CHAPTER X

INTERNATIONAL EQUITY MARKETS: FACTORS, INTERRELATIONS AND

INTEGRATION
International investing is not a new trend of the 1990s. In fact, the London Stock Exchange traces its
origin to international investing. In 1553, the first joint-stock company, where the public subscribed to
ownership in equal shares, was founded in London following advice from the famous explorer
Sebastian Cabot. Merchants at that time wanted to trade with the then mysterious Far East and were
attracted by Cabot's suggestion of searching for a North-East passage. On May 10, 1553 three ships
departed from Deptford, near London and headed north. The expedition never reached the Far East, but
one of the ships ended up in Russia. The captain of the ship was able to secure a treaty with the Czar
Ivan the Terrible, who granted freedom of trade to English ships. The Muscovy Company was founded
out of this event. The Muscovy Company was the first company where ownership and day-to-day
management were separated with shares freely traded.
In the economic development of the Victorian Age international investment played what was
recognized to be an essential part. International trade grew rapidly. Great Britain took the leading part
in this process, and the amazing expansion of British economic power in the nineteenth century is to be
attributed in on small part to international investment. The process of international investment did not
work with uninterrupted smoothness. It was punctuated by occasional financial crises (notably, for
example, the Baring crisis in the 1890s), by the 1930s depression, by currency difficulties, by wars, and
by periods of default on the part of the borrowing nations. More recently, international investing took
on a faster pace. International investment has increased four times, during the past 15 years.
International investing has probably been around for centuries, and probably for thousands of years.
This fact might be a bit surprising. International investing does not seem easy. International investors
need to analyze assets from a huge number of asset classes (stocks, bonds, derivatives, etc.), and from
different national markets and currencies. The number of individual securities to invest presents a big
challenge for international investors. The challenge is not a valuation problem since the same
techniques and models used to value a domestic asset are usually used to value a foreign asset. The
challenge is a practical once: international investors have to reduce each investment to a tractable
number of parameters, otherwise the resources spent on international investing would be too high.
Therefore, international investors need to (1) identify the major factors influencing international
security price behavior, and (2) determine the sensitivity of each security to these factors.
This chapter starts by presenting the case for international portfolio diversification. Then, this chapter
studies the technical issues faced by a firm or an investor before investing in international equity
markets. This chapter concentrates on international equity markets, but the issues are similar to those
used for other international financial markets.

I. Why Do Investors Care about International Equity Markets?

The case for international investing rests on a very simple argument: portfolio diversification.
Diversification is a risk management tool. As long as one class of domestic assets is less than perfectly
correlated with another class of domestic assets, a balanced portfolio that includes both classes of
assets is likely to exhibit more stable performance over time. This argument leads to the conclusion

X.1

that domestic investors should be well-diversified. This argument easily extends to foreign markets.
International diversification, by increasing the number of markets and assets to invest in, provides an
improved risk-return tradeoff. The case for international diversification remains as relevant for
investors today as it was one hundred years ago.
Diversification is more attractive in international markets than in domestic markets. The benefits of
diversification are driven by correlations. In general, the correlations across national markets are lower
than the correlations across securities in most domestic markets.
Example X.1: In Table X.1 we calculate the cross-country correlation coefficients for monthly stock returns for
the period January 1970 to February 2015 (542 observations). The returns are calculated from MSCI country
indexes.
TABLE X.1
MSCI Indexes: Correlation Matrix (1970-2015)
A. European Markets
MARKET
Bel
Belgium
1.00
Denmark
France
Germany
Italy
Netherlands
Spain
Sweden
Switzerland
U.K.
World

Den
0.59
1.00

B. Pacific Markets
MARKET
Australia
Australia
1.00
Hong Kong
Japan
Korea*
Singapore
Taiwan*
C. North American Markets
MARKET
Canada
Canada
1.00
U.S.
Mexico *

France
0.72
0.53
1.00

HK
0.32
1.00

U.S.
0.74
1.00

Gerrn
0.70
0.59
0.73
1.00

Italy
0.54
0.48
0.59
0.56
1.00

Japan
0.37
0.34
1.00

Mexico
0.54
0.58
1.00

Neth
0.75
0.62
0.73
0.78
0.55
1.00

Korea
0.50
0.40
0.48
1.00

World
0.77
0.88
0.56

Spain
0.56
0.51
0.59
0.58
0.57
0.59
1.00

Swed
0.55
0.54
0.57
0.64
0.50
0.63
0.57
1.00

Singap
0.51
0.57
0.39
0.46
1.00

EAFE
0.62
0.62
0.49

Switz
0.68
0.55
0.68
0.71
0.50
0.75
0.50
0.57
1.00

Taiwan
0.33
0.41
0.24
0.33
0.45
1.00

EM-LA
0.60
0.57
0.72

U.K.
0.59
0.49
0.63
0.54
0.57
0.69
0.47
0.52
0.62
1.00

U.S.
0.56
0.39
0.36
0.45
0.53
0.35

World
0.69
0.61
0.73
071
0.57
0.81
0.62
0.69
0.72
0.73
1.00

World
0.65
0.48
0.67
0.53
0.60
0.38

EM-ASIA
0.65
0.61
0.52

Notes:
*: The sample for South Korea, Taiwan, Mexico, the EM-Latin America and the EM-Asia indexes start in January 1988.

You should note that, with some exceptions, returns correlations are moderate, with an average correlation of all
the markets in Table X.1 of 0.48. The exceptions tend to be geographically closed countries, like Belgium and
Netherlands, or Germany and France. Economic integration and common economic policies play a big role. For
example, the average intra-European developed market correlation is 0.57, while the average intra-Asian develop
market correlation is 0.46.

X.2

Roll (1992) argues that the moderate correlations among international stock markets are partly
attributable to the technical procedures of index construction. Some markets indexes have a small
number of stocks (less than thirty) while others have a large number. Some national markets are
industrially concentrated while others are very diversified. These diversification elements explain part
of the observed intermarket difference in price index, not individual stock behavior.
Most stock market indices reflect the industrial structure of a country. We can think of the index from a
country as analogous to a managed portfolio with particular industry sector "bets." Therefore, countries
with different (and uncorrelated) industrial composition might display very low index intercorrelations.
Other economists disagree with Roll (1992). For example, Heston and Rouwenhorst (1994) present
evidence showing that the low correlation between country indices is almost completely due to country
specific sources of variation, not industrial structure.
We should also note that correlations are not constant. We tend to see that during periods of economic
or financial crisis, correlations around the world increase. Figure X.1 presents the correlation between
Japanese and U.S. stock monthly returns, using a 1.5-year rolling window to calculate the correlation
coefficient. For example, during the 2008 financial crisis the correlation between U.S. stocks and
Japanese stocks increases to over .80. The average correlation between U.S. and Japanese stocks is
close to 0.35.
Higher correlations during periods of crisis also occur at the domestic level. Between October 2008
and February 2009, at the height of the 2008 financial crisis, the average correlation of stocks in the
S&P 500 was around .80%. When stocks rallied in 2010, the average correlation fell to 40%, then it
spiked back over 80% during the European debt crisis.
The empirical fact that correlations tend to increase during periods of crisis is not a good one for the
advocates of international diversification: when you really want to be diversified i.e., bad times-,
diversification does not work as expected!
Figure X.1
Correlation between MSCI Japan and MSCI U.S. index returns (1970-2015)

1.A

Lower risk

X.3

Solnik (1974) was the first to quantify the risk reduction benefits of international equity diversification.
Solnik shows that U.S. domestic portfolio volatility decreases as the number of domestic stocks
increases, but asymptotically converges to a lower limit. This domestic lower limit is 27% of the
average stock volatility in the U.S. For portfolios of the same size, using both U.S. and foreign stocks,
the overall portfolio risk is substantially lower. The asymptotic lower (risk) limit is 11.7%.
Solnik's results have been verified many times, using different time periods and subsets of countries,
with similar results. Solnik's findings had a very important effect on investors and money managers in
the 1970s. In addition, in 1974 ERISA gave U.S. pension funds the freedom to invest overseas. These
factors helped U.S. investors to start to think about internationally diversifying their portfolios.
1.B

Expected Returns are Unaffected

Lowering risk is only a part of the answer to the above posed question. For example, an investor can
invest all his money in cash. This all-cash strategy has no nominal volatility, and therefore a lower
volatility than a portfolio of stocks. The expected returns, however, are also substantially lower, and
then an all-cash strategy provides no risk-adjusted benefits.
International diversification, on the other hand, seems to provide a "free lunch": it allows investors to
lower risk at no opportunity cost. The rationale for this result is simple. Investors who extend their
horizons to international markets can often find securities promising more rapid growth and more
attractive valuations. Many economists make this observation a second argument for international
diversification.
Example X.2: During the period 1978-1993, adding foreign stocks to a U.S. stock portfolio increases returns by
almost 1% and reduces volatility by almost 2%. The correlation between foreign stocks and U.S. stocks during
this period was 0.42.

The problem with the above studies and example is the use of "ex-post" returns. While variances and
covariances are estimated with reasonable precision, expected returns are not (see the Appendix to the
Review Chapter). In general, expected returns require long periods, on the order of decades, to obtain
the required statistical accuracy.
But, similar numbers are obtained using longer time periods. For example, from 1970-2011, including
25% EAFE stocks in a U.S. portfolio increased returns by an average 2.1%, while decrease volatility
by an average 2.2%.
But, there is another reason to rely on long periods of time to make inferences about the benefits of
international diversification. During short periods of time, the higher correlations during high volatility
periods (usually, bear markets) can make international diversification inefficient. From Figure X.1, it
should be clear, that during the 2008 financial crisis, U.S. investors diversifying in Japanese equity
markets did not enjoy the benefits of diversification. But, over long horizons, international
diversification works. During the 1970-2001, including 25% Japanese stocks in a U.S. portfolio,
increased returns by an average 8%, while decrease volatility by an average 3.2%. Using a
comprehensive data set from 1950, Asness et al. (2010) show that international diversification pays
off in the long run.

