New findings regarding return autocorrelation anomalies and the importance of non-trading periods

Abstract

In this paper, differences in return autocorrelation across weekdays have been investigated. Our research provides strong evidence of the importance on non-trading periods, not only weekends and holidays but also overnight closings, to explain return autocorrelation anomalies. While stock returns are highly autocorrelated, specially on Mondays, when daily returns are computed on a open-to-close basis, they do not exhibit any significant level of autocorrelation. Our results are compatible with the information processing hypotheses as an explanation of the weekend effect.