Aggregate Information in Different Anomalies: International Evidence

Abstract

This paper conducts a comprehensive empirical test of 20 prominent anomalies using the data from 32 countries/regions during the period 1999 to 2017. My results show that most of the anomalies become weaker or even diminish using data in recent years, both in the US and in other markets. This paper also compares five different information aggregation approaches to evaluate their performance of anomaly returns. The results show that although different aggregation methods may perform better in a specific country, on average the latent variable based approach and the Fama-MacBeth regression based approach outperform other approaches in capturing the return-related information. Moreover, results based on equal-weighted returns provide much more significant anomalies than those based on value-weighted returns.