X.4

1.C

Staying at Home: The Home Bias

As mentioned in the Review Chapter, the lower the correlation between the assets, the greater are the
benefits due to diversification. As we have mentioned above, international equity markets offer a
domestic manager a wonderful opportunity to improve the risk-return profile of a domestic portfolio.
The empirical evidence, however, displays a puzzle: domestic portfolios largely ignore international
markets. This allocation reflects a so-called home bias.
French and Poterba (1991) and Tesar and Werner (1995) report that portfolio compositions based on
actual data indicate a strong home bias. French and Poterba's (1991) estimates suggest that in 1989,
Americans' foreign equity investment was less than 7% of the capitalized value of the U.S. stock
market. Tesar and Werner (1995) use data on international financial transactions across five OECD
countries including the U.S., the U.K., Canada, Germany and Japan for the period 1979-1990. They
find evidence of a home bias in portfolio compositions for investors in all the five countries they
examine. While investors have increased their holdings of foreign stocks in recent years, the fraction
of the portfolio invested abroad remains far less than the share implied by standard models of
optimal portfolio choice. In 2002, a study done by UBS, reports the proportion of foreign bonds and
foreign equities in the total equity and bond portfolio of local residents for several OECD countries.
The most internationally diversified investors are in Netherlands (62%), followed by Japan (27%)
and the U.K. (25%). At the bottom of the list was the U.S., with an 11% international share. More
recent studies found that the home bias has been decreasing, but it still significant. For example, in
2010 U.S. investors allocated only 28% of their portfolio abroad, while U.K. investors allocated
50% of their portfolio abroad. More surprising is that the home bias also shows up at the
institutional investor level, with North American institutional investors investing only 25% abroad,
while European institutional investors invest 39% abroad, mainly in North America.
But, the home bias is not limited to international portfolios. Investors tend to invest locally. Coval
and Moskowitz (1999) showed that the preference for investing close to home also applies to
portfolios of domestic stocks. Specifically, they showed that U.S. investment managers exhibit a
strong preference for locally headquartered firms, particularly small, highly leveraged firms that
produce non-tradable goods.
1.D

Problems with International Diversification

Many problems associated with international investing may explain the low correlations and the home
bias. The main problems are:
(1) Currency risk.
(2) Information costs.
(3) Controls to the free flow of capital.
(4) Country or political risk.
(5) Cognitive bias
We have already studied the first problem, currency risk. It is really not a "problem," since currency

X.5

risk can be managed with the different techniques studied in Chapters VI to VIII. Furthermore,
currency risk can be diversified away in a well-diversified international portfolio. The second problem,
however, is more difficult to control. Information costs includes not only the actual monetary costs of
acquiring information, but the nonmonetary costs associated with understanding different cultures,
accounting standards, legal environments, etc. We have discussed many of these points in Chapter X.
In the next two sections, we will two of the problems: capital controls and country risk.
The last problem, the cognitive bias, belongs to the field of behavioral finance. A cognitive bias is a
pattern of deviation in judgment that occurs in particular situations, leading to perceptual distortion,
inaccurate judgment, illogical interpretation, or what is broadly called irrationality. In the home bias
case, domestic investors believe that they have better information regarding the value of domestic
companies than they have for foreign companies. Investors tend to over-invest in markets (home
markets) where they believe they have a comparative advantage. Thus, a home bias is created.
Field surveys conducted in 2003 and 2005 on Japanese institutional investors showed that they held
relatively optimistic views for their domestic stock market than for the foreign stock market. Their
one-year expected returns for the Nikkei Stock Average were on average much higher than those for
the Dow Jones Industrial Average.
1.E

International Diversification through Multinational Corporations

Domestic or international multinational companies operate in many countries. In principle, MNCs are
exposed to the domestic factors of the countries they operate. For example, in 1998, a company in the
Dow Jones Industrial Average now derives on average, about 40 percent of its revenue from outside the
US, up from 35 percent in 1988. Thus, one might think that MNC stocks are good alternative to enjoy
the benefits of international portfolio diversification. The evidence from U.S. MNCs suggests, however,
that multinational firms do not provide all the benefits available from direct investment in foreign
securities. Moreover, the evidence suggests that all MNCs do not even provide additional
diversification benefits to a portfolio of purely domestic firms (i.e., a portfolio of MNCs have a higher
SD than the S&P 500).
Jacquillat and Solnik (1978) examined firms from nine countries and found that MNC stock prices
behave very much like those of purely domestic firms. Their approach was to formulate a multi-factor
market model, where each factor represents a national market index. In Table X.2, we present their
estimates of average betas for each country. Their results show that MNC stock prices are more
strongly affected by domestic factors than foreign factors, in most cases. This is especially true for U.S.
and British firms, where the addition of foreign factors to the domestic market does not significantly
improve its explanatory power (R2). This is less true for French, Swiss, Belgian and Dutch MNCs.

X.6

TABLE X.2
Multifactor Market Model for MNC
Nationality
of MNC

National Index
U.S. NL

Single Index

Single
Index

BEL GER ITA SWE FRA

SWI UK

R2

beta

R2

U.S. MNC

.94

.12

-.05

-.01

-.04

.04

.02

-.01

-.07

.31

1.02

.29

Dutch MNC

.31

.76

.09

.16

-.02

-.28

.25

-.21

-.06

.63

0.98

.50

Belgian MNC

-.27

.07

1.04 .06

.03

.19

.06

.08

.07

.58

1.03

.45

German MNC

.24

.03

-.21

1.18 -.02

-.01

.10

-.15

-.11

.74

1.18

.65

Italian MNC

-.10

.06

.10

.01

.83

.11

-.19

-.16

.20

.51

.91

.47

Swedish MNC

.06

-.15

-.02

.08

-.10

.96

.01

.15

.02

.50

.92

.42

French MNC

-.10

.14

.33

.18

.02

-.16

.95

-.22

.03

.62

1.08

.45

Swiss MNC

-.12

-.23

-.04

-.09

-.02

.16

-.11

1.74 .16

.75

1.39

.52

British MNC

-.10

-.11

.30

.09

-.04

-.13

-.09

.07

.49

1.06

.44

.84

Source: Jacquillat and Solnik (1978), Journal of Portfolio Management

Senchack and Beedles (1980) contrasted the risk, returns and betas of portfolios of multinationals
with portfolios of domestic and international stocks and found that multinationals did not deliver
diversification benefits.
In a more recent study, Rowland and Tesar (2001) find evidence that, over the 1984-92 period,
multinational corporations may have provided diversification benefits for investors in Canada,
Germany and the United States. They also find, however, that the addition of foreign market indices
to a domestic portfolio - inclusive of multinationals - provides substantial diversification benefits in
all countries.
The impact of national control and management policy, as well as government constraints on a firm's
performance, may explain why multinationals are not a good substitute for international portfolio
diversification.

II. International Capital Market Integration

With the globalization of financial markets, it is widely believed that capital markets are becoming
more integrated. The definition of integration, however, is ambiguous. An accurate definition of
integration focuses on the pricing of assets.
Capital market integration can be defined as a situation where assets in different currencies or countries
display the same risk-adjusted expected returns.
Segmentation, in contrast, implies that the risk-return relationship, in each national market, is primarily

X.7

determined by domestic factors.

The correlation coefficients shown in Table X.1 give the impression of segmented international equity
markets. The conclusion would not be correct, given that correlation coefficients are calculated using
unadjusted returns. Comparing returns in financial markets requires an adjustment: asset prices should
be adjusted for risk. Stocks in Brazil are riskier than stocks in the U.S. and, therefore, they should not
display a high correlation. When the same asset is traded in two different markets, integration can be
directly measured, because efficient markets imply that prices, once expressed in the same currency,
should be identical across markets. In these cases, correlation coefficients are an appropriate measure
of integration. More generally, however, comparing risky assets requires an asset pricing model, such
as the CAPM or APT model, to adjust for risk. Under integration, the same risk-return relationship
should apply to domestic and foreign assets.
The question of whether capital markets are integrated has several important practical implications. For
instance, if a multinational company has a choice of raising capital in two countries, the cost of capital
can be substantially different if these markets are not fully integrated. If national stock markets are
segmented, international portfolios should display superior risk-adjusted performance because some of
the domestic systematic risk can be diversified away by investing internationally.
One approach to measuring integration is to look for direct evidence of barriers to capital movements.
Many countries impose restrictions on international investment. For example, during the early 70's,
Germany tried to stem capital inflows by taxing foreign residents' deposits in Germany. A reverse
phenomenon occurred in the early 80's as France attempted to cut off capital outflows from France.
Often, however, investors can circumvent these restrictions. The key question is to test if these
restrictions are binding. Therefore, indirect evidence may be necessary to determine whether markets
are integrated or segmented. The rest of the section is devoted to indirect measures of market
integration.
2.A

Evidence from Closed-end Country Funds

A closed-end fund is a special type of mutual fund. Close-end funds differ from open-end mutual funds
(like the Fidelity Magellan or the Templeton Growth Fund) in that they neither issue nor redeem shares
after the initial stock offering. To buy or sell shares, you have to go to the market. A country close-end
fund is an investment company that invests in a portfolio of assets in a foreign country and issues a
fixed number of shares domestically.
Each closed fund provides two distinct market-determined prices: the fund's share price quoted on the
domestic market, and its net asset value (NAV) determined by prices of the underlying shares traded on
the domestic market. A closed-end fund sells at a discount (premium) if the market price is below
(above) its NAV. Domestic closed-end funds, on average, sold at a substantial discount during the 70's
and early '80s.
Closed-end country funds invest in assets from a single country. Many international markets have
closed-end country funds trading in the U.S. For example, at the NYSE we find closed-end country
funds from Brazil (two funds), Germany (two), Korea (two), India (two), Mexico (two), Italy, the U.K.,
etc.

