This table lists the total points for each of the following political risk components out of the maximum points indicated. The final column in the table shows the overall political risk rating (the sum of the points awarded to each component).
A Government Stability 12 G Military in Politics 6
B Socioeconomic Conditions 12 H Religious Tensions 6
C Investment Profile 12 I Law and Order 6
D Internal Conflict 12 J Ethnic Tensions 6
E External Conflict 12 K Democratic Accountability 6
F Corruption 6 L Bureaucracy Quality 4
Country A B C D E F G H I J K L
Political Risk Rating
July 2008
Canada 8.0 8.5 11.5 10.5 11.0 5.0 6.0 6.0 6.0 3.5 6.0 4.0 86.0
Japan 5.0 8.0 11.5 10.5 9.5 3.0 5.0 5.5 5.0 5.5 5.0 4.0 77.5
China 10.5 7.5 7.0 9.5 10.0 2.5 3.0 5.0 4.5 4.5 1.5 2.0 67.5
United States 7.5 8.0 12.0 10.5 9.5 4.0 4.0 5.5 5.0 5.0 6.0 4.0 81.0
India 6.0 5.0 8.5 6.5 10.0 2.5 4.0 2.5 4.0 2.5 6.0 3.0 60.5
Table 25.8
Political risk points by component, July 2008
Source: International Country Risk Guide, July 2008, Table 3B.
U.S. investors have several avenues through which they can invest internationally. The most obvious method, which is available in practice primarily to larger institutional inves- tors, is to purchase securities directly in the capital markets of other countries. However, even small investors now can take advantage of several investment vehicles with an inter- national focus.
Shares of several foreign firms are traded in U.S. markets either directly or in the form of American depository receipts, or ADRs. A U.S. financial institution such as a bank will purchase shares of a foreign firm in that firm’s country, then issue claims to those shares in the United States. Each ADR is then a claim on a given number of the shares of stock held by the bank. Some stocks trade in the U.S. both directly and as ADRs.
There is also a wide array of mutual funds with an international focus. In addition to single-country funds, there are several open-end mutual funds with an international focus.
For example, Fidelity offers funds with investments concentrated overseas, generally in Europe, in the Pacific Basin, and in developing economies in an emerging opportunities fund. Vanguard, consistent with its indexing philosophy, offers separate index funds for Europe, the Pacific Basin, and emerging markets. Finally, as noted in Chapter 4, there are many exchange-traded funds known as iShares or WEBS (World Equity Benchmark Shares) that are country-specific index products.
U.S. investors also can trade derivative securities based on prices in foreign security markets. For example, they can trade options and futures on the Nikkei stock index of 225 stocks traded on the Tokyo stock exchange, or on FTSE (Financial Times Share Exchange) indexes of U.K. and European stocks.
Risk and Return: Summary Statistics
Statistics bearing on the efficacy of international diversification appear in Table 25.9 , which is divided into developed economies (Panel A) and emerging economies (Panel B).
The 44 index portfolios are value weighted and constructed of companies for which good data are available. Market capitalization is the sum of the market values of the outstanding stock of the companies included in each country index. Countries in each panel are ordered by market capitalization as of January 1, 2010.
Table 25.9 also includes average monthly excess returns for each index (returns in excess of the U.S. T-bill rate) over the period 2000–2009, standard deviation, country beta against the U.S., and correlation with U.S. returns. These statistics are computed using returns measured in U.S. dollars as well as in foreign currency. We use the table to develop insights into risk and reward in international investing.
Are Investments in Emerging Markets Riskier?
We pointed out in Chapter 24 that the appropriate measure of risk depends on whether one is evaluating the overall investment portfolio or an asset to be added to the exist- ing portfolio that might improve diversification. For the overall portfolio, standard devia- tion of excess returns is the appropriate measure of risk. In contrast, for an asset (here, a foreign-country index portfolio) to be added to the current portfolio (here, the domestic U.S. index portfolio), the covariance with (or beta against) the U.S. portfolio is the appro- priate measure.
