Investment Decision Research Which Based On Multi-factor Model

China’s capital markets are mainly based on a-share market. It has more than20year’s period which has great active china’s economy, however, in the past20years,several scholars, organizations and investors constantly query whether China’s a-sharemarket follow the laws of marketing. Confront with these people who chanted thepolicy market, uselessness of technical analysis, author also think deeply. I think, therewere malicious speculation and Chinese characteristics markets’ develop featureconstraint China’s a-share market’s value-discover function, but with the development ofour government’s policy and the advance of normalization of stock market, China’sa-share market’s basic function will be more and more normalization. Especially withthe share-merger reform push on and margin mechanism pull in, they will have bigaffects on changing Chinese stock investment conception and the traditional profitpattern. There are strong evidence that the development of quantitative trading and thedevelopment of program trading. This paper based on multi-factor model, pointing at along history period a-shares’ data, using computer processing followed forecast stocksrate of return and contrast the data after it occurrence, attempting to find some evidencebased on these data.This thesis based on China’s a-share market related companies’ transaction data andfinancial data from1996years to2011years. Then establishment of trading strategiesbased on multi-factor model in order to empirical test the achievements. Due to lack offinancial data or abnormal fluctuate of trading data (for eggs. suspension, resumption,some stock became st-stock and so on), this paper use Vfp6.0and Eviews5.0toprocessing data, including data validation, data cleaning, wealth of data and treatment ofabnormal data. It was limited due to lack of data or other realities, at last, the evidencechose the data which contains14factors concert the12month’s mobile window’s.The evidence shows, firstly, the results of statistical test can not commendablesupport this factor decision, but the realistic forecasting effect are good contrary toexpectations. secondly, these data were demarcated by share-merger reform. It shows,before the reform of non-tradable shares, the evidence don’t support risk premium theory, in other words, risk assets get high profit should not at the cost of high systemrisk; after the reform of non-tradable shares, information transfer faster in the market, itincreases China’s a-share market’s validity, so, the multi-factor model’s earning becameweaker and weaker. Thirdly, this paper choose Fama-French three-factor model toinspection this strategies, the evidence shows, high return portfolio accompanied bycompanies with big scale and low B/M, it means high return portfolio instead lowerrisk.