Program on Energy and
Sustainable Development (PESD). PESD's coal research examines the political economy of coal,
the international trade and pricing of coal, and coal's long term role in the world's energy mix.
Other research interests include carbon policy and carbon markets, carbon capture and storage,
renewable energy markets, and financial markets for energy commodities.
Richard received a B.A. in philosophy from Rice University, where he was awarded the James
Street Fulton Prize for the top graduate in the field. He has worked in commodities markets for
oil, natural gas, renewable energy, and emissions as an energy analyst and trader.
Introduction
In 2009 the global coal market witnessed one of the most dramatic realignments it has ever seen
China, long a net exporter of coal, suddenly imported a record-smashing 126 Mt tons (see
Figure 1). 1 This inversion of Chinas role in global coal markets meant that Chinese imports
accounted for nearly 15% of all globally traded coal, and China became the focal point of global
demand as traditional import markets like Europe and Japan stagnated in the wake of the
financial crisis. By the first quarter of 2010, even Colombia was defying established trade
patterns by sending cargoes to China despite its massive geographic disadvantage to export coal
into Asian markets. The middle kingdoms appetite for imported coal seems insatiable, and the
China Factor appears to have ushered in a new paradigm for the global coal market.
But China doesnt need the coal. The worlds largest coal producer cranked out 2.96 Bt of
production in 2009, backed up by 114.5 Bt of reserves. 2 While the worlds other fastest growing
importer, India, is plagued by a growing gap between coal supply and power demand that it is
unable to fill domestically, this is not the case in China. The spike in Chinese demand for
imported coal is therefore a more complex (and less easily predictable) phenomenon that
requires careful examination if the world is to understand what impact China might have on
global energy markets in the coming decade.
5 August 2010
In this paper we outline a model that explains Chinese coal import patterns and that can allow the
coal market to understand, and to some degree predict, Chinas coal import behavior. We argue
that the unique structure of the Chinese coal market creates a series of key arbitrage relationships
between Chinese domestic coal markets and international coal markets that determine Chinese
import patterns. Based on this theory of Chinas import behavior, we construct an arbitragebased import model that explains the dramatic shift in Chinese net imports over the last several
years.
The implications of this model are significant for the development of the global coal trade in the
coming decade. First, we find that Chinas import behavior does not represent a structural
shift in global markets. China, as a participant in the global coal market, is a cost-minimizer
that will be both a buyer and seller in the global market as key price relationships fluctuate.
Second, and perhaps most importantly, the arbitrage relationships that we describe directly link
the domestic price of coal in China to the global price of coal. This linkage has significant
implications for the use of coal and the cost of generating power globally in the coming decades.
Developments in Chinas domestic coal market will be a dominant factor determining global
coal prices and trade flows (and by implication power prices in many regions). This makes
understanding the domestic Chinese coal market, which operates according to a unique economic
and political logic, crucial for any participant in the global markets.3
A series of forthcoming studies on the structure and long term future of the Chinese coal market will be release by
PESD Stanford in 2010, and can be found at http://pesd.stanford.edu
5 August 2010
MonthlyMillionMetricTons
18
16
14
12
10
8
6
4
2
0
Imports
Total
Exports
Jan10
Jul09
Jan09
Jul08
Jan08
Jul07
Jan07
Jul06
Jan06
Jul05
Jan05
Jul04
Jan04
Jul03
Jan03
Jul02
Jan02
Figure 1 Chinese net imports were negative until 2008, when historical trade balances inverted
dramatically. Source: McCloskey.
Chinas coal reserves and production are concentrated in the North and West of China. Three
provinces in these regions Shanxi, Shaanxi, and Inner Mongolia have 69% of the countrys
proven reserves and were home to half of national production in 2009. That same year Inner
Mongolia surpassed Shanxi to become the largest producer at 637 Mt; Shanxi produced 615 Mt,
and Shaanxi produced 296 Mt. 4 70% of the production in these three provinces is exported
outside of its home province to supply coal demand most heavily concentrated along the eastern
and southern coasts. Figure 2 illustrates the basic geography of Chinese coal regions.