Figure 25.3 ranks developed and emerging markets by standard deviation of excess returns. As candidates for complete investment portfolios, emerging markets are clearly riskier. However, Figure 25.4 , which measures risk using beta against U.S. stocks, paints a dif- ferent picture. Of the interest- ing BRIC group (Brazil, Russia, India, and China), only Russia and Brazil are clearly riskier.
Average Country-Index Returns and Capital Asset Pricing Theory
Figure 25.5 ranks markets by average excess returns over 2000–2009; it shows a clear advantage to emerging mar- kets. This is consistent with risk ranking by standard devia- tion ( Figure 25.3 ), but not with ranking by beta ( Figure 25.4 ), as would have been expected
0 2 4 6 8 10 12 14 16 18 20
1 6 11 16 21 26
Rank
Standard Deviation (% per month)
Developed markets Emerging markets
U.S. JapanDenmark Israel Austria
Finland Turkey
Russia Brazil
India China
Malaysia
Figure 25.3 Monthly standard deviation of excess returns in devel- oped and emerging markets, 2000–2009
Note: Developed and emerging markets are ranked by standard deviation of returns (low to high).
CountryWorld Capitalization*
Excess Returns (% per Month in U.S. Dollars)Excess Returns (% per Month in Local Currency) AverageStd. Dev.Beta/U.S.Corr./U.S.AverageStd./Dev.Beta/U.S.Corr./U.S. World37,19320.015.341.000.9720.015.341.000.97 A. Developed countries U.S.12,29920.205.141.001.0020.205.141.001.00 Japan3,27320.385.520.660.6120.465.340.620.60 U.K.2,7600.045.450.920.8720.014.540.770.87 France1,8280.216.551.090.8520.155.590.900.82 Canada1,4310.767.111.150.830.385.320.860.84 Hong Kong1,3510.356.740.850.650.356.730.850.65 Germany1,2660.247.821.280.8420.136.911.090.81 Australia1,1021.006.661.030.790.604.090.580.73 Switzerland1,0490.355.480.860.8120.054.830.740.79 Spain7730.717.121.110.800.346.120.910.77 Italy6630.186.811.020.7720.185.750.840.75 Korea6471.0510.231.470.740.908.331.040.64 Netherlands4590.196.981.140.8420.166.270.960.79 Sweden3980.368.831.380.800.107.471.010.70 Singapore3960.477.561.080.740.276.760.950.72 Belgium2480.117.641.110.7520.276.580.930.73 Norway2301.138.731.290.760.747.381.090.76 Finland18020.1210.681.430.6920.4510.321.240.62 Denmark1620.716.481.000.790.365.830.810.72 Israel1480.697.490.820.560.556.860.630.47 Greece11720.109.451.120.6120.468.440.930.57 Austria1010.808.501.180.710.397.191.000.71 Portugal930.276.580.800.6220.095.520.610.57 Ireland6420.547.591.080.7320.867.230.900.64 New Zealand330.567.120.950.680.144.660.490.54 Table 25.9 Risk and return across the globe, 2000–2009continued
CountryWorld Capitalization*
Excess Returns (% per Month in U.S. Dollars)Excess Returns (% per Month in Local Currency) AverageStd. Dev.Beta/U.S.Corr./U.S.AverageStd./Dev.Beta U.S.Corr./U.S. B Emerging markets Brazil1,1501.9110.931.570.741.557.431.030.71 India9921.3810.131.200.611.348.991.030.59 Russia6861.7112.031.460.631.5911.471.380.62 China*5720.969.131.080.610.969.121.080.61 Taiwan4640.138.680.990.590.097.910.870.57 South Africa3651.138.421.070.661.105.950.680.59 Malaysia2300.655.780.480.420.535.300.360.35 Turkey1981.3015.691.880.611.7112.731.330.54 Chile1961.096.710.810.620.954.990.520.53 Indonesia1841.6311.331.210.551.618.830.880.51 Thailand1430.9810.051.060.540.788.940.950.55 Columbia1352.609.830.920.482.538.200.570.35 Poland1091.0010.471.310.640.478.210.930.58 Philippines760.307.950.720.460.357.120.650.47 Peru652.239.931.030.532.159.761.000.53 Czech Republic462.118.930.970.561.447.730.760.51 Argentina410.8812.231.000.421.9712.580.900.37 Hungary301.2010.421.320.650.738.291.010.63 Pakistan251.4511.240.250.111.8210.850.220.10 Table 25.9— continued Risk and return across the globe, 2000–2009 *Billion of $U.S. as of January 1, 2010. Source: Datastream.
by applying the CAPM to world assets. According to the CAPM, country average returns should line up by their betas against the world portfolio, which is expected to be the most efficient portfolio globally.