X.8

Several governments in emerging markets allow limited access to their capital markets through closedend country funds. International investors from outside these markets must use the closed-end country
funds as the only instrument to invest in these markets.
International investment restrictions can affect the ratio of a country fund's price to its NAV if they are
binding. All other things constant, binding restriction will raise the price of a fund's share relative to its
NAV by approximately the amount the marginal domestic investor is willing to pay to avoid these
restrictions.
Example X.3: On January 13, 1989, the Korea Fund's share sold at a 65% premium relative to the fund's net
asset value. During that time, South Korea was a country with tight restrictions on foreign investment. On the
same day, however, the Brazil Fund sold at a 35% discount. In 1989 the Brazil Fund was the only way through
which a U.S. investor could invest in the Brazilian stock market.

Table X.3 presents the sample statistics for Premiums for 14 Closed-End Country Funds.
TABLE X.3
Statistics for Premiums for Closed-End Country Funds (1981-1989)
Fund or Portfolio

Mean

SD

Brazil

-28.82

9.64

.92

Mexico

-13.78

33.72

.97

France

-20.18

8.41

.92

Germany

-4.32

5.90

.77

U.K.

-21.37

6.74

.72

Japan

-11.73

10.50

.96

Korea

44.35

20.86

.93

Malaysia

-7.46

19.75

.97

Taiwan

40.96

36.24

.96

Thai

25.46

12.45

.93

Country Funds

-4.54

11.88

.89

Domestic Funds

-11.22

5.58

.97

The evidence in Table X.3 seems to support the hypothesis of International Market Segmentation.
Some funds investing in restricted countries like Korea, Thailand and Taiwan sell at a substantial
premium, while less restricted countries like Germany and U.K. sell at a discount.
These numbers are consistent with the hypothesis of a positive relation between the level of a fund's
premium and the severity of international investment restrictions. The relation, however, is not
monotonic: both Brazil and Mexico impose restrictions and sell, on average, at a discount. The last two

X.9

rows of Table X.3 show that a Portfolio of Country Funds sells at a discount of 4.54%, while a
Portfolio of pure Domestic Funds sells at a discount of 11.22%.
More evidence in favor of segmentation is provided by a paper published in the Journal of Finance in
1990, by Bosner-Neal et al. estimated that the announcement of changes in investment restrictions
(more liberal rules) decreased closed-end country fund premiums by an average of 6.8% in recent
years.
Financial Restrictions Work
The evidence favors the hypothesis of International Market Segmentation. Financial restrictions to
foreign investment work. Therefore, for segmented markets, country funds represent a good portfolio
diversification candidate.

III. Reducing the Complexity of International Investing I: Country Analysis

As we mentioned in the introduction, it is very important to reduce the complexity of international
investing to a limited number of variables. The apparent segmentation of international equity markets
points out to one very important factor: the country where an investment is planned. It has been shown
that country (currency) selection is the key decision in an active portfolio strategy. An active allocation
strategy requires the study and forecast of changes in at least three macroeconomic variables:
currencies, interest rates, and stock markets.
The first two variables affect the performance of bond and stock portfolios. Exchange rates are one of
the many variables indirectly affecting local asset prices. Exchange rates, however, have a direct
impact on the performance of foreign investments. The issue of currency forecasting was addressed in
Chapter V, where we stressed that it is difficult. It is not easier to forecast the relative performance of
national stock markets. In each country economists try to monitor a large number of economic, social,
and political variables such as
(1) anticipated real growth
(2) monetary policy
(3) wage and employment rigidities
(4) social and political situations
(5) competitiveness
(6) fiscal policy (including fiscal incentives for investments).
Real economic growth is probably the major influence on a national stock market. Philippe Jorion, in a
1999 working paper, analyzes a database of about forty stock markets going back to the 1920s. Jorion
establishes that an important long-term determinant of equity returns is real GDP growth per capita.
Jorion reports that over the twentieth century, countries that have grown at a faster rate have also
enjoyed greater stock market returns. The relationship is such that a one percent increase in real GDP
per capita is associated with close to a one percent increase in equity price growth.
Many studies have found that expected returns in international stock markets vary across time. In
addition, this variation has a predictable component. Harvey (1991) found that U.S. and national
variables can help predict future stock returns in international markets. The predictive variables are the
term spread (the difference between long-term and short-term interest rates), the default-risk spread

X.10

(the difference between yields on short-term interest rates), and the interest rate differential. Although
the predictable component of international stock returns has been found to be small, several companies
believe that an asset allocation can be improved by incorporating macroeconomic variables in their
analysis.
Investment and consulting companies invest a great deal of their resources (50% according to Frank
Russell) to country analysis. Many of these companies produce country reports for their clients. First, a
country report presents an overview of the economic situation of the country analyzed. Second, a
country report will highlight growth sectors of the economy. In the Appendix Country Report (see my
website), we discuss how to write a professional country report.

X.11

Example X.4: Smith Barney Shearson publishes a monthly report for global investors. Below, we reproduce
their Economic Overview of Hong Kong (February 1994).
HONG KONG
Geography
Hong Kong is a small British territory on the southeastern coast of China, halfway between the
islands of Taiwan and Hainan. The territory includes not only Hong Kong Island but also more
than 200 smaller islands in the South China Sea, as well as the peninsula nearby, which is part of
the mainland but juts out into the sea. Hong Kong will cease to be a British colony on July 1,
1997, when it becomes a special administrative region of China. About 99% of Hong-Kongs
population is Chinese. The frontier with Mainland China lies about 32km inland from Kowloon.
When Hong Kong Island was handed over to Britain by the Chinese, after its defeat in a war over
the opium drug trade in 1842, the territory was almost uninhabited and known chiefly as a lurking
place for bands of pirates. The British wanted to turn Hong Kong into a marketplace free from
China's control where traders of all nations might meet and do business with the merchants of
southern China. At first, Hong Kong served simply as an enormous kind of warehouse, where
goods could be kept in safety while merchants bargained with each other. But as time went on,
some of the inhabitants began to repair ships and then to build them. When the Japanese
captured the territory in 1941, hundreds of thousands of Chinese returned to China. As a result,
there were only 600,000 people in Hong Kong when World War II ended. After that, however,
Chinese refugees escaping the civil war in China entered the territory, and the population
expanded significantly. Victoria and Kowloon were so overcrowded that thousands of squatters
built shacks both inside and outside the cities. Eventually, hundreds of factories and workshops
grew up and today these make such goods as textiles, rubber shoes, ropes, flashlights,
transistors, tape recorders, televisions, vacuum flasks, and goods of ever-increasing variety.
However, Hong Kong has always earned its living mainly by handling goods from elsewhere. The
territory is an important center for trade in Asia.
Joint Declaration
With the U.K.'s lease on the New Territories expiring in 1997, the Sino-British Joint Declaration
regarding the future of Hong Kong was signed in 1984. British administration and jurisdiction over
Hong Kong will continue until June 30, 1997, and on July 1, 1997, Hong Kong will become a
Special Administrative Region (SAR) of the People's Republic of China. The joint declaration
provides that for 50 years after 1997, Hong Kong's lifestyle, including its capitalist system, low
taxes and existing contracts, will remain unchanged, and that Chinas socialist system and
policies will not be practiced in the SAR --hence the phrase "One country, two systems." The
agreement is registered with the United Nations Secretariat in New York. An additional
document, the Basic Law, effectively a post-1997 constitution for Hong Kong, was put in place in
1990. The Sino-British Joint Liaison Group (JLG) meets regularly to implement the joint
declaration.
Today, at least at an economic level, Hong Kong is an integral part of the southern Chinese
economy and it is this relationship that is the driving force behind the territories current economic
performance. Manufacturing has migrated from Hong Kong to its hinterland, except for the textile
industry, in which country-specific quotas still dictate that this industry has to be resident in Hong
Kong. As such, the economy is increasingly service oriented.

