This study delves into the reversal effects in the U.S. stock market using American stock data listed on the New York Stock Exchange, American Stock Exchange, and NASDAQ from 1970 to 2022. The aim is to answer two key questions: What characteristics make certain groups of stocks exhibit stronger reversal effects? And what market conditions contribute to stronger reversal effects?To begin with, the paper examines whether growth stocks exhibit stronger reversal effects compared to value stocks from the perspective of growth stocks. The study uses the price-to-earnings ratio (P/E ratio) to measure stock growth, with high P/E ratio stocks classified as growth stocks and low P/E ratio stocks classified as value stocks. The findings reveal: 1) The reversal effects of growth stocks are significantly stronger than those of value stocks; 2) After a substantial market decline in the previous year, the reversal effects of stocks are significantly stronger; 3) Across different market environments, the reversal effects of growth stocks are consistently stronger than those of value stocks, and growth stocks exhibit the most pronounced reversal effects in markets following significant declines. Furthermore, the paper explains why the reversal effects of growth stocks are stronger from three perspectives: market risk exposure, interest rate sensitivity, and profit volatility. The study discovers that the market BETA, duration, interest rate sensitivity, earnings per share volatility, and positive correlation with the Purchasing Managers' Index (PMI) for growth stocks are all significantly higher than those for value stocks. This helps explain why, during stock market rallies/interest rate declines/economic expansions, the rebound strength of growth stocks' prices/profits is higher than that of value stocks, leading to stronger reversal effects. Finally, the study finds that the phenomenon of "stronger reversal effects in growth stocks" also holds true in the A-share market, which serves as an emerging market.

Based on the common phenomenon of style rotation in domestic and foreign stock markets, this paper aims to study and answer which factors jointly drive style rotation and whether style rotation is predictable. Based on the dividend discount model, this paper selects variables that may explain style rotation from the three dimensions of capital cost, risk premium and performance growth. At the same time, this paper innovatively introduces the capital flow variables of institutional investors and northward funds to help explain and predict style rotation from the perspective of "smart money".First, based on the A-share market, this study uses the daily frequency value factor yield data from January 1, 2015 to December 31, 2020 to carry out temporal regression of the variables that may affect the value factor yield. It is found that the macroeconomic leading indicator and Wind A dynamic dividend yield can significantly affect the yield of value factor, and the impact is positive, that is, the rise of the macroeconomic leading indicator and the rise of the dynamic dividend yield of A shares both lead to the rise of value factor yield. In addition, based on daily frequency, this paper also found that value factor yield and northbound capital scale is significantly negatively correlated, but this relationship does not exist on monthly frequency. Secondly, this study further uses the daily frequency value factor yield data from January 1, 2015 to December 31, 2020 to carry out temporal regression of each explanatory variable of the previous day, trying to study whether these variables can predict the value factor yield. It is found that the leading macroeconomic indicator and Wind All-A dynamic dividend yield can positively predict the yield of value factor. Specifically, if the leading macroeconomic indicator rises in the previous trading day or the Wind A dynamic dividend yield rises in the previous day, on average, the value factor yield will rise in the next trading day. This finding is consistent with the results of synchronous temporal regression in the previous section. In addition, this paper does not find that the size of northbound funds can significantly predict the return rate of the value factor. Finally, this study uses variables that have significant predictive effect on the value factor rate of return to build a model. Based on this model, the out-of-sample value factor rate of return is predicted, so as to timing the value factor. The results show that the rate of return of value factor investment strategy based on model timing is twice as high as that of long-term holding value factor.