Northern coastal Chinese coal demand is served by a network of truck routes and railways that
move coal east and south from western and northern production centers. But rail and truck
capacity to supply coal to Southeast China is both insufficient and prohibitively expensive.
Therefore coal supply for Southeast China is first transported east on rail lines like the Da-Qin
and Shuo-Huang to eastern ports like Qinhuangdao, Huanghua, Rizao, next loaded onto boats,
and finally shipped south via sea routes. Figure 3 illustrates Chinas major coal transport
infrastructure segments. This rail-to-sea link is still much cheaper than moving coal overland
5 August 2010
from North to South. Though the costs associated with this transport route are still high; the
transportation cost of moving coal from Shanxi to Guangzhou can be as high as 50-60% of the
price of coal delivered to Guangzhou. 5 The high cost of moving coal to the heavily industrialized
coastal area that includes the Pearl River Delta and the Yangtze River Delta opens windows for
import coals to compete with domestic coals.
Southeast China is also the closest region in China to two major global coal exporting nations,
Indonesia and Australia. Coal buyers in Southeast China therefore are often confronted with two
options: buy domestic coal delivered by sea from Northern Chinese ports, or buy international
coal. This arbitrage opportunity allows Chinese coal buyers to take advantage of price
differentials between domestic Chinese coal and international coal prices. Until 2009, those
differentials had not favored imports.
Figure 2 Map of Chinese coal planning regions, as described by NDRCs Coal Industry Policy of
2007. Source: Kevin Tu.
This is an estimate. Costs fluctuate according to the price of coal and the price of shipping coal.
5 August 2010
Figure 3 Schematic of Chinas major coal transportation infrastructure. Source: Kevin Tu.
We argue that modeling arbitrage spreads between domestic and international coals for coal
buyers in Southern China explains Chinas 2009 import spike and can also be used to reliably
analyze Chinas national import behavior under future market conditions. In this section we
describe the parameters of our model, called the China Coal Import Arbitrage Model
(ChinaCoalArb for short), and demonstrate how its results can be used to interpret Chinas
import trends.
Chinese coal buyers in Southeastern China can buy coal from multiple markets, and price
discrepancies between different markets create profit opportunities. For a portion of spot market
demand, buyers will compare the CIF 6 cost of coal landed in Guangzhou from multiple
F
destinations and, all other things equal, will take the cheapest coal. 7 The differentials between
CIF Guangzhou coal prices from multiple origins therefore create arbitrage opportunities for
CIF is a coal market term indicating that a price is for the delivered location, and thus includes all freight costs.
For purposes of modeling we have slightly simplified the dynamics of Chinese coal buying behavior to focus on
the domestic vs. import tradeoff. For instance, we have not included term contracts or prices at port stockpile in our
model.
5 August 2010
Chinese buyers who can shift their purchasing patterns to capture the differential between
domestic and international markets under different conditions. The model calculates these
arbitrage relationships of domestic to key international coals. We then compare these arbitrage
relationships to historical imports, demonstrating that import levels have broadly tracked these
arbitrage trends, increasing where price spreads favored international coals over domestic coals.
3.1
Chinese domestic prices in ChinaCoalArb are represented by FOB 8 prices at the Qinhuangdao
port. Qinhuangdao port is mainly supplied with coal from Shanxi, Shaanxi and Inner Mongolia
and is Chinas largest coal port. Qinhuangdao throughput in 2009 was 206.33 Mt, and the total
throughput of the seven major coal ports serving Northern China was 433Mt. 9,10 Coal loaded in
Qinhuangdao and delivered to Guangzhou is transported south down the Chinese coast via
maritime shipping. Figure 4 below shows the historical development of prices at Qinhuangdao.
We select three key international coal supply countries as the models suppliers. Australia,
Indonesia, and Russia were the largest exporters of coal to China in 2009. 11,12 All three
countries are major exporters to the international market and benefit from reasonably proximity
to Chinese import markets. Table 1 shows 2009 total Chinese imports from these origins. FOB
coal prices in each of these markets are derived from bids and offers at those locations. 13 Figure
4 shows the historical development of Russian, Indonesian, and Australian export coal prices
from 2005.
