Estimating Betas and Stock-Return Correlations From Monthly Data: A Warning Note.

The empirical finance literature makes extensive use of 'monthly' stock returns, where a monthly return is the change in stock price between one particular day of the calendar month - which we term the reference day - and the corresponding day of the following month. We show that estimates of betas and stock-market correlations are highly sensitive to the choice of reference day and we suggest that studies based on such estimates can be unreliable. We support this claim by carrying out two small-scale empirical studies showing in each case that the results of critical tests are dependent upon the choice of reference day.