LAUSR

This paper examines the performance of three nested factor pricing models for different market trends, such as bull, bear or consolidation, in China’s stock market. Specifically, the models we consider are the Capital Asset Pricing Model (CAPM), the Fama-French threefactor model and the Fama-French extended five-factor model. Empirical results show that these models can explain the time-series variations in excess returns on a range of portfolios in the bear markets reasonably well, but have more difficulty explaining the cross-sectional variations in returns across the portfolios. We further adopt instability tests based on Hansen’s statistics and recursive regressions, from which we derive the following two findings. First, we find the models to be more unstable in the bear and bull markets (trending markets) under time-series regression. Although the models in the time-series regressions usually perform well, this is due to the higher stock price synchronicity in trending markets, especially when it is bearish. Second, because the cross-sectional analysis of the Fama-MacBeth approach requires factor loadings to be estimated first by time-series regressions, the models inherit greater instability from trending markets. This causes the unitary parameter estimates to be less reliable, which in turn brings about difficulties in explaining the cross-sectional portfolio returns in trending markets.