This paper incorporates two specific features in the test of capital asset pricing model (CAPM). The first, or empirical, one is to allow the systematic risk β to come from two different regimes to capture the instability found in the previous studies. The second, or institutional, one is to consider the censoring effect caused by the implementation of price limit regulation. Under price limit regulation, the observed returns need not be equal to the equilibrium ones if the closing prices reach limits. The caused econometric problems seem to be very challenging but can be resolved by the Gibbs sampler with data augmentation algorithm. A simple illustrative example is provided by using the data from the Taiwan Stock Exchange where the price limit system is adopted. The general results suggest that β's are unstable over time and the data may be consistent with CAPM in one regime but inconsistent in the other regime.
International Review of Economics and Finance 12(3), pp.305-326