FOB is a shipping term meaning Free On Board, and in the coal market it indicates the price quoted for coal
loaded on the vessel at the port of origin.
9
Qinhuangdao Port, China Coal Transportation and Distribution Association.
10
These ports include: Qinhuangdao, Tangshan, Huanghua, Tianjin, Rizhao, Lianyungang, and Qingdao.
11
China has historically imported coal from Vietnam, but due to increased Vietnamese domestic consumption and
price increases imports from Vietnam are expected to decline.
12
Russian imports are not necessarily delivered into Southern China as the port of Vostochny is north of the Chinese
border. But even though Russian ports are north rather than south, the general arbitrage principal applies and thus
we have included Russia in our model even if Russian material is not always imported to Guangzhou. Russian
exports to China are still comparatively small, but increased from 0.76Mt in 2008 to 11.8 Mt in 2009.
13
There are multiple price indices used in the coal market. We have used here indices provided by McCloskey and
Reuters.
5 August 2010
USD/metricton
180
Qinhuangdao(6,000kcal/kgNAR)
160
Newcastle(6,700kcal/kgGAD)
140
VostochnnyRussia(6,700kcal/kgGAD)
120
KalimantanIndonesia(4,900kcal/kg
NAR)
100
80
60
40
20
0
Apr10
Jan10
Oct09
Jul09
Apr09
Jan09
Oct08
Jul08
Apr08
Jan08
Oct07
Jul07
Apr07
Jan07
Oct06
Jul06
Apr06
Jan06
Oct05
Jul05
Apr05
Jan05
Australia
Indonesia
Vietnam
Russia
Mongolia
126.491 Mt
44.602 Mt
30.461 Mt
23.932 Mt
11.787 Mt
6.002 Mt
Canada
North Korea
USA
South Africa
New Zealand
Other
4.093 Mt
3.599 Mt
.805 Mt
.732 Mt
.303 Mt
.145 Mt
5 August 2010
3.2
The port of Guangzhou in Guangdong province functions as the demand center of our arbitrage
model. Guangdong is a heavily industrialized zone that has historically been the largest coal
importing province. Coal imports by province for 2009 are shown in Table 2. Guangzhou
ports coal handling capacity reached 56.5Mt/year in 2008. Nearby import centers witness
similar price relationships between domestic and international markets and exhibit import
patterns. Arbitrage relationships for Guangzhou can therefore be used as a proxy for arbitrage
relationships for all of Southeastern China. 14 Other major coal ports of Southeastern China
include Shanghai, Ningbo, Fuzhou, Xiamen, Quanzhou, Shantou, and Beihai ports.
Table 2 China coal import by provinces, 2009
14.14
33.72
Fujian
Zhejiang
Shanghai
All
other
Total
12.00
10.26
3.13
52.75
126
3.3
Freight Prices
The delivered cost of coal in Guangzhou (CIF) is calculated by adding freight costs between
loading and discharge ports to the FOB coal cost. 15 Dry bulk freight rates from Indonesia,
Russia, or Australia into China largely track international dry bulk freight markets which are
volatile, internationally traded commodity markets (see historical freight rates in Figure 5). 16
The model uses specific freight prices quoted from each FOB loading port and delivered into
Guangzhou port provided by AXS Marine.
14
The basis difference between Guangzhou and other cities in Southeast China will be roughly the freight costs
differential between those two locations. Thus while Guangzhou is the center of coal imports in China and is
indicative of market conditions for imports in general, some minor basis differentials will exist for other locations.
15
There are a few other adjustments which we describe later.