X.12

Hong Kong To Maintain Steady Growth Rate

Despite indications of a moderation in growth during the second half of 1993, the Hong Kong
economy is expected to maintain a steady growth rate of 5.5% in 1994. Growth will be driven by
exports as well as domestic demand, the later supported by private consumer spending and
infrastructure development. Export growth is expected to remain at double-digit levels in the
coming year, led by surging exports to China in both directions. Domestic exports will remain
depressed, reflecting the shift of Hong Kong's manufacturing base to China. The stock market
boom and increase in incomes will help sustain the robust growth in private consumer spending.
Building and construction activity should continue to accelerate as work on Hong Kong's new
airport gathers momentum. Private sector investment in construction is expected to remain
subdued.
Export Growth To Remain in Double Digits
The slowdown in the Chinese economy in the second half of 1994 affected the territorys exports
as exports to China moderated. However, export growth is likely to be remain at double-digit levels
in 1994, as economic growth in China is likely to be sustained, following the ending of the
economic austerity program there, and latest statistics show that the U.S. economy is getting
stronger. China and the U.S. are Hong Kongs two largest export markets. Exports to China for its
own consumption, which account for half of Hong Kong's total exports to the country, are expected
to continue to register double-digit growth rates. Exports to other markets in Asia should also
remain significant. Meanwhile, the economic strengthening in the U.S. will also fuel demand for
products made in China for Hong Kong companies.
Investment Spending Underpinned by New Airport
Investment spending will be underpinned by public-sector spending on infrastructural projects,
which will offset slower growth in private sector construction and investment in plant and
machinery. Work on the new airport will support construction activity. A large number of
government-funded airport-related projects costing HKD 40 billion have already been contracted
out. Work on these projects will continue over the next few years, as they do not require
private-sector financing and are therefore not affected by the political dispute between Britain and
China over pre-democracy in Hong Kong. However, the breakdown in Sino-British talks will affect
construction of the airport and Hong Kong's economic performance beyond 1995.
Inflation To Stay High
Hong Kong's sustained economic expansion and tight labor market are likely to keep inflation high.
Inflation is expected to remain around 8.5% in 1994. Despite large decreases in manufacturing
employment. The unemployment rate is expected to remain at the low level of 2%. Labor market
conditions will remain generally tighter in the services sector than in the other areas, due to the
acute shortages in professional, managerial, clerical, sales and service workers. So far, the shift of
Hong Kong's manufacturing base to China and increased capital investments have helped to
maintain Hong Kong's relative competitiveness. Export prices are expected to remain stable for a
second consecutive year in 1994.

X.13

Ranking* of Selected Stocks

Average Daily Volume

Market Cap

Shrs/day(m)

HKD(m)
I . HSSC

2. HK Telecom
3. Hang Song Bank
4. SHK Prop
S. Hutchison
6. China Light
7. Cheung Kong

8. Swire Pacific
9. Henderson Land
10. Wharf
11. HK Land
12. HK Electric
13. New World Devl
14. Jardine Math
15. CMC Pacific

1. Masshan
2. Shanghai Petroche
3. Hopewell
4. HK Telecom
5. Sino Land
6. Great Eagle
7. Hutchison
8. Choung Kong
9. HK Land
10.Hang Lung Devi
11.CITIC Pacific
12. Guangzhou Ship
13. Now World Devl
14. HSIBC
15. HK & China rim

261,080.0
169.522.6
138,102.3
129,987.0
116.370.8
99,027.4
97.243.8
93,662.5
79,401.0
69.921.2
67.704.8
56,164.3
50,832.6
47,523.0
43,958.7

Prospective EPS Growth

Prospective PER (%)

I. Sino Land
2. Hopewell
3. Swire Pacific
4. HK Land
S. China Light
6. Wheelock Marden
7. Now World Devi
8. Cathay Pactfic
9. Henderson Land
10. Shangn-La Asia
11. HKS Hotels
12. Shun Tak
13. HSBC:
14. Dao Hong Bank Group

67.0
63.2
50.2
46.0
38.9
35.4
33.0
30.3
28.8
28.3
25.1
23.3
22.5
22.0

1. UOL

2. Cheung Kong
3. Jardine Int'l Motors
4. Sime Darby (14K)
5. Now World Devi
6. Kumagai Gurni (HK)
7. Hang Lung Devi
8. HS8C
9. Hopewell
10. SCM Post
11. Tian An
12. Jardine Math
13. Shun Tak
14. Cathay Pacific

7.0
9.5
10.2
10.9
11.1
11.3
11.5
11.5
11.9
12.2
12.4
12.5
12.7
12.8

15. Tian An

21.1

15. Dairy Farm

13.3

X.14

42.075
38.414
15.667
15.629
14.249
7.681
7.510
6.105
5.452
5.329
5.281
4.445
4.274
4.177
4.162

KEY INDICATORS
1990

1992

1993E

Market Indicators
EPS Growth (%)+
Mkt Cap (USDbn)

6.0 23.0 25.0

83.5 122.0 162.0

19.0
-

20.0
-

Mkt PER (x)

Mtd. Yield (%)
Price/Book Value (x)
Price/Cash Earnings (x)

10.1
5.3
2.6
26.6

8.7
2.5
2.5
17.7

5.8
2.8
2.3
15.5

Financial Conditions
Foreign Reserves (USDbn)
M2 Growth (%)
Gvt Bud Surplus (USDbn)
USD Cross-Rate (yr-avg)
Macro Indicators
Real GOP Growth
Inflation (cpr%)
GOP (nominal USDbn)
Capital Formation (%GDP)
Trade
Current Account (USDbn)
Export Growth (nominal)

Import Growth (nominal)

n.a
16.8
1.2
7.8

3.2
9.8
71.6
27.4

1991

11.9
4.7
2.6
21.6

11.4
3.6
2.8
18.0

15.0
(0.4)
7.7

1994F

n.a
26.2
0.5
7.8

n.a
12.2
2.6
7.7

15.0
(2.1)
7.8

4.2
12.0
$2.2
27.4

5.0
5.5
5.5
9.4
8.5
8.5
96.6 109.9 126.4
27.6 30.0 32.0

3.8
2.5
2.6
2.5
12.2 19.7 20.8 14.0 13.0
14.2 21.2 22.6 14.0 13.5
235.0 245.0 255.0 250.0 261.4

Total Trade (% of GDP)

+Based on HSI aggregate of 33 stocks

F: Forecasts as of 14/01/94 (HSI = 10.774)

X.15

3.A

Country Risk Analysis

Sovereign debt crises and defaults have been around for centuries, see Graph X.1 for a more recent
history. The first recorded default in history occurred in the fourth century B.C., when ten Hellenic
city-states defaulted on loans from the Delian league. But, the dramatic increase in international
lending in the 1970s gave rise to the discipline of country risk analysis. During the early 1980s, many
developing countries began to incur large balance of trade deficits, largely because of a global
recession. The large balance of trade deficits caused deteriorated conditions in developing countries.
Loans to Latin American countries were largely denominated in dollars. Loan repayments became
more difficult when the dollar strengthened against the borrowers' currencies and when USD interest
rates began to rise in the early 1980s. As a result of this situation, more than 25 countries requested a
reschedule of their debts. International banks and investors suddenly realized the relevance of
improving country risk analysis.
Graph X.1
Sovereign External Debt1800 - 2006 Taken from Reinhart and Rogoff (2011)

Country risk analysis is limited, international money managers recognize that their decisions might be
incorrect. To reduce the damage, many institutions place a cap on the total amount invested in a
particular country. This would limit the sensitivity of the overall portfolio to any single country's
problem. Diversification also helps to protect the overall portfolio from substantially declining in
value, since any country's economic and political problems will affect only a small portion of the
portfolio. However, some countries tend to have risk ratings that move together. For example, the
country risk of a portfolio of investments in the Middle East will tend to move together with the price
of oil. There is a more recent set of tool to reduce country risk: derivatives. There is an active market
for credit and political risk derivatives. In Chapter XIV, we will cover the most active of these
derivatives, the credit default swap.

X.16

Diversification and Country Risk

Taken From Profits in a Time of War, The Economist, Sep 20, 2014.

An excessive concentration on one country is a classic mistake. After Chinas revolution in 1949
HSBC, then a purely Asian bank, lost half its business. Irans nationalisation in 1951 of the AngloIranian Oil Companys assets devastated the firm, a precursor of BP.
There are modern echoes of these episodes. Repsol, a Spanish oil firm, fell in love with Argentina,
leaving it vulnerable when YPF, the firm it bought there, was nationalised in 2012. First Quantum,
of Canada, had made a third of its profits from a mine that the Democratic Republic of Congo
nationalised in 2009. But as they have expanded over the past two decades, multinationals have
spread themselves more. Only a dozen big, global, listed firms have over a tenth of their sales in
Russia. BP is the countrys largest foreign investor but gets only about 10% of its value from its
stake in Rosneft, an oil giant. McDonalds Moscow outlets, once a symbol of dtente, are
temporarily shut, victims of a diplomatic tit-for-tat. Even so, the burger giant makes less than 5% of
its profits in Russia.
This picture is true in other hotspots. Telefonica, a Spanish firm, and Procter & Gamble (P&G),
together have billions of dollars trapped in Venezuela, which has introduced capital controls. But it
represents less than 5% of their sales. Ben van Beurden, the boss of Royal Dutch Shell, recently said
diversification is the only way to inoculate yourself.

