Efficient estimation of price adjustment coefficients

The price adjustment coefficient model of Amihud and Mendelson (1987) is shown to be suitable for estimation by the Kalman filter. A techique that, under some commonly used conditions, is asymptotically efficient. By Monte Carlo simulations it is shown that both bias and mean squared error are much smaler compared to the estimator proposed by Damodaran and Lim (1991) and Damodaran (1993). A test for the adeqacy of the model is also proposed. Using data from four minor, the nordic countries except Iceland, and one major, US, stock markets the results are that the markets under-react to new information, but for most of the nordic countries, the model is not adequate.