Time-varying risk and long-term reversals: A re-examination of the international evidence

Long-term reversals (LTRs) have been extensively documented as existing in international markets. The predictability of returns opens the question whether investors should continue to diversify across international markets. Such predictability also has implications for the corporate decision of where and when to raise capital. This paper re-examines the international evidence for long-term contrarian profits using a longer time horizon than all previous studies (81 years of data from 1925 to 2005), and finds that the long-term contrarian anomaly disappears when time-varying (conditional) alphas are considered. This is true even without transaction costs. For the marginal trader, conservative transaction costs subsume LTR profits, whether returns are risk-adjusted or not. This suggests that benefits from actual trades on LTRs do not negate a strategy based on diversification. Although time-varying beta has been extensively documented as ineffective in describing LTRs, this paper demonstrates that it is important to control for time variation in alpha. A full conditional capital asset-pricing model captures the LTR premium. This suggests that macroeconomic factors are important for understanding LTRs and their associated risk.