Suppose that investors in each country actually were uninter- ested in international diversifi- cation. In that case, as explained in Chapter 9, expected excess returns on the index for each country, R C , would depend on home country variance:
E(RC)5AsC2
(25.4) where A is the country aver- age coefficient of risk aversion and sC2 is the variance of the
country-index excess return. So if risk aversion does not vary too much across countries, Figure 25.3 (in which average return increases with country standard deviation) would be consistent with a theory of capital asset pricing in a world absent international diversifica- tion motives.
In contrast, a simple version of a world CAPM would imply that the capitalization- weighted portfolio of world risky assets is “the” efficient portfolio, and individual-country- index expected returns should line up by beta against this portfolio. But this prediction ignores the fact that investors save for consumption in different currencies. We might get around this problem by assuming that investors would hedge currency risk and hence that world portfolio returns, as well
as betas against it, should all be estimated in local currencies.
But if investors were to follow this practice, they would find themselves with investment port- folios that are extremely heav- ily tilted toward foreign assets.
Even U.S. investors would have to hold two-thirds of their portfo- lios in foreign assets (and hedge all these currencies accordingly), as the U.S. stock market capital- ization is only about one-third that of the entire world.
In fact, we observe inves- tors in each country exhibiting home bias, that is, they seem to favor home-country assets rather than seeking pure efficient
⫺1.0
⫺0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Monthly Average Return
Developed markets Emerging markets
Ireland
U.S. U.K.
France
Israel
Norway Colombia Brazil
Russia India
China Taiwan
1 6 11 16 21 26
Rank
Figure 25.5 Average dollar-denominated excess returns of developed and emerging markets, 2000–2009
Note: Developed and emerging markets are ranked by average monthly returns (low to high).
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
1 6 11 16 21 26
Rank
Beta
Developed markets Emerging markets
Pakistan
U.K. U.S.
Germany Turkey Brazil Russia
India China Israel
Japan Korea
Figure 25.4 Country index dollar return beta on U.S. stocks, 2000–2009
Note: Developed and emerging markets are ranked from low to high beta against U.S. stocks.
d iversification. In addition to issues of psychology, regulation, and extra expense, home bias may also arise from the following motivation. 6 Investors evaluate their standard of living against a reference group that is most likely to consist primarily of their compatriots.
This alone is a reason to tie portfolio returns to the success of home firms that supply the labor/management income of the reference group.
On these grounds, we expect that country-index expected returns would be influenced by both beta against the world portfolio and the variance of home country returns. In addi- tion to these two variables, we might also consider country size (capitalization) as another variable that explains returns, since larger markets tend to exhibit better regulation and transparency.
Unfortunately, a regression of average country return on beta, variance, and size poses problems in implementation. First, statistical imprecision in estimates of beta and vari- ance degrades the regression estimates. Precision also suffers from the small number of countries (40–50) that offer sufficient data. The results of such regressions using data from Table 25.9 are shown in Table 25.10 . The first column includes all countries, and the coef- ficient on return variance is the only one to attain statistical significance. In this regression, the coefficient on beta actually has the “wrong” sign, implying that returns fall with beta.
Separate regressions for developed and emerging markets in the next two columns suggest these populations exhibit different behavior. Here, the coefficients on beta are positive, but neither is significant. Country size, measured by log of capitalization, has the expected sign in both regressions, but neither is significant. All we can conclude from these regres- sions is that neither model adequately explains the data.
Country return variance, rather than beta, may well be the dominant variable, suggest- ing the importance of home bias. Notice that we used returns measured in U.S. dollars in the regressions of Table 25.10 , so the returns would apply most directly to U.S. investors.
Investors in other countries would have to translate all returns to the home currency.