The basic idea behind country risk analysis is to determine from a big data set (with a lot of
economic, socioeconomic and political variables and observations) a single measure that describes
the riskiness of investing in or lending to a country. We will call this single measure country risk.
This measure is usually expressed as a letter (say, A to F). The lower the letter is, the worse the
ranking of a country.
There are two main approaches to determine country risk: a qualitative approach and a quantitative
approach.
(1) Qualitative Approach: It is based on the opinions of experts (politicians, union members,
economists, political analysts, etc.) to form a consensus opinion about the risk of a country. The
consensus opinion becomes the grade.
(2) Quantitative Approach: It is based on a model. It starts by determining some quantifiable factors
that influence country risk. Then, a formula is used to determine numerical scores for each factor.
Finally, a weighted average of the factors numerical scores is calculated. This weighted average
determines the final grade.
The quantitative approach tends to be considered objective; however, many decisions, like the number
of factors and the weights, are based on some degree of subjectivity.
There are many institutions and publications that publish country risk ratings. The rating agencies
Moody s, S&Ps and Fitch IBCA also produce country risk ratings. The Institutional Investor
publishes a summary of country risk ratings for 150 countries based on a survey of some 100

X.17

international banks. This summary is revised every six months. The risk rating scale ranges from a high
of 100 (best level) to a low of zero. Euromoney has a panel of 40 leading economists in international
financial institutions evaluating country performance in the financial markets. Based on these panels,
Euromoney publishes ratings for 180 countries, every six months.
The main advantage of country risk is its simplicity. It provides a single, easy to understand single
measure. It allows cross-country and across time comparison. But, many critics state that it is too
simple. In practice, we tend to observe that country ratings tend to converge. This phenomenon is
called herding, which is also a common feature in credit ratings. Because many of the factors are slow
to change, we tend to see a lot of persistence in the country ratings. For example, in 2011, for
Euromoney the world ranking of China was 40th. It is almost the same world ranking it had in 2001
(45th), even though China has done very well through the global financial crisis of 2008.
The main criticism of country risk analysis is the lack of predictive power. For example, no rating
agency or publication was able to predict the Asian crisis of 1997. South Korea had the same rating as
Sweden as late as October 1997. Then, after the magnitude of the crisis emerged, South Korea was
suddenly downgraded.
To be fair, we should note that country risk, like all risks, is an unobservable variable. We observe
factors that influence risk, but we do not observe the unobservable risk. The single measure country
risk will attempt to estimate the true country risk. But, keep in mind always, that country risk ratings
refer to an unobservable variable. It is difficult to estimate unobservable variables.
We will present a brief description of two approaches to analyze country risk: the Risk Rating Model
and the Prince Model. Different variations of the Risk Rating Model are widely used to evaluate the
degree of risk of a country as whole. The Prince Model evaluates the political risk.
3.A.1

Risk Rating Method

John B. Morgan, in an article published in Banker's Magazine, describes an assessment method used by
banks to measure country risk. This system is based on four major aspects of a country:
i.
ii.
iii.
iv.

Economic indicators, to evaluate the country's financial condition

Debt management, to measure the country's ability to repay debt
Political factors, to evaluate political characteristics and political stability
Structural factors, to measure socioeconomic conditions, such as human resource base.

Short-term and medium term models of these four aspects need to be developed. The segmentation into
two time horizons is used because a country's economic outlook may vary with the time horizon used.
Each of the four models assigns a score between 0 and 100. The scores (between 0 and 100) for each
factor are a function of fundamental data. For example, the economic indicators grade depends on
GDP per capita, GDP growth, inflation, productivity, interest rates, etc.
A specific formula is used to compute the scores. For example, for the economic indicator (EI) we can

X.18

use the following linear model for country j:

Score(EI)j = 0 + 1 GDP growthj + 2 Inflationj + 3 Productivityj + ....
Statistical methods, like regressions or Bayesian analysis, are used to determine the coefficients (0, 1,
2,...). A practitioners experience will also play an important role in adjusting the model and the
coefficients, if needed. For example, experience indicates that GDP growth should have a positive
impact on the economic indicator factor. If the coefficient is negative, a practitioner will likely
conclude that there is something wrong with the data or with the formulation of the model.
Once the four models are complete, the overall rating is determined by weighing the importance of the
models. The overall numerical grade is converted into a rating.
You should note that the factors and weights assigned to the four factors are subjective. Different
institutions use different factors and weights, and, therefore, reach different conclusions.
Example X.5: Euromoney produces semi-annual country risk analysis of 185 countries using a panel

of experts. They rate nine categories with a score (0 to 100).

Categories and weights:
Economic performance: GDP growth
Political Risk
Debt indicators: Debt/GDP
Debt in default or rescheduled
Credit rating: Moody s or S&Ps or Fitch IBCAs rating
Short-term credit market access
Access to bank finance: Commercial bank credit
Access to Capital markets
Discount default factor: Spread over US Treasury bills:

- 25% weight
- 25% weight
-10% weight
-10% weight
-10% weight
- 5% weight
- 5% weight
- 5% weight
- 5% weight.

As mentioned above, the overall score is converted into a rating (letter). The conversion process
converts grades into categories similar to those in Standard and Poor's rating system on securities.
Table X.4 presents a standard conversion table. Figure X.2 presents the world country risk ratings on
January 30, 2012. On that day, the world country risk weighted average was 44.50, for a B rating.

X.19

TABLE X.4
Conversion Table of a Country's Grade into a Rating
Overall grade

Rating

Rating

91-100

AAA

Excellent

81-90

AA

71-80

61-70

BBB

51-60

BB

41-50

31-40

CCC

21-30

CC

10-20

0-10

Average risk

Low quality

Excessive risk

Notes:
A rating of BBB or better is considered investment grade.
A rating of BB or less is considered junk (also called high-yield, speculative grade). In the U.S.,
the usual spread of junk debt is between 400 to 600 bps over 1-yr T-bills. Range is very wide: Spreads
can go over 2600 bps.
Each rating category has three subcategories, usually with a plus or a minus, say, AA+, A, and
AA- for the AA rating category. Low quality or below categories (CCC to D) have no subcategories.

X.20

Figure X.2
Euromoney World Country Risk, Janauary 30, 2012

Ratings are used to determine the interest rate a borrower should be charged. At the domestic level,
usually each rating is associated with a credit spread range, which is expressed in basis points (bps).
For example, for a company rated AA, the credit spread may be between 40 bps and 85 bps. Then, the
credit spread is added to a base rate, usually the yield-to-maturity of government bonds with similar
maturity to the maturity of the loan under consideration. For countries, the method of determining the
interest rates governments should pay is similar. The terminology, however, is different. When we
consider countries, the credit spread is called country risk spread or sovereign spread. The base rate is
usually the U.S. T-bill or T-bond rate.
Example X.6: Bertoni Bank evaluates the country risk of San Marcos. The following table illustrates the use of
the risk rating method.
Short-term Horizon
Factor

Weights

Grade

Weighted

Medium-term Horizon
Weights

Grade

Grade

Weighted
Grade

Economic

.30

80

24

.30

70

21

Debt manageme

.30

90

27

.20

70

14

Political

.20

60

12

.30

50

15

Structural

.20

75

15

.20

60

12

Total

78

X.21

62

San Marcos has a better ranking in the short-term than in the medium-term. The short-term ranking is A and the
medium-term horizon ranking is BBB.
Now, suppose the government of San Marcos wants to issue short-term debt. Suppose, the sovereign spread
range for a country rated A is [90 bps, 135 bps]. Based on the country risk analysis, Bertoni Bank determines that
the appropriate sovereign spread for San Marcos is 95 bps. Suppose the U.S. T-bill is 3%, then San Marcos
should be charged an interest rate of 3.95% for short-term debt.

The risks associated with investment-grade countries are considered significantly higher than those
associated with first-class government countries. The difference between rates for first-class
government bonds and investment-grade bonds is called investment-grade spread. The range of this
spread is an indicator of the market's belief in the stability of the economy. The higher these
investment-grade spreads are, the weaker the economy is considered.
Country Risk Analysis: The Times Are Changing
The 1990s wave of international capital flows has changed the focus of country risk assessment. One of
the principal lessons from the 1970s was that traditional financial analysis was too narrow to capture
the forces at work, the result has been a broadening of the factors analyzed. Narrow financial ratios
have been replaced by broad-gauged economic analysis. Joyce Chang, director of international
emerging markets fixed-income research at Merrill Lynch, says, "one of the key components of
looking at sovereign risk is to have political analysts look at market events." Merrill Lynch has used
political experts at local universities "who know nothing about markets," as Chang acknowledges, but
who can help evaluate local political conditions and asses the prospects for stability.
Source: Institutional Investor, September 1997.

3.A.2

The Prince Model

William D. Coplin and Michael K. O'Leary, from Syracuse University, developed an integrated
framework that allows comparisons across countries called the Prince Model. The Prince model has
been used by the C.I.A., the U.S. State Department, and many multinational corporations.
3.A.2.i

Political Risk Factors

The relevant factors are:

(1) Regime change: A change in key government personnel through normal electoral or authorized
political process, or through illegal means.
(2) Political turmoil: General levels of politically inspired violence, including violent strikes,
demonstrations, riots, terrorist activities, guerrilla actions, or civil war.
(3) Government policy: Decisions with respect to fiscal and monetary policies, trade restrictions, or

X.22

foreign investment regulation

(4) External events: Other countries' actions that affect the country of concern.
Government policy has the most frequent and negative effects. Coplin and O'Leary developed a rating
system that places countries in the categories A (less risk), B, C or D (more risk) for two government
policy areas:
(1) Financial Transfer Risk. This refers to the risk from financial transfer, non-convertibility from the
local currency, and the transfer of foreign currency out of the country. The transfer could be for the
payment of exports, repatriation of profits or capital, etc.
(2) Direct Investment Risk. This refers to the risk to foreign investment in wholly owned subsidiaries,
joint ventures, and other forms of direct ownership of assets in a country.
3.A.2.ii

Background Data

Sources: International Financial Statistics (IFS), published by the IMF; Foreign Economic Trends
(FET), published by the U.S. Department of Commerce; local government publications, local press,
direct contacts, U.S. Embassy officials.
Economic Indicators: GDP real growth, GDP per capita, capital investment, budget balance, changes in
real wages, unemployment rate.
International: Debt service ratio, current account, exports, imports, principal exports and imports,
currency changes.
Social Indicators: Energy consumption, population growth, infant mortality rate, urban population,
literacy, persons under 15, income distribution, work force distribution.
Political: Constitutional organization of the country, main government officials, schedule of election,
status of the press, currency exchange system, sector of government participation in the economy.
3.A.2.iii

The Prince Political Forecasting System

The Prince Model forms the basis of many forecasting models for international business. The Prince
Model is used to generate probability scores for the most likely regime, turmoil, and restrictions on
international investment and trade.
Example X.7: Prince Chart for Regime Stability in the Dominican Republic, October 1987.

