Behavior Index Factor Strategy Based On Investors' Speculating And Trading Behavior

Previous works of researchers have shown that it is possible to arbitrage over stocks when the market is over-speculated or driven by irrational bubbles.Also many works have proved the ineffectiveness of Chinese security market where value-investing could subsequently be fading out.While quantitative investing has been booming over 30 years overseas,it's necessary to apply these advanced methods to our market to develop more scientific and controllable ways in investing.Taking the majority of noise traders in the young Chinese stock market into consideration,we choose this speculation-based multi-factor model to generate a profitable strategy.There are many findings about multi-factor based on corporate characteristics,but in this paper,we focus on the factors that related more to the speculated and over-traded stocks.Our purpose is to develop a multi-factor model that could possibly measures the investors'over-speculating influences on stocks as well as gains effective profits when we use this strategy.We try to find factors that reflect the over-speculated stocks as well as stocks with irrational bubbles after referring to Asset Pricing Models and Behavior Finance Theories.The first factor,IVR,comes from Fama-French multi-factor model which proved stock returns had shared variation with 3 factor including an overall market factor and factors related to firm size and book-to-market equity.Many researchers have proved its consistence in Chinese stock market using Shanghai&Shenzhen market data.While the model could generally explain the stock returns variation,we think that the unfitting part of the model basing on monthly regression which equals to 1 minus R-squared probably represents the over-speculated degree of a single stock.And the fact is that the over-speculated stocks always lose in the subsequent days and we could consequently making arbitrage strategy out of this.The second factor widely proved is turnover rate but adjusted with market value in which way we get the pure factor over the strong"small-size-effect".The third factor is rather unique because we refer a statistical way in which we try to get the over-speculated stocks among their similar terms.What really takes much work in this part is that we use the correlations over 250 previous returns of two to judge if they are two of the same kind,and we collect 10 similar stocks to every stock so that we could possibly calculate it's reference price,of course,in separate months.Apart from building the three factors above,we also proved their efficiency in forecasting stock returns and examined their low-correlated relations.Most importantly,we find an enhanced effect in building portfolio that could had have always profited in last 9 years when we combined them together by scoring method.