Mechanism Analysis And Prediction Of Fund Manager Investment Risk Behavior

In recent years, the size of open-ended funds is increasing each years, it has gradually become the most important part of China’s financial markets. As the main person responsible for the fund investment, the investment risk behavior of fund managers has an important influence on the stability of long-term performance of the fund. Mechanism analysis and prediction of fund managers’investment risk behaviors can provide reference for the decision-making of investors and regulators, at the same time, make a great significance for long-term positive effect of open-ended fund development.Most scholars both domestic and foreign, use the volatile of fond performance, namely the standard deviation of fund managers’investment risk. According to the research, the managers will make an adjustment on their investment risk by their own history relative rank. In this paper, based on this research, according to a large number of open-ended fund operation data and characteristics, the multiple regression method and neural network were used respectively to make a mechanism analysis and prediction on the manager’s investment risk behavior. The empirical test shows the method is effective and universality.At first, we make the mechanism analysis on the managers’investment risk behavior from career concerns, for self-interested reasons, managers invest and the starting point of decision making is more to consider their career to determine whether increase or decrease the investment risk, the managers who have the low previous rank face on more employment risk will select high risk strategy, but the managers who have the high previous rank has no employment risk, in order to maintain the reputation on the labor market,they will select low risk strategy.Next, for the panel data of China’s 2011 open-ended equity funds, we select the fund relative performance ranking and other factors to analyze the managers investment risk by using multiple linear regression method, through stepwise regression method to solve the regression coefficient.regression prediction model is established, using 2012 open-ended equity funds data is verified.Finally, in order to improve the prediction precision, we respectively use BP neural network and RBF neural network to predict managers risk behavior. Through the contrast analysis of the two methods shows that RBF neural network prediction model has significant improvement both on the accuracy and stability.