16
The Baltic Dry Index, which is widely used to indicate the cost of chartering dry bulk freight vessels, reached
highs of 11,459 in 2008 before crashing to lows near 670 in late that year. The Index is comprised of charter rates
for four types of ships: capsize, panamax, supramax, and handysize. For further details see the Baltic
Exchange: http://www.balticexchange.com/default.asp?action=article&ID=1
A price history is available from Bloomberg:
http://www.bloomberg.com/apps/quote?ticker=BDIY&exch=IND&x=15&y=11
5 August 2010
Freight prices for the Chinese maritime coastal shipping market are not fully connected to
international freight prices, however. Smaller boats are typically dedicated to the domestic
market, though larger boats can switch into the international shipping markets when prices are
attractive. 17 The Shanghai Shipping Index measures the cost of sending coal from Qinhuangdao
to Guangzhou (see Figure 3). 18 Model freight price assumptions assume capesize vessels for all
international shipping routes, and 40-50,000 dead weight ton (DWT) vessels for Chinese coastal
shipping prices. Representative port handling charges for loading and discharge are added for all
ports. 19
NewcastletoGZO
50
QinhuangdaotoGZO
USD/metricton
40
KalimantantoGZO
VostochnytoGZO
30
20
10
0
Apr10
Feb10
Dec09
Oct09
Aug09
Jun09
Apr09
Feb09
Dec08
Oct08
Aug08
Jun08
Apr08
Feb08
Dec07
Oct07
Aug07
Jun07
Figure 5 Dry bulk freight rates from FOB ports to Guangzhou (GZO) port in ChinaCoalArb. Source:
AXS Marine, Reuters, authors analysis. Rates are based on historical quotes for specific shipping routes.
All international quotes are based on capesize vessels, China domestic quotes are based on 40-50,000 dwt
vessels.
17
We benefited from discussion with Jon Windham of Macquarie Securities who offered his views of the Chinese
domestic shipping market and generously shared his research and data.
18
More information on the Shanghai Shipping Index is available at http://en.chineseshipping.com.cn/html/index.asp.
19
We assign port fees based on rates provided by AXS Marine.
5 August 2010
3.4
We make the following adjustments in order to more accurately reflect real market conditions:
First, coal buyers are buying energy. We therefore adjust all FOB prices to an energy equivalent
USD/metric ton basis of 6,700 kcal/kg gross air dried (GAD). 20
FOB indices are shown in Figure 4 (above). Energy equivalent prices are shown in Figure 6
(below).
Second, relevant taxes are added to all coal prices. Chinese VAT of 17% is added to the CIF
price of all coals. We do not adjust for Chinas import tax because it has been phased out by
2007, before Chinas import surge. 21
Third, all coal and freight prices are adjusted for historical exchange rates between RMB and
USD.
Fourth, Chinese buyers face increased transaction costs when purchasing coal from the
international market as compared to the domestic market. 22 Transaction costs for Chinese
buyers associated with import vs. domestic coal include raising letters of credit (LCs), dealing
with foreign sellers and more onerous contracts. Though it is impossible to know precisely the
increased transaction costs for individual buyers, we add a $3/ton advantage to Qinhuangdao
coal over all international coals. 23
20
We assume a linear relationship between price and energy content, though in some cases pricing may not exactly
follow this method, especially pricing for price lower CV coal. For instance, historically, lower CV Indonesian
coals have traded at a deeper discount to higher CV material. For more information see PESD Stanfords
forthcoming study of the Indonesian coal market by Bart Lucarelli at
http://pesd.stanford.edu/publications/the_history_and_future_of_indonesias_coal_industry_impact_of_politics_and_
regulatory_framework_on_industry_structure_and_performance/
21
Chinas coal import tax for all coals with the exception of coking coal was 6% prior to April 1, 2005, 3% till Nov.
1, 2006, and 1% until May 31, 2007. The import tax was abolished after June 1, 2007. Coking coal import taxes
were 3% until January 1, 2005, at which point they were abolished.
22
Stuart Murray of London Commodity Brokers shared his insights with us on what transaction costs Chinese
buyers of international coal face.
23
Results are not highly sensitive to this assumption because shifts in arbitrage relationships that we describe here
are much larger than $3/ton. The overall impact is to make domestic coal slightly more favorable in all
circumstances. For instance, the highest import disadvantage prior to 2009, which was for Newcastle against
Qinhuangdao, would only drop from highs of $70/ton to $67. Conversely the highest import advantage after 2008,
which was for both Indonesian and Russian material, would only increase from highs of about $43/ton to about
$46/ton. Though domestic coals at the margin that are less than $3/ton more profitable than import coal may
become unprofitable against imports if this assumption is eliminated.
