The Optimization and Empirical Study of Stock Excess Return

Abstract

In order to explore the influencing factors of stock excess return and provide a particular reference for predicting the development of stocks in the country, I comprehensively refer to relevant economic theories and academic research and use the 2018 and 2019 CSI 300 index data provided by the wind to establish an econometric model. Moreover, carry out inspection and evaluation. I choose the logarithm (R) of the stock excess return as the explained variable, and the risk coefficient, market capitalization, book-to-market value ratio, turnover rate, etc., as the explanatory variables for regression analysis. According to the support of economic theory and the observation of the scatter plot trend, I removed the two explanatory variables of price-earnings ratio and net profit. In order to reduce the influence of outliers on model parameter estimation, I use the SPSS case diagnosis function to remove outliers. I performed a least-squares estimation of the initial model and performed t-tests and F-tests for the parameters and the overall equation for the remaining data. The variance inflation factor shows that the model has a multicollinearity problem, which I remedy by removing the year*ln_ME term. The model did not pass White's heteroskedasticity test. I used Eviews to perform White's heteroscedasticity correction on the model to correct the model standard error and reduce the degree of heteroskedasticity. After remediation, I used the RESET test on the model again, and although the F value increased slightly, it was still in the receptive field, so the model remediation did not make the model setting severe problems. The Jacques Bella test is shown that the model residuals do not completely obey the normal distribution, but due to the relatively large sample size, the model residuals can be approximated by a normal distribution according to the central limit theorem. In addition, I also obtain numerical prediction results through point prediction, and the scatter plot shows a high degree of interval overlap. Finally, I summarize the form presented by the final model and return to the practical significance, which proposes a reference standard for evaluating excess return from the market environment and company size, and further affirm the economic significance of excess return as a diversified evaluation of stock quality.