無 The Partial Adaptive Estimation of CAPM with Censorship in an Emerging Market
The Department of International Trade National Chengchi University
The daily data of stock returns in Taiwan stock market suffer from thin trading, price limits and non-normality that either cause specific estimation problems due to the daily characteristic or violate the assumption of traditional CAPM. These violations of the traditional market model could cause serious biases in estimation of beta. This paper takes use of the Aggregate Model, Partial Adaptive Model, and Two-limit Logit Model to tackle the problems resulting from the violations. The empirical results are consistent with previous literatures, and indicating that the partial adaptive estimator with censorship has the highest likelihood value. Meanwhile, the result also reveals that size and liquidity do play important roles in affecting the behaviors of stock returns in different ways.