Do emerging markets still provide effective diversification? Evidence from ETF correlations.
Foreign stock markets—and especially emerging stock markets—are often suggested as a means for American investors to diversify their portfolios. The possible lack of correlation between emerging markets and the American stock market suggests that combining securities from these two markets should reduce the variability of the return experienced by an American portfolio. Such diversification then reduces the portfolio's risk without necessarily reducing its return.
This study examines the correlation between U.S. domestic equity markets and exchange-traded funds (ETFs) specializing in emerging financial markets. The results generated correlation coefficients that consistently exceeded 0.6 between the domestic market and the emerging market ETFs. These results suggest there is at best only marginal benefit from using emerging market ETFs to diversify a domestic equity portfolio.
Risk reduction through the construction of diversified portfolios is the foundation of modern portfolio theory. Harry Markowitz's pioneering work (1952) developed a theory of portfolio construction that maximized returns for each level of risk. Markowitz's model led to the development of financial models that answered questions such as how many securities are required to produce an optimal, well-diversified portfolio and what characteristics are necessary to reduce risk.
The answer to the first question may be surprisingly few. For example, Evans and Archer (1968) found that a portfolio of 8 to 10 securities was reasonably well diversified and that the inclusion of more securities in the portfolio only had a marginal improvement in diversification. Adorna, Carver and Mayo (2008) found that portfolios composed of ETFs required even fewer securities to achieve similar levels of diversification.
While the number of securities to achieve diversification may be modest, the returns on the securities cannot be highly correlated. Correlation coefficients relating returns need to be low or even negative for diversification to occur. For example, combining AT&T and Verizon in a portfolio should have little impact on diversification since their returns should be highly correlated.
The argument that investing in securities from emerging markets will diversify an American portfolio has been supported by evidence that there is low correlation between the returns in these markets and the returns earned through investments in American markets. Keppler and Lechner (1997) reported low correlation for the period 1989-1995. For example, they found correlation coefficients with American returns of 0.08, 0.14 and 0.00 for Chile, Korea, and India, respectively (p. 96). Mark Mobius (1996) reported similar findings for the period 1991-1995. His correlation coefficients for Chile, Korea and India were 0.26, 0.00 and -0.08 (pp. 216-217).
A more recent study by Goetzman, Li and Rouwenhorst (2005) found that the benefits of global diversification continued to exist. This encompassing study extended the time period through 2000, and found correlation coefficients that were relatively low, often less than 0.3 for emerging markets. These low correlations in all three studies suggest that the inclusion of emerging market securities would contribute to the diversification of an American portfolio.
How an American investor can take advantage of these findings poses a problem. The investor cannot buy all the securities in each index. Specialized closed-end investment companies and mutual funds are one possibility. These funds, however, do not necessarily track a foreign index, so the results depend on the asset selection of the funds' managers. A particular fund may not produce the same correlations generated by the index.
ETFs, however, may offer a solution.
Methodology And Data Analysis
This study examines the correlation between ETFs investing in emerging markets and the SPDR (Spider), which tracks the performance of the Standard & Poor's 500 Index. Since the SPDR replicates the S&P 500, it reflects the performance of the U.S. domestic equity market. Emerging market ETFs hold securities in less-developed countries and track the returns earned by those markets.
The correlation coefficients are calculated between the returns on the SPDR and on 13 emerging market ETFs. (A list of the specific ETFs is provided in Exhibit 1.) The period covered is 2003 through 2007, and the returns were computed using adjusted monthly prices. The period covered is short because emerging market ETFs are a relatively new phenomenon. Only three of the ETFs have been in existence for four or more years, and several (e.g., GAF, RSX and FNJ) commenced trading during 2007, so only six months of price data were available.