5 August 2010
10
USD/metricton
200
180
Qinhuangdao
160
Newcastle
140
VostochnyRussia
120
KalimantanIndonesia
100
80
60
40
20
0
Apr10
Jan10
Oct09
Jul09
Apr09
Jan09
Oct08
Jul08
Apr08
Jan08
Oct07
Jul07
Apr07
Jan07
Oct06
Jul06
Apr06
Jan06
Oct05
Jul05
Apr05
Jan05
Figure 6 Energy equivalent coal price indices (6,700 kc/kg GAD). Source: McCloskey, Reuters,
authors analysis.
3.5
Several necessary assumptions in ChinaCoalArb may impact the performance of the model and
should be noted when comparing model results to real market outcomes.
First, the model does not explicitly separate thermal from coking coal. 24 Though some coals can
switch between these markets making the distinction blurry at times, buying behavior in these
markets will differ under certain conditions. Most worth noting is that some demand for highquality material is likely less responsive to price movements than demand for thermal coal
because high-quality coking coal supplies are much tighter in China and internationally. We
suspect this accounts for a large share of Australian imports into China pre-2009 when CIF
prices for most Australian coals compared to Qinhuangdao coals were significantly higher.
24
In 2009 total coking coal imports were 34.5 Mt out of 126 Mt total imports.
5 August 2010
11
Second, by using the Qinhuangdao spot price to represent the Chinese domestic price, the model
does not take into account discrepancies between spot prices and term prices for power
generators in the Chinese market. Although in theory the two should be tightly correlated since
the deregulation of all coal prices in 2006 (spot was deregulated in 2002 but the NDRC still
directly capped term prices for power generators until 2006), this is not always the case in
practice. As evidenced by the June 2010 NDRC price cap on term prices, 25 China can in reality
have a two-tiered coal market under certain conditions. Thus if demand for import coal in the
power sector could replace either spot domestic coal or term domestic coal, it may be useful to
consider any price discrepancies between these two domestic markets as they will affect
arbitrage relationships. Though we argue that the current model capably represents Chinese
buying behavior in the aggregate, under certain circumstances modeling of power generators
buying behavior may be improved by using the NDRC capped term price as the domestic price
instead of spot. 26
Third, China announced that it would relax foreign exchange controls in June 2010, which has
led to minor RMB appreciation. Possible RMB appreciation going forward could increase
Chinas buying power for foreign coal and thus make imports more attractive. 27
Fourth, while the indicative energy-equivalent coal ton that we have created reflects differences
in energy content, the relationship of coal price to energy content may not always be linear (as
we have already discussed). Thus lower CV coals that are priced at a deeper discount coals may
present a greater arbitrage opportunity than our model indicates.
Fifth, while energy content is arguably the most significant variable impacting coal pricing, the
model does not reflect several key variables that can also impact price. Ash, moisture, volatiles,
sulfur, and other coal properties will also influence price. It is therefore important to note our
energy-equivalent coal ton should be broadly indicative of buying behavior but cannot capture
all variables that impact coal purchasing and pricing.
25
5 August 2010
12
Sixth, while we argue that Guangzhou is the best proxy to represent Chinas national import
behavior, other importing locations will witness slightly different arbitrage relationships that
might impact purchasing decisions. For instance, Shanghai is slightly farther from Indonesia and
Australia than Guangzhou while slightly closer to Qinhuangdao and Russia. This will increase
freight costs from the former two and favor coal shipped from the latter two.
3.6
Several macro-level factors that could impact coal imports should be noted.
First, there are technical limitations to imports that prevent a full switch to imports even if it was
clearly the most profitable option. Power plant boilers are designed to burn specific
specifications of coal, which is almost always domestic coal for energy security considerations.
However, power generators can blend imports and domestic coal supplies. For each boiler there
will be a technical limitation on the amount of blending that can occur. 28 Other industrial
applications, like cement or steel making, are likely also designed to burn domestic
specifications of coal and may not be able to fully switch to imports.
Second, import port capacity is also a theoretical limit on possible Chinese coal imports (one that
has not yet been tested).