6 For a formal analysis of this idea see Peter M. De Marzo, Ron Kaniel, and Ilan Kremer, “Diversification as a Public Good: Community Effects in Portfolio Choice,” Journal of Finance 59 (August 2004).
Markets
All Countries Developed Emerging Coeffients Intercept
Ln(size) Beta on U.S.
Variance
0.67 0.00 1.16
20.05 20.06 20.08 20.33 0.95 0.34
0.94 20.50 0.15
t-statistic Intercept Ln(size) Beta on U.S.
Variance
1.26 0.00 1.41
20.60 20.79 20.45 20.75 1.15 0.50
2.94 20.60 0.30
R-square 0.30 0.31 0.28
Observations 44 25 19
Table 25.10
Regressions of country average monthly excess returns on size, beta, and return variance, 2000–2009
Benefits from International Diversification
Table 25.11 presents correlations between returns on stock and long-term bond portfolios in various countries. Panel A shows correlation of returns in U.S. dollars, that is, returns to a U.S. investor when currency risk is not hedged. Panel B shows correlation of returns in local currencies, that is, returns to a U.S. investor when the exchange risk is hedged.
First and foremost, these correlations as a whole suggest that international diversification should be considered, at least for active investors. Although correlations between most of the country stock portfolios shown in Table 25.11 are quite high, a few are low enough to imply meaningful benefits from diversification, particularly among bond portfolios and stock/bond portfolios.
Comparison of correlations between currency-hedged and unhedged portfolios shows that, at least for U.S. investors, some currencies should be hedged while some hedges are better forgone. Clearly, taking full account of currency risk would make construction of optimal diversified international portfolios cumbersome and expensive. This brings up a more basic question: Will passive investors benefit from international diversification in the future? We see in Table 25.12 that globalization seems to have encouraged a trend toward higher cross-country correlations. (This trend is the subject of the nearby box.) The lowest correlations are exhibited by emerging markets, but these too have been rising steadily.
Since emerging markets are also more volatile (higher standard deviations), lower correla- tions are needed to achieve significant benefits.
The observed high correlation across markets calls into question the common claim of large diversification benefits from international investing. This conventional wisdom is depicted in Figure 25.6 , which is based on data for the period 1961–1975. It suggests that international diversification can reduce the standard deviation of a domestic portfolio by as much as half (from about 27% of the standard deviation of a single stock to about 12%).
This improvement may well be exaggerated if correlation across markets has markedly increased, as data from recent years suggest, while standard deviations of country indexes have actually decreased. Still, while benefits from international diversification may be sig- nificant, we first need to dispose of a misleading, yet widespread, representation of poten- tial benefits from diversification.
It’s one of the golden rules of investing: Reduce risk by diversifying your money into a variety of holdings—stock funds, bonds, commodities—that don’t move in lockstep with one another. And it’s a rule that’s getting tougher to obey.
According to recent research, an array of investments whose prices used to rise and fall independently are now increasingly correlated. For an example, look no fur- ther than the roller coaster in emerging-markets stocks of recent weeks. The MSCI EAFE index, which measures emerging markets, now shows .96 correlation to the S&P, up from just .32 six years ago.
For investors, that poses a troubling issue: how to main- tain a portfolio diversified enough so all the pieces don’t tank at once.
The current correlation trend doesn’t mean investors should go out and ditch their existing investments. It’s just that they may not be “getting the same di versification”
they thought if the investment decisions were made some time ago, says Mr. Ezrati, chief economist at money- management firm Lord Abbett & Co. He adds that over long periods of time, going back decades, sometimes var- ied asset classes tend to converge.
One explanation for today’s higher correlation is increased globalization, which has made the economies of various countries more interdependent. International stocks, even with their higher correlations at present, deserve some allocation in a long-term investor’s hold- ings, says Jeff Tjornehoj, an analyst at data firm Lipper Inc.
Mr. Tjornehoj is among those who believe these correla- tions are a temporary phenomenon, and expects that the diversity will return some time down the line—a year or few years.
Source: Shefali Anand, “Investors Challenge: Markets Seem Too Linked,” The Wall Street Journal, June 2, 2006, p. C1. © 2006 Dow Jones & Company, Inc. All rights reserved worldwide.
