Stock Portfolio Returns, Autocorrelations, and Nonsynchronous Trading

Abstract

Short-horizon portfolio return autocorrelation is often attributed to nonsynchronous trading. Nonsynchronous trading arises as last trades of most stocks occur randomly before the market close. However, prominent studies cast doubt on whether it can entirely explain the observed magnitude of index autocorrelation. We study if there is a market wide component to nonsynchronous trading, as it is often viewed as idiosyncratic to a stock. The results indicate that there is a significant market-wide factor and all stocks are sensitive to this factor. Further, higher size decile stocks are far more sensitive to the market-wide factor than lower size deciles. Theoretical models also imply that portfolio return autocorrelation should increase in lagged dispersion in last trade time; but we find no such sensitivity for equally and value weighted portfolios. Similar to the known relationship between stock price changes and volume, we find reliable evidence that longer trading days garner positive returns while shorter trading days yield negative or zero returns.