~U T T E R V~I 0 R T H Journal of lnternational Money and Finance, Vol. 14, No. 4, pp. 597-615, 1995 Elsevier Science Ltd I'~[E ! N E M A N N 0261-5606-95-0005-4( ) Printed in Great Britain A search for long memory in international stock market returns YIN-WONG CHEUNG University of California, Santa Cruz, CA 95064, USA and City Universityof Hong Kong, Hong Kong AND KON S LAI* California State University, Los Angeles, CA 90032, USA A major issue in financial economics is the behavior of stock returns over long as opposed to short horizons. This study provides empirical evidence from the perspective of long memory analysis. International evidence on long memory is explored using the Morgan Stanley Capital International stock index data for eighteen countries. Two tests that are robust to short-term dependence and conditional heteroskedasticity are employed: a modified rescaled range test and a fractional differencing test. The empirical results in general provide little support for long memory in international stock returns. The findings are not sensitive to inflation adjustments in stock returns, data sources, and statistical methods used. (JEL G12). The behavior of stock returns over long as opposed to short horizons has been a hotly contested issue. Shiller (1984) and Summers (1986) show in simple models of fads that large and slowly decaying swings in stock prices away from fundamental values can occur, without significant autocorrelation in short-term returns. According to the Shiller-Summers models, market inefficiency can exist, but statistical tests on short-term returns will fail to detect it. Stambaugh (1986) observes, nonetheless, that the long swings away from fundamental values suggested by the Shiller-Summers models imply the presence of nega- tive autocorrelation in long-horizon returns. This is because, to the extent that the swings away from fundamental values are not permanent, prices will be mean-reverting and returns will exhibit negative serial correlation over long horizons. Fama and French (1988) examine autoregressions of multiperiod returns on diversified portfolio of NYSE stocks and report empirical findings seemingly consistent with the Shiller-Summers models. It is reported that while autocor- relations in returns are close to zero at short horizons, they become strongly *The authors are grateful to two anonymous referees for useful comments and suggestions. An earlier version of the paper was circulated under the title 'A search for long memory in stock market returns.' This research was supported by the GICES.