A Bias-Correcting Procedure For Beta Estimation In The Pres

That beta estimates are biased in markets characterized by thin trading is well known. Two techniques designed to reduce this bias, one by Dimson (1979) and the other by Scholes and Williams (1977), involve the aggregation of leading and lagged beta coefficients calculated from measured returns to arrive at consistent beta estimates. An alternative method is proposed for obtaining consistent estimates of beta in the presence of thin trading. The procedure is a statistical approach, but security-specific information is required to implement it. The proposed estimator is tested on simulated data, and the results are compared to the methods of Dimson and Scholes and Williams. The new estimator is found to have approximately the same bias as the other estimators, but it has a considerably lower variance.