Asset Pricing Problems Research Based On The Behavior Analysis Of Investor Overconfidence

In this paper, under the investor overconfidence, we do our best to research how to build a better asset pricing model and carry out its empirical test.First, through the comprehensive reviews of many literatures about investor overconfidence, we find the effect of overconfidence on asset pricing be influenced by market states, signal types and other factors. In the bullish, investors are easily overconfident. This behavior always makes them overreaction to the good news and under-reaction to the bad, which would lead asset price to be overestimated. Nevertheless, investors are fond to under-confidence in bearish that makes them overreaction to the bad news and under-reaction to the good, which would cause asset price underestimated. Moreover, the integrated relationship among these factors can also be expressed by contingency tables and figure.Secondly, according to the results in Daniel et al. (2001), we develop a new dynamic multi-factors asset pricing model under investor overconfidence. This model not only embodies the asymmetric reaction of the overconfidence on different signals; but also reflects the noise trade components in the beta. What’s more, it could be developed to inter-temporal states. End, this dynamic model also implies, as other researcher arguments, that overconfident investors would earn more expected profits and utilities than the rational traders because bear more risks created by themselves.However, we derive measurable variable named relative amplitude through data mining from the highest and lowest prices. With this variable, the dynamic multi-factor asset pricing model could be simplified as two factors asset pricing model, which would be named the capital assets model of relative amplitude (ACAPM). According to return data of individual stock from the Shanghai and Shenzhen A-share market, the empirical test results of ACAPM show that addition of relative amplitude not only improve the fitness of the pricing model to the empirical data, but also to realize effective measurement on the abnormal return of individual stocks. At the same time, the empirical results also show that Chinese investors are probably overconfident in the bullish, and this overconfident is relevant to the specific industry in some time.In short, the existence of relative amplitude only shows that noise trading could not be deleted by the reverse arbitrage of rational investors, which let the effect of noise-trade left and becomes a part of earnings, and just being the source of abnormal returns. The coefficient is exact residual ratio of noise trade.