Stock return autocorrelation, beta, and data frequency

In this thesis, I empirically test the autocorrelation function of stock returns and how the frequency of data affects the measurement of the beta of stock returns. Different data frequencies, from daily to monthly, of stock and market returns from the Center for Research in Security Prices over a 30-year period from 1983 to 2012 are used in the empirical analysis. I find that infrequent rebalancing generates a certain pattern in the autocorrelation function of stock returns and that return autocorrelations can switch sign and become positive. Furthermore, idiosyncratic risk strongly affects the detection of autocorrelation. In addition, the stock beta increases with the measurement time interval. The findings suggest that beta depends heavily on the shape of stock returns' autocorrelation function due to short-term stock reversal