Choose A Model Empirical Analysis In China's Stock Market

Chinese market presents an ideal testing ground for the quantitative equity portfolio management that is commonplace in developed countries. With the quarterly financial data and monthly stock market data from January 2002 to December 2010, we test the performances of the hypothetical equal-weighted and value-weighted top and bottom portfolios based on each of the 14 factors’ Z-score screening process, then use the short list of factors to formulate an equally weighting aggregate Z-score screening model and an optimal weighting aggregate model and test the performance of hypothetical value-weighted top and bottom portfolios based on the models. We provide extremely detailed factor by factor results to show the exact stock selection ability of each factor and the two aggregate score screening models.The overall performance for our stock selection model are quite encouraged and evident in the rationality and mature of China’s stock market. Both the score screening models based on monthly, quarterly and semiannual rebalancing frequencies provided relative high top bottom portfolio annualized return spread and excess return compared with benchmark, with high consistency and even lower levels of portfolio risk and systematic risk. According to the bull and bear market effects, the top portfolios significantly outperformed the bottom portfolios in the bear market, but in some period of the bull market such as 2006 and 2007, our model cannot provide sufficient evidences to add value.Apart from the models, we also generates some findings from the univariate back-testing results of each factor. The Shenzhen A-share market has a significant large capitalization size effect recent years, as a result, the valuation factors such as E/P and B/M ratio do not have acceptable stock selection ability. However, the accounting factors, especially the operating profitability factors such as ROE, ROA and ROS all shown relative high stock selection ability; while the operating efficiency factors provide the middle selecting ability compared with others, besides, all the accounting factors shown relative high correlation co-efficiency. Shorter term momentum factors generate a large return premium of around 20% a year compared with benchmark portfolio, but the momentum effect does not seem to persist in longer duration strategies and has high correlation with other factors.