Improving the Parkinson Method of Estimating Security Price Volatilities.

The author proposes a new method for estimating the volatility parameters of security prices, which is an improvement of the estimation method by M. Parkinson (1980). The author assumes that the security prices follow the geometric Brownian motion. However, contrary to the setting of Parkinson, the geometric Brownian motion may have nonzero drift terms. The author shows that the efficiency of his estimator is about 10 in comparison with the standard sample variance estimator. Since the efficiency of the estimator by Parkinson is about 4.91, his estimation method may considerably improve the estimation methods already known in financial economics. Copyright 1992 by University of Chicago Press.