A. Returns in U.S. dollars StocksBonds U.S.U.K.JapanFranceCanadaGermanyAustraliaU.S.JapanU.K.FranceCanadaGermanyAustralia Stocks U.S.1 U.K.0.871 Japan0.610.611 France0.850.900.601 Canada0.830.820.640.801 Germany0.840.860.550.960.771 Australia0.790.840.650.810.830.781 Bonds U.S.20.1320.030.0320.0920.1320.1520.061 Japan0.000.020.130.0320.010.010.040.451 U.K.0.190.400.230.310.270.260.370.500.381 France0.150.310.230.330.200.270.310.630.530.801 Canada0.430.530.440.540.630.500.560.410.310.580.631 Germany0.120.270.200.300.170.240.280.660.530.791.000.621 Australia0.430.530.430.530.510.470.630.480.460.720.810.760.801 Table 25.11 Correlation for asset returns: Unhedged and hedged currenciescontinued
B. Returns in local currency (eqivalent to U.S. dollar returns plus fully hedged currency risk) StocksBonds U.S.JapanU.K.FranceCanadaGermanyAustraliaU.S.JapanU.K.FranceCanadaGermanyAustralia Stocks U.S.1 Japan0.601 U.K.0.870.591 France0.820.630.891 Canada0.840.610.740.741 Germany0.810.570.840.950.701 Australia0.730.640.740.740.680.691 Bonds U.S.20.1320.1620.1320.2520.1120.2820.181 Japan20.0320.2720.0520.1620.1420.1220.090.361 U.K.20.0620.1320.0120.1320.0820.1520.070.790.381 France20.1820.1920.1020.2120.2020.2420.190.770.320.771 Canada20.0420.090.0120.070.0120.1220.120.830.380.770.741 Germany20.2120.2220.1220.2220.2220.2520.210.750.310.760.990.741 Australia20.2320.3020.2020.2720.2420.3320.270.780.390.710.720.750.711 Table 25.11—concluded Correlation for asset returns: Unhedged and hedged currencies
Misleading Representation of Diversification Benefits
The baseline technique for constructing effi- cient portfolios is the efficient frontier. A useful efficient frontier is constructed from expected returns and an estimate of the covari- ance matrix of returns. This frontier combined with cash assets generates the capital alloca- tion line, the set of efficient complete portfo- lios, as elaborated in Chapter 7. The benefit from this efficient diversification is reflected in the curvature of the efficient frontier. Other things equal, the lower the covariance across stocks, the greater the curvature of the efficient frontier and the greater the risk reduction for any desired expected return. So far, so good.
But suppose we replace expected returns with realized average returns from a sample period to construct an efficient frontier; what is the possible use of this graph?
The ex post efficient frontier (derived from realized returns) describes the portfolio of only one investor—the clairvoyant who actu- ally predicted the precise averages of realized returns on all assets and estimated a covari- ance matrix that m aterialized, precisely, in the
27 11.7
1 10 20 30 40 50
U.S. Stocks Global Stocks
Number of Stocks
Percent Risk
0 20 40 60 80 100
Figure 25.6 International diversification. Portfolio stan- dard deviation as a percentage of the average standard deviation of a one-stock portfolio
Source: B. Solnik, “Why Not Diversify Internationally Rather Than Domestically.” Financial Analysts Journal, July/August 1974, pp. 48–54.
Copyright 1976, CFA Institute. Reproduced and republished from Financial Analysts Journal with permission from the CFA Institute. All rights reserved.
Table 25.12
Correlation of U.S.
equity returns with country equity returns
Sample Period (monthly excess return in $U.S.) 2000–2009* 1996–2000* 1991–1995* 1970–1989**
World .97 .92 .64 .86
Sweden .80 .60 .42 .38
Germany .84 .66 .33 .33
France .85 .63 .43 .42
United Kingdom .87 .77 .56 .49
Netherlands .84 .63 .50 .56
Australia .79 .64 .36 .47
Canada .83 .79 .49 .72
Spain .80 .59 .51 .25
Hong Kong .65 .63 .33 .29
Italy .77 .44 .12 .22
Switzerland .81 .56 .43 .49
Denmark .79 .56 .36 .33
Norway .76 .58 .50 .44
Belgium .75 .49 .54 .41
Japan .66 .54 .23 .27
Austria .71 .53 .19 .12
*Source: Datastream.