X.23

Actor

Orientation

Certainty

Power

Salience

Prince Score

Armed Forces

+48

Cuba

-12

PLD

-40

PRD

-50

International Community

+24

Labor

+18

Radical Left

-25

PR -Reformist Party

+100

Catholic Church

+27

Rural Workers

+16

U.S.A.

+100

University Students

-24

Total Positive Scores:

Total of All Scores:
Probability:

333.
484.
333/484 = 70%.

Notes:
Orientation: The general attitude of the actor.
Certainty: The firmness of the actor's orientation.
Power: Degree to which the actor can exert influence in support of his/her position.
Salience: The importance attached to supporting the actor's position, relative to all other concerns facing the actor.

The "Prince Score" is the product of the firmness of the actor's orientation (certainty), the player's ability to
influence the outcome (power), and the importance of the specific risk factor to the actor (salience). Support
(opposition) is indicated by a positive (negative) sign next to the score.
The above scores are based on the "consensus" of a team of country specialists. The probability of the regime
remaining in power for the next 18 months following October 1987, was 70%.

The Prince Model was also used in other areas. For example, the Prince Model aided the C.I.A. in
estimating the probable position of 52 countries on various issues at the 1979 World Administrative
Radio Conference.
Country Risk and The Asian Crisis
In a study of the Asian Crisis, Clark L. Maxan (1998) shows that the usual measures of country risk did
not help to predict the 1997 Asian market crisis. In fact, most Asian countries saw improvements in
terms of risk (that is, they became less risky) up to and shortly before the crucial turning point around

X.24

September 1997. Moreover, other forward-looking measures of country risk completely failed to show
any sign of recognizing the coming market collapse until it was underway. Source: Emerging Markets
Quarterly, Winter 1998.

3.A.3.
Other Country Risk Indicators
Given the lack of predictive power of CR, a single indicator may not be enough to capture the overall risk of a
country. There are other indexes that may be help to signal the true riskiness of a country i.e., indicators that
can be correlated with the underlying true country risk.
There are many indicators that practitioners use to gauge the global economic, social and political
environment of an economic. The most popular indicators are the following:
- Globalization index (GI), produced by A.T. Kearny. It measures a countrys global links. It looks to
all kind of data: foreign direct investment (FDI), international travel, internet servers, number of
foreign embassies, etc.
- FDI confidence index, produced by A.T. Kearny. It is based on a survey of senior executives from the top
1,000 MNFs indicating the likelihood of investment in specific markets.
- Global competitiveness index (GCI), produced by World Economic Forum. It uses two indexes: a global
competitiveness index and a microeconomic competitiveness index to rate each country growth environment
and opportunities.
- World competitiveness index, produced by the Institute for Management Development (IMD). It uses over
300 economic and social indicators to determine the ability to grow and sustain growth in a country.
- Opacity index, produced by PWC and The Milken Institute. It covers 50 countries. It is based on 5
components: Corruption in government bureaucracy, Laws governing contracts or property rights,
economic (fiscal, monetary, and tax-related), accounting standards, and business regulations.
Together, these factors form the acronym CLEAR (Corruption, Legal, Economic, Accounting,
Regulatory). A higher level of opacity in any area increases the cost of doing business and makes
investment capital riskier.
- Index of economic freedom, produced by the Heritage Foundation and the Wall Street Journal. It takes
into account different institutional factors determining economic freedom: corruption, non-tariff
barriers to trade, the economic (regulatory and non-regulatory) burden of government, the rule of
law, restrictions on banks, black market activities, etc.
Table X.5 presents a summary of some country risk indicators for 5 countries: Brazil, China, Japan,
the U.K., and the U.S. In general, the rankings from the different indexes tend to converge, but
convergence is not always the case. The economic freedom rankings of Brazil and China are far
away from the others.

X.25

TABLE X.5
Summary of Different Country Risk Indicators
Country

Brazil
China
Japan
UK
USA

Euromoney
(2011)

41
40
25
17
15

Globalization
(2007)

GCI
(2011)

World
Competitive
(2011)

Opacity
(2009)

Economic
Freedom
(2011)

53
26
9
10
5

44
19
26
20
1

28
45
16
2
6

99
138
22
14
10

67
66
28
12
7

IV. Reducing the Complexity of International Investing II: International Factors and Linkages
Recall that international investing involves a huge number of variables and factors. Investors would
like to have this huge number of factors reduced to only a few key factors. If this were possible, it
would greatly simplify the task of structuring a well-diversified international portfolio. The first step in
this process is to determine whether the price of an individual security is primarily affected by
international or purely domestic factors.
4.A

Domestic versus International Factors

Different studies have shown that domestic factors are more important than international factors.
International industry effects appeared weak compared to national effects. A simple approach to
determining the relative importance of each factor is to separately correlate each individual stock with
i.
ii.
iii.
iv.

the world stock index

the appropriate (international) industrial sector index
the currency movement
the appropriate national market index.

The first three factors may be regarded as international, and the last one as domestic.
Example X.8: Solnik and de Freitas, in a paper published in Recent Developments in International Finance
and Banking in 1988, performed the following experiment. They used monthly observations on a sample of 279
international firms from eighteen countries over the period December 1971 to December 1984. The country,
industrial, and world indexes come from MSCI. The currency movement is that of the local currency relative to
the USD. They regressed each individual stock on each factor and obtained its R2.
Recall that a measure of correlation is the R2. The R2 is a measure of correlation that tells us how much of the
variability of the dependent variable is explained by the independent(s) variable(s).
Table X.6 reports the average R2 for all companies from a given country in the first four columns. The average
R2 for the multiple regression (multi-factor Market Model) is reported in the last column.

X.26

TABLE X.6
Average R of Regression on Factors
2

Single-Factor Model

All Factors

Market
World

Industrial

Currency

Domestic

Belgium

.07

.08

.00

.42

.43

Germany

.08

.10

.00

.41

.42

Norway

.17

.28

.00

.84

.85

Spain

.22

.03

.00

.45

.45

Sweden

.19

.06

.01

.42

.43

France

.13

.08

.01

.45

.60

Italy

.05

.03

.00

.35

.35

Netherlands

.12

.07

.01

.34

.31

U.K.

.20

.17

.01

.53

.55

U.S.

.26

.47

.01

.35

.55

Canada

.27

.24

.07

.45

.48

Australia

.24

.26

.01

.72

.72

Hong Kong

.06

.25

.17

.79

.81

Japan

.09

.16

.01

.26

.33

Singapore

.16

.15

.02

.32

.33

All

.18

.23

.01

.42

.46

Note:
1. The various correlations do not add up; the four factors are correlated with each other.
2. The world factor and industrial factors explain an average of 18% and 23% of the variability of stock returns.
3. Domestic factors are the most important influence on stock returns.
4. The currency factor is almost negligible, with the exception of Hong Kong and Canada.
The simple regression of stock returns on the domestic market index return has an average R2 of .42. When we
add the three international factors, the average R2 increases to .46. This is a rather small change in R2. However
the story differs among countries; the increase is rather large for the U.S. and France.
A detailed analysis of the results indicates that the marginal contribution of the international factor is generally
positive and significant. The marginal contribution of the currency factor is extremely weak but positive, and
appears to be country specific but not company specific. A local currency appreciation tends to be good for the
local stock market.

X.27

Other studies have used different methodologies, however the results are similar. In an article
published in the Journal of Portfolio Management in 1989, Grinold, Rudd and Stefek find that some
industry factors are more "global" than others. For example, they found that the oil industry factor is
highly significant, which is not the case of the factor for consumer goods.
4.A.1

Currency Factor and Hedging

In Example X.8 we found that the currency factor on average was negligible. The average R2 was .01,
which indicates that only 1% of the variation in local currency is explained by changes in the exchange
rates. The low explanatory power of the currency factor has currency hedging implications. If stock
returns are independent of exchange rate changes, currency risk is not a systematic factor in the APT
model and, then, is not priced in the stock market. That is, hedging currency risk cannot affect the
systematic risk of multinational firms.
Jorion (1991) analyzes the importance of the currency factor using a simple CAPM model with a
currency factor added. He also uses a similar APT model than the one used by Solnik and de Freitas.
Jorion uses more sophisticated tests and obtains a similar result as Solnik and de Freitas: the currency
factor is already incorporated into the other factors. Therefore, currency risk is diversifiable and is not
priced in aggregate in the stock market.
Jorion also analyzes the importance of the currency factor at a disaggregate level. Jorion finds that
different firms and industries have different and significant exposures to the currency factor. Exporters
tend to benefit from a depreciation of the domestic currency, while importers tend to benefit from an
appreciation of the domestic currency. In another paper published in the Journal of Business in 1990,
Jorion finds that the sensitivity of a U.S. firm's returns to the currency factor is positively related to its
percentage of foreign operations.
4.B

