By G. Andrew Karolyi, Professor of Finance, Fisher College of Business, Ohio State University

An understanding of the magnitude and dynamics of return correlations among international stock markets is critical for a sound global asset management program. How low these correlations are among different markets, of course, limits the potential benefits to global investors of portfolio risk diversification strategies. But, knowing how they change over time can also be an important ingredient of a successful strategic or tactical global asset allocation program. The objective of this article is to survey briefly the most recent evidence on measuring international stock return correlations. I outline a series of facts that researchers have uncovered. I also address a genuine concern expressed by market regulators, policymakers, issuers and investors alike: do recent increases in equity market correlations reflect contagion effects?

Research on international stock market correlations and its usefulness in making the case for the benefits of international diversification of risk has its beginnings in the 1970s. For a number of years, studies attempted to understand the factors underlying the low international correlations by focusing on institutional factors and, specifically, the fact that national market indexes have a very different industrial composition. After all, risk diversification by investing in the Australian and Canadian market for U.S. investors may not stem from their economic growth rates, monetary/fiscal policies or exchange rate movements, but from the fact that they are heavily resource-based markets. The early evidence had clearly documented that country factors were more important than industry factors, but the debate was sparked again in the 1990s by a series of papers that re-established the growing importance of industry factors. No doubt this is an intuitive outcome with the growing integration of international equity markets through the growth in cross-border listings, international mutual and closed-end country funds and the introduction of new currency blocs. Researchers have learned three key facts.

FACT 1: COUNTRY FACTORS STILL DRIVE INTERNATIONAL STOCK RETURN CORRELATIONS, BUT INDUSTRY FACTORS ARE GROWING IN IMPORTANCE.
Correlations in international equity returns are unstable over time. Important contributions have established that these correlations dynamics have interesting regularities, such as slowly autoregressive patterns at the monthly and even intraday frequencies. What these studies have been less successful in uncovering is a systematic pattern that relates closely to economic fundamentals, such as changes in interest rates, dividend yields, exchange rate changes, capital flows, liquidity and macroeconomic factors. The inability of these fundamental factors to capture the time-variation in equity correlations has led some to question whether behavioural forces, like contagion effects, play a role.

FACT 2: STOCK MARKET CORRELATIONS VARY OVER TIME, BUT IN A WAY ONLY WEAKLY RELATED TO FUNDAMENTAL MACROECONOMIC AND CAPITAL MARKET FACTORS.
Several recent studies have shown that the instability in international equity correlations is associated with periods of high stock market volatility and, particularly, during bearish market volatility. This threshold effect (higher correlations with larger returns) and asymmetric effect (higher correlations with large negative or bad-news returns) is important for investors as it implies that the benefits of the safety net of international diversification is lost when it is needed most. Some studies empirically explore these patterns with high-frequency intraday data, while others define formally and measure the concept of extreme correlations using extreme-value statistical theory.

FACT 3: INTERNATIONAL CORRELATIONS OF LARGE STOCK RETURNS, ESPECIALLY NEGATIVE ONES, ARE HIGHER THAN THOSE FOR USUAL RETURNS.
Could the magnitude and dynamics of correlations reflect contagion effects? That fundamental factors are only weakly related to these dynamics and that these extreme, asymmetric patterns exist has led experts to suggest that contagion--rational or irrational--can be the only remaining explanation. Dornbusch, Park and Claessens (2000) offer just such a definition for contagion:

"Contagion, in general, is used to refer to the spread of market disturbances--mostly on the downside--from one country to the other, a process observed through co-movements in exchange rates, stock prices, sovereign spreads and capital flows."

Some researchers caution that inferences about contagion effects from time-varying correlations may be flawed because of natural statistical biases due to time-varying volatilities, omitted variables and endogeneity. Still others propose a new framework of analysis for contagion that draws on co-incidences of extreme returns in international markets, but, more importantly, explicitly does not draw on traditional measures of stock return correlations. Their results also suggest that their measures of international financial contagion may not be as large as what others may have perceived. Until researchers get a better handle on the economic importance of global stock returns correlations and, especially, the association with contagion, caution is advised. Stay tuned.

ADDITIONAL READINGS
Bae, K.H., G. A. Karolyi, and R. M. Stulz, 2000, "A New Approach to Measuring Financial Contagion," Ohio State University working paper.

Cavaglia, S., C. Brightman, and M. Aked, 2000, "The increasing importance of industry factors," Financial Analyst Journal, September/October, 41-54.

Forbes, K. and R. Rigobon, 2000, "No contagion, only interdependence: Measuring stock market comovements," MIT working paper.

Griffin, J. and G. A. Karolyi, 1998, "Another look at the role of the industrial structure of markets for international diversification strategies," Journal of Financial Economics 50, 351-373.

Longin, F. M., and B. Solnik, 2000, "Correlation structure of international equity markets during extremely volatile periods," Journal of Finance, forthcoming.