Estimating conditional betas and the price of risk for a thin stock market

This paper examines the Sharpe-Lintner Capital Asset Pricing Model (CAPM) in which time-varying-parameter models are altemative to the static market model. Prior evidence does not support the CAPM and suggests that market risk is not priced or the price of the beta risk is significantly negative for a thin European stock market, e.g. the Finnish stock market. We show that this phenomenom is due to static ordinary least squares beta estimates that are spurious. We reduce the errors-in-variables problem by estimating the firm-specific betas using the Kalman filter technique and employ the forecasted beta values in cross-sectional analysis. It tums out that in our analysis of pooled data the sign for caefficient of the price of risk becomes positive and we are no longer able to reject the mean-variance efficiency of the market index. The data cavers all Finnish cammon stocks listed on the Helsinki Stock Exchange throughout the years 1972-1989.