3.7
Model Results
Figure 7 displays the results of the arbitrage model. We compare arbitrage relationships to total
monthly imports in order to demonstrate the relationship. The left axis indicates the price
advantage on a $/ton basis of import coals landed in Guangzhou compared to domestic coals
landed in Guangzhou. Negative values indicate a profit advantage for domestic coals and
positive values indicate a profit advantage for imports. The right axis indicates million tons of
28
From interviews with coal and power experts at Yudean Group in Guangdong. This is a complex engineering
issue which we will not address in detail here as there are multiple variables that impact blending ratios at coal
plants, though we do want to note that some blending limitations may exist.
5 August 2010
13
national coal imports by month. Figure 8 displays the only arbitrage relationships in greater
50
18
ImportPriceAdvantage USD/metricton
16
30
14
10
MillionTonsImports
12
10
Russia
Indonesia
10
8
Australia
QHDNEWC
30
QHDINDO
QHDRUS
4
50
2
70
0
May10
Apr10
Mar10
Feb10
Jan10
Dec09
Nov09
Oct09
Sep09
Aug09
Jul09
Jun09
May09
Apr09
Mar09
Feb09
Jan09
Dec08
Nov08
Oct08
Sep08
Aug08
Jul08
Jun08
May08
Apr08
Mar08
Feb08
Jan08
Dec07
Nov07
Oct07
Sep07
Aug07
Jul07
Jun07
Figure 7 Arbitrage relationships drive Chinese coal import patterns. Note: All arbitrage values are the
CIF Guangzhou from Qinhuangdao minus CIF value of imported coals. Source: McCloskey, Reuters,
AXS Marine, Shanghai Shipping Index, authors analysis.
5 August 2010
14
45.00
ImportPriceAdvantage USD/metricton
35.00
25.00
15.00
5.00
5.00
15.00
25.00
QHDNEWC
35.00
QHDINDO
45.00
QHDRUS
55.00
65.00
75.00
Apr10
Feb10
Dec09
Oct09
Aug09
Jun09
Apr09
Feb09
Dec08
Oct08
Aug08
Jun08
Apr08
Feb08
Dec07
Oct07
Aug07
Jun07
Figure 8 Weekly arbitrage values from 2007 show the dramatic shift in the relationship of domestic to
international coals. Note: All arbitrage values are the CIF Guangzhou from Qinhuangdao (QHD) minus
CIF value of imported coals. Source: McCloskey, Reuters, AXS Marine, Shanghai Shipping Index,
authors analysis.
The results explain the dramatic shift from China importing 40.4 Mt in 2008 to importing 126 Mt
in 2009. Prior to the fourth quarter of 2008, international coal prices were disadvantaged
compared to domestic coal prices. In summer 2008 Australian and Russian imports were out of
the money against Qinhuangdao by as much as $65/t and $30/t respectively. While minor
amounts of Australian material was still imported likely specific qualities of coking coal not
readily available in the Chinese domestic market Chinas imports were negligible because
importing wasnt profitable. The partial exception to this description is Indonesian coal, which
came in and out of the money against Qinhuangdao in Southern China. This can be attributed to
two primary factors, both of which give Indonesia competitive advantage exporting into China.
First, Indonesias geographic proximity to Chinese markets means that it pays a smaller freight
penalty than Australia and Russia (assuming Russian material is delivered into South China,
5 August 2010
15
which in reality it may not always be). Second, Indonesian FOB prices were historically slightly
lower than Australian and Russian on an energy adjusted basis. Thus as freight and domestic
and international coal prices fluctuated, import windows opened for Indonesian coal in Southern
China.
At the end of 2008 this historical relationship of domestic to international coal changed
dramatically. In the wake of global recession, the historical relationship of domestic to
international coals in Southern China inverted and a massive arbitrage opportunity arose. By late
2009, Indonesian coal was as much as $40/ton more profitable than domestic coal, Australian
coal was as much as $29/ton more profitable, and even Russian coal which suffers from a huge
rail transport penalty to move coal from central Russia to eastern ports was pricing into
Southern China against domestic coal at $40/ton better than Qinhuangdao. Imports skyrocketed,
cresting in winter 2009-2010. International prices have since recovered and the arbitrage
window began to close by summer 2010.
