What approaches to global diversification work best? In their July 2011 paper entitled “What Matters in International Equity Diversification? “, Chun-hung Chen, Tom Goodwin, and Wenling Lin use mean-variance spanning and optimization tests of indexes to compare benefits of alternative approaches to global diversification of the equity portion of an investor’s portfolio. Specifically, they investigate potential contributions to global diversification from: (1) American Depository Receipts (ADR); (2) multinational firms; (3) global versus country-specific delineation of large capitalization versus small capitalization; and, (4) non-U.S. developed small and micro capitalizations, emerging market small and micro capitalizations and frontier markets. Using monthly returns for the Russell 1000 and Russell 2000 indexes for a benchmark U.S. portfolio, and for various global and country indexes as available through February 2011, they find that:
- The clearest and strongest finding is that global diversification via any of the international indexes considered provides a statistically and economically significant improvement to a U.S.-only portfolio. However, there is no clear winner among these indexes. In other words, home bias is costly even in an era of rising correlations. Specifically:
- Over the available July 2002 through February 2011 sample period, country stocks beat ADRs by a nose, with both substantially enhancing return and slightly reducing the risk of a U.S.-only portfolio.
- Over the available August 1996 through February 2011 sample period, these is some evidence that country stocks offer better diversification of a U.S.-only portfolio than multinationals.
- Over the available August 1996 through February 2011 sample period, while the cut-off between small and large varies widely across countries, there is not a strong case for differentiating based on country-specific versus global size rankings. The latter may offer slightly better diversification potential to a U.S.-only portfolio.
- Over the available January 1998 through February 2011 sample period, extension to emerging market small stocks enhances performance of portfolio already diversified across large and small U.S. stocks and large non-U.S. developed and emerging market stocks, even after consideration of liquidity. However, the incremental diversification and return benefits of further extension to non-U.S. developed market small and micro stocks, emerging market micro stocks and frontier markets are small.
In summary, evidence from mean-variance tests over available sample periods indicates that: (1) almost any international exposure enhances the performance of a U.S.-only portfolio; and, (2) emerging market small stocks (capitalization $1-5 billion) is likely a good way to enhance a portfolio that already includes small and large U.S. stocks and large non-U.S. developed and emerging market stocks.
Cautions regarding findings include:
- Sample periods are short relative to such potential drivers of cross-country correlations as economic cycles and political conflicts.
- Using indexes for analysis extends sample periods backward in time, but inclusion of the management and trading costs required to create and maintain liquid index-tracking assets (as reflected in mutual funds and exchange-traded funds) may affect results.
- Equity class correlations are not particularly stable over time and have generally been increasing (consider “Effects of Creeping Indexation?”). Continuation of this trend would dampen diversification benefits compared to the sample periods.
- The methodology assumes that index return distributions are well-behaved (Gaussian-like). To the extent that the distributions are wild, interpretations of the statistics applied break down.