Linkages among Stock Markets

As it was mentioned in Section I, the positive but moderate correlation coefficients in international
stock markets are the main reason behind internationally diversifying portfolios. The low correlation in
some markets is surprisingly low, given the increasing global financial integration.
The analysis of correlation coefficients might not be the correct tool to study the issue of linkages
between international markets -or the issue of integration versus segmentation. For example, one could
conceive of a situation where no movement of capital is allowed between national stock markets, but
common shocks to growth or monetary policies induce positive correlations between the two markets.
In such a case, ex ante, or expected, returns could be very different across markets, even with highly
correlated ex post, or realized, returns. Thus, it is very useful to study the linkages among international
stock markets without focusing on the correlation coefficients.
Big and unusual stock market movements (extreme events) seem to drive markets together. During
these events all markets move together in the same direction and with similar changes. The Crash of
October 1929 is one of these events. Another event that highlighted the links between international

X.28

stock markets during big market movements was the Crash of October 1987. The Crash of October
1987 is a recent and very well studied event, which might provide an opportunity to understand
international linkages. In the next section, we will analyze the Crash of October 1987 and its
consequences.
4.B.1

Extreme Linkages: The Crash of October 1987

Roll (1989) points out that the October 1987 Crash was the only month during the 1980's where all the
stock markets around the world moved in the same (negative) direction. The international transmission
of the crash started in non-Japanese Asian countries and continued through European markets, the U.S.
and finally Japan.
In Table X.7, we reproduce the daily returns during the Pre-Crash period, the Crash period and the
Post-Crash period by country.
TABLE X.7
Daily Returns (percent/day) by Country
Country

1/2/87-10/12/87

10/12/87-10/30/8

11/2/87-3/31/89

Australia

.2239 (0.850)

-3.5160 (8.315)

.0475 (1.216)

Hong Kong

.2218 (1.121)

-5.4174 (12.072)

.1083 (1.353)

Japan

.1543 (1.274)

-0.9777 (5.567)

.0810 (0.946)

Malaysia

.2821 (1.171)

-3.6080 (6.026)

.0128 (2.754)

N. Zealand

.0291 (1.091)

-2.0473 (5.296)

-.0755 (1.366)

Singapore

.2508 (1.075)

-3.9675 (10.182)

.1004 (1.327)

Austria

-.0202 (0.736)

-0.8255 (1.663)

.0699 (0.557)

Belgium

.0808 (0.814)

-1.6531 (4.316)

.0906 (0.965)

France

.0114 (0.920)

-1.6526 (4.568)

.1018 (1.254)

Germany

-.0296 (1.251)

-1.5913 (4.178)

.0254 (1.292)

Italy

-.0338 (1.017)

-1.3943 (3.184)

.0293 (1.149)

Netherlands

.0672 (0.993)

-1.5985 (5.296)

.0633 (1.301)

Spain

.2143 (1.276)

-2.4154 (3.286)

.0555 (0.927)

Sweden

.1272 (1.009)

-1.8998 (4.534)

.1202 (1.242)

Switzerland

.0156 (0.917)

-2.0706 (5.409)

.0025 (1.305)

U.K.

.1852 (0.865)

-2.0759 (4.947)

.0524 (0.962)

Canada

.1143 (0.689)

-1.5150 (5.413)

.0405 (0.772)

X.29

Mexico

.9831 (2.509)

-3.4050 (6.892)

.0128 (2.754)

U.S.

.1213 (0.965)

-1.4128 (7.253)

.0428 (1.094)

Notes to Table X.7: Standard deviation in parenthesis.

There is a major question of interest: what were the causes of the Crash? Roll (1989) analyzes four
candidates that were extensively discussed after the Crash by academic researchers, politicians and
journalists:
(1) Portfolio insurance and computer systems?
Just after the Crash, many journalist and politicians blamed the Crash on a variety of sources ranging
from portfolio insurance to inadequate computer systems. Empirical studies have found these claims
totally unfounded. For example, many studies have found that markets with portfolio insurance crashed
less than markets without it.
(2) Futures markets?
Another allegation points to all uses of stock index futures or other related futures contracts, not just
portfolio insurance. The argument seems to be that irrational speculators cause instability. However,
stock markets with related futures markets crashed in the same way as countries without futures
exchanges.
(3) Specific event?
The search for a triggering event has been very unsuccessful. Several reasons have been advanced in
the U.S.: announcement on October 14 of a worse than expected trade balance, poor performance of
Asian markets in the week before the Crash, etc.
The most likely event was the introduction in the U.S. Congress of a tax bill that would have severely
penalized corporate takeover, leverage buyouts, and other similar activities.
The evidence for this event is the most persuasive for all the events advanced; however, it is difficult to
believe that it had such an extraordinary effect in other markets.
(4) Speculative bubble?
Eugene Fama, from the University of Chicago, says that the most questionable aspect of 1987 was not
the Crash itself, but the incredible market advance during the previous five years. There is strong
evidence that fundamental factors from January through September of 1987 could not explain such
large increases as were observed in the U.S. The same seems to be true for other markets.
Example X.9: Go back to Table X.7. The rate of return from January 1, 1987 to October 9, 1987 was 37.62% for
the U.S., 44.45% for the U.K., 36.12% for Japan, and a spectacular 687% for Mexico.

This apparent behavior has been attributed to a speculative bubble. Under this view, the most plausible
theory for the Crash is that a speculative bubble burst in October 1987.
A Bubble at Work: The South Seas Bubble

X.30

In March 1711 the Chancellor of the Exchequer announced to Parliament plans to convert the National
Debt, then GBP 9 million, into shares in a joint-stock company called "The Governor and Company of
Merchants of Great Britain trading to the South Seas and other parts of America, and for the
encouraging of the Fishery." This company was better known as the South Seas Company. In exchange
for GBP 9 million, the South Seas Company was going to receive a 6% annual payment from the
government, plus the monopoly rights to trade with South Seas territories. The issue was quickly fully
subscribed. Within four years the capital of the company was increased to GBP 10 million. By 1717,
nearly half of the wealth of the country was invested in the company. In 1720, Sir John Blunt, a wellknown banker and also director of the South Seas Company, proposed to the Government a merger
with the remaining National Debt -at that time GBP 31 million. In April 1720, the Parliament approved
this idea and the South Seas Act was passed. Within days of the Act receiving royal assent, the
company announced its first money subscription at a price of GBP 300 for GBP 100 par. After
several promises from the company of dividends of 30% annually and even 50% annually, the issue
was oversubscribed. Approximately GBP 1.5 million was subscribed in less than a day -even the King
and the Prince of Wales bought stock at 400%. The interest in this company was such that other
"bubble" companies were set up. It was calculated that the value of all these bubble companies, valued
at market prices, was close to GBP 500 million -five times the value of all the cash in Europe. The
bubble soon burst.
Some investors made huge profits with the bubble, like the Prime Minister, Sir Robert Walpole, and
the Prince of Wales. Others were ruined. After the collapse of the South Seas shares, the Bank of
England and the East India Company took over the South Seas Company, and its shares traded for
another 100 years. The company, however, was never successful.
As a result of the South Seas bubble, the British government put severe restrictions on joint stock
issues. These restrictions left two insurance companies, the Bank of England, the East India Company
and what remained of the South Seas Company as the main components of the English stock market
for the rest of the century.
Another result of the South Seas bubble was a long poem written called The Bubble by Jonathan
Swift, the author of Gullivers Travels, published in January of 1721. The poem included the following
line:
While some build Castles in the Air,
Directors build em in the Seas;
Subscribers plainly see um there,
For Fools will see as Wise men please.
It is difficult to test for bubbles. It is very difficult to measure the structure of the speculative bubble.
Some popular tests are based on the structure of the autocorrelations. During a speculative bubble, the
degree of serial correlation could be highly non-stationary, swinging up and down and yet still being
positive during most of the bubble's expansion. (Traditional methods, like the Q test, usually assume
stationarity and they may have weak detecting power as a consequence.)
Several studies have used non-standard tests (two popular test are the so-called variance ratio test and
the BDS -Brock, Dechert, and Sheinckman- test) and have dismissed it as a plausible explanation for

X.31

the October 1987 Crash.

Can a Crash be avoided?
The immediate consequence of the Crash was a couple of reports by official agencies, the most famous
one is the report by the Presidential Task Force on Market Mechanisms. Not surprisingly, these reports
call for more controls and regulations, especially of futures markets. The objective of these measures is
to reduce volatility.
The proposed measure included:
(1) Increase in margin requirements.
(2) Imposition of price limits.
(3) Differential taxing for short and long positions.
There is no evidence that margin requirements or price limits have any impact on stock price volatility.
The Crash was an international event. In Table X.7, we have countries with different regulations,
controls, taxes and trading systems. However, all experienced a significant negative shock in October
1987.

4.B.2

Average Linkages

We know, from the previous section, that markets display similar movements during periods of
extreme stock market movements. We also know that on average correlations are low to moderate. Are
there any patterns in those correlations? Some studies have pointed out certain regularities. Big
movements (not of the extreme kind) increase the correlation between international markets. There
seems to be a lead-lag relation between the U.S. market and the rest. Markets around the world seem to
pay close attention to the U.S. market open. An important question for investors and traders is the
following: are changes in stock markets related across time? For example, if there is a big (surprising)
change in the U.S., do we expect a similar change in Japan, the next day? That is, the nature of
correlations among international markets might be very important for investors. We should already
suspect that the answer to this question has very important portfolio implications.
4.B.2.i

Big Movements, Higher Correlations

When price changes are moderately big, transaction costs become relatively unimportant. Transaction
costs are a barrier for instantaneous arbitrage. Therefore, big price changes will bring world markets
together.
Example X.10: DuPont and IBM are U.S. stocks with listings on the Tokyo and the London exchanges. From
October 21, 1987 to June 15, 1988, the stock movements were tightly linked across international markets only if
the price movement was bigger than the transaction cost. For example, if the price movement was bigger than the
roundtrip transaction costs, the response to a change in Tokyo or London exchanges in the NYSE was one to
one. If the price movement was less than the roundtrip transaction cost, the response was close to zero.