This model therefore provides credible explanation of Chinas coal import behavior and explains
Chinas record imports in 2009. We conclude Chinas coal buying behavior follows the logic of
a cost minimizer and Chinas coal imports will fluctuate according to the arbitrage differentials
between domestic and international coal prices.
3.8
Analyzing causes of the dramatic inversion of the historical relationship between Chinas
domestic coal market and the international coal market provides a more detailed understanding
of the market conditions that caused this shift, and thus enables coal market observers to examine
how likely these conditions are to carry forward. We argue that in the wake of the global
financial crisis of 2008 seven principal drivers caused the inversion of historical price
relationships of international to Chinese coal prices.
First, the macroeconomic impact of the global financial crisis was comparatively smaller on
China than many other coal consuming nations. Chinas GDP growth rate was 9.6% in 2008 and
5 August 2010
16
9.1% in 2009, declining by only 0.5%. 29 By contrast, real global GDP growth was estimated to
be a negative 2.1% in 2009. 30 This meant that Chinese macroeconomic activity sustained a
comparatively high level of energy demand relative to other coal importing economies. And in
China energy means coal.
Second, after the global financial crisis Chinese domestic freight prices remained higher than
international freight prices, giving imports an advantage over domestic coal. Figure 9 shows
Chinese domestic freight compared to international freight rates as a percentage of their prefinancial crisis levels. 31 This meant that the freight component of the delivered price of imports
decreased relative to its domestic counterpart.
Third, international FOB prices declined more than Chinese domestic FOB prices when
measured as a percentage of their pre-financial crisis levels. Figure 10 illustrates this trend. This
shift, combined with the relative freight advantage, in put imports at a significant advantage to
domestic coal.
29
Both are numbers after adjustment announced by China National Bureau of Statistics.
http://www.stats.gov.cn/tjdt/zygg/sjxdtzgg/t20100702_402654527.htm
30
World Bank global outlook:
http://web.worldbank.org/external/default/main?theSitePK=659149&pagePK=2470434&contentMDK=20370107&
menuPK=659160&piPK=2470429
31
We chose September 2008 as the pre-financial crisis point, as equities markets began their precipitous drop in that
month and Lehman Bros filed for bankruptcy. For a history of the performance of the S&P 500 see Bloomberg:
http://www.bloomberg.com/apps/quote?ticker=SPX:IND
5 August 2010
17
NewcastletoShanghai
140%
QinhangdaotoShanghai
120%
KalimantantoShanghai
100%
Vostochny(RUS)toShanghai
80%
60%
40%
20%
Jun10
May10
Apr10
Mar10
Feb10
Jan10
Dec09
Nov09
Oct09
Sep09
Aug09
Jul09
Jun09
May09
Apr09
Mar09
Feb09
Jan09
Dec08
Nov08
Oct08
Sep08
0%
Figure 9 Cost of freight indexed to pre-financial crisis levels in September 2008. Source: Shanghai
Shipping Index, AXS Marine.
120%
Qinhuangdao
110%
Newcastle
100%
Vostochny
90%
Kalimantan
80%
70%
60%
50%
40%
30%
20%
May10
Apr10
Mar10
Feb10
Jan10
Dec09
Nov09
Oct09
Sep09
Aug09
Jul09
Jun09
May09
Apr09
Mar09
Feb09
Jan09
Dec08
Nov08
Oct08
Sep08
Figure 10 FOB coal prices indexed to pre-financial crisis levels in September 2008. Source:
McCloskey, Reuters.
5 August 2010
18
Fourth, the relative strength of Chinas domestic coal prices was largely due to a series of
regulatory events in key production regions that curtailed supply and supported prices.
Consistent with national coal mining policies, Shanxi province embarked on a major campaign
of mine consolidation. The government began a program of closing small mines or
consolidating them into larger mines and implementing more rigorous safety standards. The
targets of that program are shown in Table 3. The result was a shut-in of traditional supply that
supported prices. There is evidence in summer 2010 that the program was not completely
successful and that government officials may re-implement similar measures. 32
Table 3 Mining consolidation targets in Shanxi.
