Bidirectional Prediction Research Of Fund Manager Management Ability And Fund Performance

Since 2022,the number of public fund products in China has exceeded 10,000,and the number of public fund managers currently exceeds 3,000.The phenomenon of multiple management is becoming more and more common,which to some extent speeds up the pace of fund research to fund manager ability research.Individual investors are often confused about whether to follow the fund or the manager.To get the answer to this question,it is necessary to first clarify the relationship between the management ability of fund managers and fund performance,and whether there is a two-way influence between the two.This problem can be divided into two oneway influence subjects for research: first,whether the management ability of fund managers is affected by the performance of funds they have managed in the past;Second,the extent to which the management ability of the fund manager influences the fund performance.Whether the management ability of a fund manager is affected by the performance of the funds it has managed in the past can be abstracted to explore whether the management ability index of a fund manager can be predicted by its historical performance.The weighted yield index of 88 fund managers from 2014 to 2021 is selected as the data set,and the LSTM model is used to forecast the time series data.The model achieved the prediction effect of correctly fitting the change direction of time trend,and the MAPE value was 1.7.By extracting the data information on the time node according to the decile,88 fund managers were sorted and reorganized into 4 groups with different management abilities.The accuracy of predicted grouping was as high as 82%,which supported that the management ability of fund managers was affected by the performance of funds they had managed in the history.That is,the management ability of fund managers can be reflected by fund performance.This paper studies the influence degree of fund manager’s management ability on fund performance,selects random forest method to model and forecast fund performance,and turns the problem into the contribution degree of fund manager’s management ability index to fund performance prediction.66 fund indicators and2 fund manager indicators are used to build a feature system,and a total of 124 funds are managed by 88 fund managers who meet the fund life from 2014 to 2021 and are in the stage of previous proven management talent.By comparing the modeling effect with and without the characteristic data of related indicators of fund managers,the model learning effect after adding the index of fund managers’ management ability is slightly better than that of the model without adding the index of fund managers’ management ability,but it is not very significant.The evaluation of the importance of variables found that some funds have a slightly lower weight of fund manager indicators,maintaining only about a third of the position.Therefore,it can be inferred that the management ability of fund managers has a certain impact on the performance of their managed funds,and some of the insignificance of the conclusions may be due to the defects in feature selection.To sum up,it can be believed that the management ability of fund managers and fund performance have a two-way effect.The performance of funds managed in history can reflect the management ability of fund managers,and the management ability of fund managers will also have a certain impact on the performance of funds managed by them.