**Source: Campbell R. Harvey, “The World Price of Covariance Risk,” Journal of Finance, March 1991.
actual realizations of the sample period returns on all assets. Obviously, we are talking about a slim to empty set of investors. For all other, less-than-clairvoyant investors, such a frontier may have value only for purposes of performance evaluation.
In the world of volatile stocks, some stocks are bound to realize large, unexpected average returns. This will be reflected in ex post efficient frontiers of enormous apparent
“potential.” They will, however, suggest exaggerated diversification benefits. Such (elu- sive) potential was enumerated in Chapter 24 on performance evaluation. It has no mean- ing as a tool to discuss the potential for future investments for real-life investors.
Realistic Benefits from International Diversification
While recent realized returns can be highly misleading estimates of expected future returns, they are more useful for measuring prospective risk. There are two compelling reasons for this. First, market efficiency (or even near efficiency) implies that stock prices will be difficult to predict with any accuracy, but no such implication applies to risk measures.
Second, it is a statistical fact that errors in estimates of standard deviation and correlation from realized data are of a lower order of magnitude than estimates of expected returns.
For these reasons, using risk estimates from realized returns does not bias assessments of the potential benefits from diversification.
Figure 25.7 shows the efficient frontier using realized average monthly returns on the stock indexes of the 25 developed countries, with and without short sales. Even when the (ex post) efficient frontier is constrained to preclude short sales, it greatly exaggerates the benefits from diversification. Unfortunately, such misleading efficient frontiers are still presented in articles and texts on the benefits of diversification.
A more reasonable description of diversification is achievable only when we input rea- sonable equilibrium expected returns. Absent superior information, such expected returns are best based on appropriate risk measures of the assets. The capital asset pricing model (CAPM) suggests using the beta of the stock against the world portfolio. To generate expected excess returns (over the risk-free rate) for all assets, we specify the expected
World Austria
Korea New Zealand
−0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
0 2 4 6
U.S.
8 10 12
Standard Deviation (% per month)
Realized Average Monthly Excess Return (%)
Efficient frontier with short sales Efficient frontier w/ no short sales Country portfolios
Figure 25.7 Ex post efficient frontier of country portfolios, 2001–2005
excess return on the world portfolio. We obtain the expected excess return on each asset by multiplying the beta of the asset by the world portfolio expected excess return. This procedure presupposes that the world portfolio will lie on the efficient frontier, at the point of tangency with the world capital market line. The curvature of the efficient frontier will not be affected by the estimate of the world portfolio excess return. A higher estimate will simply shift the curve upward.
We perform this procedure with risk measures estimated from actual returns and fur- ther impose the likely applicable constraint on short sales. We use the betas to compute the expected return on individual markets, assuming the expected excess return on the world port- folio is .6% per month. This excess return is in line with the average return over the previous 50 years. Varying this estimate would not qualitatively affect the results shown in Figure 25.8 (which is drawn on the same scale as Figure 25.7 ). The figure shows a realistic assessment that reveals modest but significant benefits from international diversification using only devel- oped markets. Incorporating emerging markets would further increase these benefits.
Are Benefits from International Diversification Preserved in Bear Markets?
Some studies suggest that correlation in country portfolio returns increases during peri- ods of turbulence in capital markets. 7 If so, benefits from diversification would be lost exactly when they are needed the most. For example, a study by Roll of the crash of
7 F. Longin and B. Solnik, “Is the Correlation in International Equity Returns Constant: 1960–1990?” Journal of International Money and Finance 14 (1995), pp. 3–26; and Eric Jacquier and Alan Marcus, “Asset Allocation Models and Market Volatility,” Financial Analysts Journal 57 (March/April 2001), pp. 16–30.
World U.S.
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 2 4 6 8 10 12
Standard Deviation (% per month)
Expected Monthly Excess Return (%)
World CML Efficient frontier w/ no short sales
Efficient frontier w/ short sales Country portfolios
Figure 25.8 Efficient frontier of country portfolios (world expected excess return 5 .6% per month)