X.32

Cross-market correlations tend to be positively correlated with measures of price volatility.

4.B.2.ii

Leads, lags and the U.S. market open

In a study published in the Journal of Financial and Quantitative Analysis, in 1989, Eun and Shim
examine the international transmission mechanism of daily movements of stock prices by estimating a
vector autoregressive (VAR) model of nine major stock markets. A VAR model is a simultaneous
equation model where all the variables are considered endogenous. The explanatory variables are
lagged endogenous variables. This approach allows us to decompose returns into a component due to
previous domestic innovations, and a second component representing previous foreign returns.
Their result shows that a substantial part of daily movements in innovations is attributable to other
countries. On the other hand, the magnitude of the impact by those countries is not symmetric; the U.S.
market is found to be by far the most influential country.
The Eun and Shim (1989) study uses daily data. Using daily data has a disadvantage, some markets in
the world have overlapping trading hours. It is very difficult for some markets to distinguish what is a
common movement (caused by world factors) and what is the influence of a specific international
market. Therefore a positive co-movement between NY and London (they share 2:30 hours of trading)
might either reflect common information or the influence of one specific market in the other.
In other studies, using hourly data, a U.S. market opening effect has been discovered. The intra-daily
evidence between New York and London is that they only affect each other around the time New York
is opening (9:30 AM, EST). At all other times, they behave like independent units. There is evidence
for an opening effect connected to NYSE.
4.B.3

Application: High Volatility, Correlations and Portfolio Choice

As we have seen above, in Section I, the home bias is a puzzle. The papers that discovered the home
bias puzzle examine gains to diversification using a time invariant correlation structure. Changes in the
correlation structure, no doubt, will affect the composition of optimal portfolios across time. Bicksler
(1974) shows that if correlations among international stock markets change over time, the gains from
international portfolio diversification may not necessarily be realized. As we have discussed in Chapter
V, variances and covariance are time-varying and ARCH models are popular ways to model them.
Several articles find that, in general, international correlations are unstable over time.
As discussed above, in this Section, higher interrelations are found when markets are more volatile.
Karolyi and Stulz (1995) find that while co-movements exhibit day-of-the-week effects,
macroeconomic shocks do not adequately explain these co-movements; neither does controlling for
industry effects. More importantly they find that covariances are high when returns on the national
indices are high and when "markets move a lot."
The above results suggest that while variances and covariances across markets are changing over time,

X.33

the spillover effects are also a function of the magnitude of the volatility shocks. In other words,
variances, covariances and correlations could be both time and state varying.
In Table X.8, the correlations between the U.S. and other major markets are calculated for two U.S.
volatility regimes: high volatility and low volatility. In other words, using the U.S. market as the home
market and based on the state of the variance of the home market return, the correlation between the
home and foreign markets is calculated. The results show that except in the case of Italy and Sweden,
higher correlations between the U.S. and other major stock markets are associated with periods of high
domestic volatility and vice versa. On average the correlations are 1 to 2.6 times higher in the U.S. high
volatility state.
The last column of Table X.8 also provides an indication of the proportion of times that a positive
foreign market return would have hedged a negative U.S. return when the U.S. is in a high volatility
regime. For instance, a positive return on the U.K. market would hedge a negative return in the home
(U.S.) market about 11% of the time when the home market is in the high volatility state. Similarly, a
portfolio with the U.S. and the world weighted in terms of market capitalization would have provided
no risk reduction benefits to a U.S. investor when the U.S. is in a high volatility regime.
As the U.S. investor is concerned, the benefits of diversification change depending on the state of the
volatility structure. For example, during the U.S. high volatility state, the correlation between U.S. and
EAFE returns significantly increases, but since the EAFE volatility is below the U.S. volatility, the
EAFE weight significantly increases. This result might explain a finding reported by Tesar and Werner
(1995). They point out that one would expect that U.S. investors would decrease their purchases of
equity from a market when that market covaries more strongly with the U.S. market. Tesar and Werner,
however, cannot find evidence for such a pattern in the data, which is consistent with the above
findings.
The results discussed above do not provide a complete explanation of the home bias phenomenon. If
investors, however, forecast that during high volatility periods the correlations increase, but the foreign
market variance increases more than the domestic variance, the portfolio weights would show an "expost" home bias.

X.34

TABLE X.8
Correlations between the U.S. and other markets in the two volatility regimes
High volatility

Low volatility

% Hedging

U.K.

0.7251

0.3124

11.11%

Germany

0.4275

0.2546

22.22%

Japan

0.2400

0.2344

22.22%

Canada

0.7983

0.6386

2.78%

Spain

0.2775

0.1481

19.44%

France

0.5471

0.2079

22.22%

Hong Kong

0.3208

0.2144

22.22%

Australia

0.2400

0.2344

19.45%

Denmark

0.3103

0.1980

16.67%

Belgium

0.3937

0.2114

22.22%

Italy

0.1160

0.1632

30.55%

Netherlands

0.7382

0.4454

19.44%

Norway

0.4735

0.2873

27.78%

Singapore

0.3762

0.2105

19.44%

Sweden

0.1730

0.2095

13.88%

Switzerland

0.5763

0.3318

13.88%

EAFE

0.5098

0.3474

11.11%

World

0.8628

0.7658

0.00%

V. Looking Ahead
In this chapter we have argued that investors should care about international equity markets for a very
simple reason: portfolio diversification. The moderate to low correlations makes the argument for
international portfolio diversification easy to state. Formal and informal capital controls and
information problems help to maintain the low correlations. These two issues are very relevant for a
specific group of international equity markets: emerging markets. Emerging markets are the equity
markets of the developing countries. These markets have enjoyed extraordinary returns in the 1990s,
which along with the very low correlations with the equity markets of developed countries makes them
very appealing to international investors. The next chapter characterizes emerging markets and
discusses the problems and risks associated with emerging markets.

X.35

Related readings
Asness, Clifford S., Israelov, Roni and Liew, John M. (2010). International Diversification Works
(Eventually), forthcoming Financial Analyst Journal.
Bicksler, J. L. (1974), "World, national and industry factors in equity returns," Journal of Finance, 29,
395-398.
Coval, J. D.; Moskowitz, T. J. (1999). "Home Bias at Home: Local Equity Preference in Domestic
Portfolios". Journal of Finance 54 (6): 20452074.
French, K. and J. M. Poterba (1991), "International Diversification and International Equity Markets,"
American Economic Review, 81, 222-226.
Harvey, C. R. (1991), "The World Price of Covariance Risk," Journal of Finance, 44, 111-157.

Heston, S. L. and K. G. Rouwenhorst (1994), Does Industrial Structure Explain the Benefits of
International Diversification? Journal of Financial Economics, 36, 3-27.
Jacquillat, B. and B. Solnik (1978), Multinationals are Poor Tools for International Diversification,
Journal of Portfolio Management.
Jorion, P. (1991), The Pricing of Exchange Rate Risk in the Stock Market, Journal of Financial and
Quantitative Analysis, 26, 363-376.
Karolyi, A. and R. M. Stulz (1995), "Why do Markets Move Together?", Working Paper, Ohio State
University.
Maxan, C. L. (1998), Do Country Risk Measures Foresee Financial Market Instability? The Asian
Crisis, Emerging Markets Quarterly, 2, 28-37.
Roll, R. (1989), `Price Volatility, International Markets Links and their Implications for Regulatory
Policies,' Journal of Financial Services Research, 3, 211-246.
Roll, R. (1992), Industrial Structure and the Comparative Behavior of International Stock Market
Indexes, Journal of Finance, 47, 3-42.
Senchack, A. J., Beedles, W.L., (1980). Is Indirect International Diversification Desirable?,
Journal of Portfolio Management, 6, 4957.
Solnik, B. (1974), Why Not Diversify Internationally? Financial Analyst Journal, 20, 48-54.
Tesar, L., Werner, I.M., (1995). Home Bias and High Turnover, Journal of International Money and
Finance, 14, 467-492.

X.36

Tesar, L., Werner, I.M., (1998). The Internationalization of Securities Markets Since the 1987

Crash, Brookings-Wharton Papers on Financial Services, The Brookings Institution.

Toshino, M., (2008) More evidence of Home Bias, Waseda University Institute of Finance, working
paper.
For a fascinating discussion of stock market crashes, bubbles and panics, read the different articles on
Crashes and Panics: The Lessons From History, edited by Eugene N. White, 1990.
For a good discussion of portfolio risk management issues in an international context, see Financial
Risk Management: Domestic and International Dimensions, by Philippe Jorion and Sarkis Joseph
Khoury, published by Blackwell, 1996.

X.37

Exercise:
1. Given the correlations in Table X.1, what are the best candidates for international diversification?
2. How can you take advantage of the closed-end country fund puzzle? Describe an arbitrage strategy.
3. You are the CFO of an MNC. International stock markets seem segmented. Where would you raise
capital?
4. Can you provide a transaction costs explanation to the co-movements in Table X.8? Explain.

X.38