Extreme Bound Analysis of Emerging Stock Market Anomalies

Abstract

Studies of emerging stock market anomalies are based on underspecified models. Extreme bound analysis (EBA), a technique to remedy specification bias, indicates that no anomaly is robust, given panel data covering sixteen countries from March 1988 through January 1995. Only under a relaxed decision rule does the author find that five of the fifteen factors studied are sturdy: price/book long–run lagged returns, population demographics, country risk, and relative market size. What is more sobering, time series EBA produces no sturdy aggregate determinants.