Research On Herd Behavior Of Mutual Funds In China

The development of institutional investors is the benchmark for the maturity of the securities market.As the scale of China's capital market continues to rise,the market position of institutional investors has become increasingly prominent.As the mainstay of institutional investors,mutual funds have drawn more and more attention from researchers and regulators and have more and more significant influence on the development of the securities market.However,as the financial system is gradually maturing,due to the problems of asymmetric information and imperfect supervision,a series of "financial anomalies" represented by herd behavior gradually emerge.These anomalies will directly lead to abnormal fluctuations in the stock market price,impact on market stability and even cause financial crisis.The purpose of this paper is to measure the existence of herd behavior when Chinese mutual funds invest in individual stocks and sector,try to use Social Network Analysis to identify the "bellwether" of mutual funds,and provide effective practical significance for the operation and supervision of securities markets.Using the quarterly data of Chinese open-ended mutual funds covering the stocks of 18 industries,we empirically test the convergence behavior of the funds' investment,which is called herding behavior.With calculating Herding Index,we find obvious existence of herding behavior in the investment of mutual funds at individual and industrial level.Herding Index of current period is 14.7%,and the Industry Herding Index is about 9.4 percent,and particularly the Buying Index is much higher than the Selling Index.Accounting to the cross-period investment,there is a significant continuity between two adjacent periods and the coefficient of correlation is 0.218,the industry coefficient of correlation is 0.112.Divide the cross-period investment in two parts: implicit-transactions and follow–transactions.The former represents the transactions following the funds' own trades and the latter represents the transactions following other ones.We find that the follow-transaction accounts for 66.7% or more of the cross-period investment.A group test classified with mutual funds' activity shows that the more active the fund is,the less significant the herding behavior will be.Analyzing herding behavior distribution among industries,we find great divergences both in current investments and in cross-period ones.With step-wise regression model,we demonstrate that stock price volatility,diversity in stock performance and policies may be the important signals that cause those divergences.Finally,Social Network Analysis is used to construct a weighted fund network model to identify the "bellwether" of mutual funds in herding,so as to provide reference value for explaining investors' decision-making behavior.