| Along with the rapid development of China’s capital market,funds are gradually coming into the public eye as an investment tool,and more and more investors are turning to them as one of their investment targets.Researchers have been profiling the performance of funds,trying to find the factors that best represent asset returns to explain various market anomalies.As more and more scholars are working in the field of asset pricing,the development of pricing models has become more advanced.The number of factors has increased from the initial single factor(CAPM model)to multiple factors(Fama-French three-factor model);the factor structure no longer consists of simple common factors that affect the returns of all assets,but also introduces group-specific factors that affect only some assets;and the sources of factors include not only those factors that can be observed,but also slowly shift to the study of potential factors.Inspired by the article by Ando and Bai(2015),this paper constructs a multi-factor pricing model that encompasses both observable and unobservable factors based on the characteristics of the Chinese capital market to study the structure of factors affecting fund returns in a new dimension.This paper selects monthly return data of Chinese open-end funds from December 2012 to December 2022 and clusters the factor data using the Macro PCA estimation method with the intention of building a better asset pricing model to fit the actual return movements of funds.The study confirms that there are potential cluster-specific factors affecting fund returns,as the intercept α of the pricing model with unobservable factors clustered is closer to 0 than that of the unclustered model,reflecting the neglect of unobservable factors in the traditional pricing model and the shortcomings in the accuracy of asset pricing.Further,this paper finds that the potential factors can be divided into three clusters,with three specific influencing factors within each cluster.In addition to finding the pricing model that best fits the fund performance,the truthfulness of fund performance is also an interesting research direction.Many investors base their investment strategies and asset allocation on the fund’s past returns.Yet a large body of research suggests that this is not a wise direction to refer to.In fact,not all of the good funds we observe have the ability to pick stocks,and may be influenced by good luck to show positive excess returns.And for investors who mistakenly choose good luck funds,they will suffer losses once the luck stops.Similarly,a bad fund performance does not necessarily mean that the fund manager does not have the management ability,but may just be unlucky.In order to identify the part of luck and skill in fund performance,this paper uses the FDR method proposed by Barras et al.(2010)to control for misjudgment in fund performance evaluation,and the Bootstrap program to simulate the data distribution so as to remove the effect brought by luck on fund performance.It is found that only a small fraction of funds are truly skilled in our fund market,and most of them do not have the ability to beat the market.In addition,we find that both skilled and unskilled funds are scattered in the left and right tails of the performance distribution. |
