Abstract: Evidence on the Presence and Causes of Serial Correlation in Market Model Residuals

Extract

The paper first presents evidence on common stock returns autocorrelation and on the deterioration of the market model R² as the returns measurement period is shortened. The empirical analysis uses returns intervals, ranging from 1 to 20 days, spanning an identical 1000 day period for a random sample of 20 NYSE firms in the Standard and Poor's 500. It is then shown that the negative R², differencing interval relationship can be explained by the joint occurrence of negative autocorrelation in market index returns.