States. Thus, most internationally diversified equity portfolios would have had a
lower return than the U.S. market index over this period. During this period interna-
tional diversification had the advantage of lowering risk but resulted in lower aver-
age returns.
The results for long-term bonds are similar. The equally weighted portfolio of
country return indexes (excluding the United States) did slightly worse than the U.S.
market index. The value-weighted portfolio performed better. This was due primarily
to the performance of Japanese bonds. In yen, Japanese bonds returned about 8.13%
but over this period, the dollar value of the yen increased by 3.62% resulting in an
11.75% return to U.S. investors. A fair number of countries underperformed the U.S.
bond market. Thus many international portfolios would have also underperformed a
portfolio of U.S. bonds.
For three-month T-bills the return on the equally weighted index was slightly
worse and value-weighted index was slightly better than the return on U.S. T-bills.
Given the higher risk discussed earlier, many international portfolios would have
been inferior to an exclusive U.S. portfolio.
Although these results are appropriate for the period discussed, it is useful to ex-
amine other periods. Solnik (1988) studied equity indexes for 17 countries for the
years 1971–1985. For all but two countries the return on the foreign index expressed
in dollars was greater than the return on the U.S. equity index. The exchange gain
from holding foreign equities added 0.2% on average to this return. For long and
short bonds only, Canada and the United Kingdom had a lower return when return
was expressed in U.S. dollars. For bonds, however, a major factor contributing to the
return being above the U.S. return was exchange gains. The 1980s was a better pe-
riod for non-U.S. markets and many international portfolios would have outper-
formed their U.S. counterparts.
For portfolio decisions, estimates of future values of mean return, standard devia-
tion, and correlation coefficients are needed. The correlation coefficients between in-
ternational markets have been very low historically relative to intracountry correla-
tions. As Europe integrates its markets and as all countries move toward greater
integration, these coefficients are likely to rise.^5 However, they are still likely to be
low relative to intracountry correlation. For example, the correlation coefficient be-
tween countries whose economies are relatively highly integrated, such as Canada
and the United States, the Benelux countries, or the Scandinavian countries is still
much lower than the intracountry correlation coefficients. Thus international diversi-
fication is likely to continue to lead to risk reduction in the foreseeable future. How-
ever, we know of no economic reason to argue that returns in foreign markets will be
higher or lower than for domestic markets.
11.6 EFFECT OF EXCHANGE RISK. Earlier we showed how the return on a foreign
investment could be split into the return in the security’s home market and the return
from changes in exchange rates. In each of the prior tables we separated out the effect
of changes in the exchange rate on return and risk. In Exhibit 11.11, the column enti-
tled “Exchange Return or Exchange Risk” calculated the effect of converting all cur-
rencies into dollars. Obviously if we were presenting the same tables from a French
11.6 EFFECT OF EXCHANGE RISK 11 • 13(^5) In particular, exchange rates between European currencies are fixed. Although European currencies
will continue to fluctuate with the U.S. currency, any advantage in diversifying across currencies will be
eliminated.