An Application Of Scaled Asset Pricing Models To China’s A-Share Market

Static CAPM has long been proved invalid in the U.S. stock market. Subsequent consumption-based CAPM performs no better and at most times worse than CAPM, though its theoretical advantages. Recently developed conditional, or scaled asset pricing models have a good performance both empirically and theoretically. In those models scaling factors zt describes time-varying risk premia, and determines conditionally portfolios’expected returns.Seeking a desirable scaling variable thus, becomes a key issue in academic financial studies. This paper studies the effect of interaction between household housing wealth and their human capital, and believes they can act as a conditioning variable in the scaled asset pricing models, as those two are usually the most important assets for most normal families. The ratio between the two is defined as a collateral ratio, which changes the risk-sharing status in the society, and determines conditionally the return of stock portfolios. This paper first constructed the series of collateral ratio using macro data, then collected stock return information in China from the first season of 2002 to the fourth of 2012 and devised 25 portfolios sorted by capitalization and P/B ratio. By means of standard Fama-MacBeth procedure, this paper tested the applicability of CAPM, consumption-based CAPM and housing CAPM, along with their scaled versions in China’s A-share stock market. Major findings are as below:First, CAPM generally works well in China, which is totally different from current American situations. Also, consumption-based CAPM and housing CAPM performs no better and usually worse than CAPM; Second, Scaled asset pricing models with collateral ratio as a conditioning factor, improve their static counterparts considerably in the cross section, and their average pricing errors are much lower. This means scaled CAPM is also very successful in explaining China’s market. Third, although scaled CAPM works well in China, the predicted collateral effect is totally contradictory with that of the U.S.. In the U.S., when collateral ratio increases risk premia decrease, while in China when collateral ratio increases risk premia increase too. This implies that in China the high collateral ratio state is actually riskier while a higher collateral ratio in the U.S. lowers the risk of households. Why this happens might be accredited to some of China’s institutional factors, such a non-marketized interest rate system, and government’s frequent intervention in the real estate market. This obviously needs more in-depth research.