2008 Target
2009 Target
2010 Target
Mine number
2600
1053
1000
300,000 ton/yr
900,000 ton/yr
900,000 ton/yr
Number of firms
2200
130
4 at 100 million scale
3 at 50 million scale
100
Fifth, the traditional negotiations for term coal contracts between coal producers and power
generators, which in 2008 allocated 1.1 Bt of coal (40% of Chinas total coal consumption that
year), broke down in 2009 when an agreement on price could not be reached. This led some coal
buyers that would otherwise have purchased domestic coal to look overseas.
Sixth, Chinese national policy on resource use facilitated increased imports. The so called Two
Markets, Two Resources 33 policy encourages coal buyers to import coal when the economics
justify it.
Finally, temporary factors in China like weather interruptions of transport and weak hydro
generation due to droughts contributed to higher domestic coal prices.
32
5 August 2010
19
There are a number of long term factors that will determine price relationship of Chinese coal to
international coal. Several forthcoming studies from Stanford address a set of key issues and
reforms that will impact Chinas domestic market such as the coal power conflict and resulting
policy and market reforms, the possible vertical integration of Chinas coal and power industries,
the consolidation of coal producing into larger mines, the adoption of more efficient power
generation technologies, and the construction of major coal-power bases that will produce over
100 Mt/year each. 34
4
Once a largely isolated coal market, China now plays a central role in determining global trade
flows and prices. Understanding Chinese import behavior under current and future market
conditions is therefore imperative for any analysis of the global coal trade. We have put forward
a theory of Chinese import behavior based on arbitrage relationships between China and the
global market and proved that theory in the ChinaCoalArb model.
Our findings indicate that China is a cost minimizer in the international market that will import
heavily when the price is right as it was in 2009 due to a confluence of circumstances we have
described here and largely rely on domestic coal when imports are unattractive.
The nature of Chinese demand for international coal is therefore fundamentally different from
India, the other source of dramatic demand growth in international coal markets. India is
structurally short coal because demand growth (mainly for power) has outstripped domestic coal
supplies. China, on the other hand, is now the worlds largest coal arbitrage trader. This means
that the relationship between Chinas domestic coal price and the international coal price will be
one of the key factors in determining global trade flows in the coming decade as China could just
as easily buy 15-20% of internationally traded coal as it could buy very little.
Chinas role as worlds largest coal arbitrageur has a hugely significant implication for the global
coal market: it links the international price of coal to Chinas domestic price. Chinas buying
and selling activity on the margins of its massive domestic coal market bring domestic and
global prices closer to parity (though at present not to complete parity). In other words, what
34
Forthcoming studies from PESD Stanford by Huaichuan Rui, Kevin Tu, and Yu Yuefeng address these issues.
See http://pesd.stanford.edu.
5 August 2010
20
happens in the mines of Shanxi will impact the price of power in Munich. The unique politics
and economics of the Chinese coal market are now therefore by necessity the politics and
economics of the global market, and whether or not China imports coal in a given year, the
China factor will increasingly define how the world sells, buys, and uses coal.
Acknowledgments
The authors would like to thank a number of people who made this analysis possible. Our
colleagues at PESD, including Frank Wolak and Mark Thurber, as usual provided invaluable
insights. We also greatly benefitted from discussion with many people in the coal and financial
industries. The Macquarie Securities team, including Adam Worthington, Yeeman Chin, Jon
Windham, Albert Saputro, shared their own insightful analyses of these issues for which we are
quite grateful. Bart Lucarelli also helped sharpen many of the ideas here. Stu Murray of London
Commodity Brokers, who is unmatched in his insight into Chinese buying behavior, shared his
experiences with us. Thanks also to many Chinese scholars and analysts, includingYang
Xianfeng, Li Hongjun, Hu Pingfan, Cai Guotian, Wu Lixin, Zhang Guoqiang, Wu Wenhua, Li
Hong, Gao Shixian. Finally, thanks to Kathy Lung and Sunny Wang for their support.
5 August 2010
21