Optimization Of A Multi-factor Stock Selection Model Based On Residual Coverage Model

In a multi-factor stock selection model,investors use the level of analyst coverage to listed companies as the analyst coverage factor,which is constructed on the basis that analysts tend to focus on companies with good prospects,which tend to have better fundamentals and higher expected returns.The analyst coverage factor is constructed by using the total frequency of analysts’ coverage to listed companies.However,simply constructing the factor based on this principle will lead to biased stock selection results because analysts’ selection of companies for research will be influenced by companylevel,industry-level and market-level factors,resulting in the performance of companies with high analyst coverage not necessarily being good.In this paper,we analyze these influencing factors,extract the current performance information of listed companies and analysts’ expectation of their future performance information from analysts’ coverage level by establishing residual coverage model,construct residual coverage factor,and then use residual coverage factor instead of coverage factor to capture stock selection information from analysts’ coverage level and establish multifactor stock selection model to reduce stock selection bias This paper improves the stock selection ability of the model.In this paper,the residual coverage factor is constructed by establishing the residual coverage model of analysts with CSI 300 constituent stocks as the research sample.The mean IC value of the factor is 0.036 and the IC＿IR value is 0.54,and the results prove that the factor has the ability to obtain alpha returns stably.Based on the test results,a single-factor stock selection backtest was conducted on this factor and compared with the backtest results of the analyst coverage factor.The backtest results show that the residual coverage factor has an annualized return of 21.12%,a Sharpe ratio of 3.10 and an information ratio of 1.08,proving that the residual coverage factor possesses better stock selection and reduces stock selection bias compared with the analyst coverage factor.In this paper,a multi-factor stock selection model including the residual coverage factor is further developed,and the model is back-tested in a realistic way and out-of-period.The results show that the portfolio constructed according to the model outperformed the market throughout the backtest period,with a cumulative return of 181.27%(22.98%annualized),a Sharpe ratio of 3.50,and an information ratio of 1.08.The return of the portfolio rose more when the market index rose and fell less when the market index fell.The out-of-period backtesting results demonstrate the robustness of the model’s stock selection effect,indicating that the multi-factor stock selection model based on analysts’ residual coverages achieves excellent stock selection results in the current Ashare market.