A Research On The Timing Ability Of Chinese Full-blown Active Open Funds

Mutual fund is gradually becoming a much more important institutional investor in security market. How to reasonably evaluate mutual fund performance attracts considerable interest, both from academics and practitioners. As an important part of the evaluation system, the timing ability of fund manager related to the evaluation of management level among the mutual funds. Besides it can tell the investors the composition of performance, which thus provides an important investment basis.This paper took empirical research methods as the main method and comparative analysis as subsidiary to search the timing ability of China open funds from two aspects, hoping to obtain a comprehensive and unbiased reflection of fund managers'timing ability and identify the composition of the excellent performance. Firstly we discuss the timing ability of mutual funds from three facets:over the whole evaluation period, in terms and in the up and down markets. Secondly we research how the chosen of models, benchmarks and data frequency impacted the empirical results and the validity of the tests.This paper selected all the flexible funds'daily yields data from July 2000 to the end of 2008 as a sample to make a general research of their timing abilities. In this paper, we used traditional H-M model based on single factor CAPM benchmark as a control, then we applied its extending four factors benchmark, which we called Carhart4 factors benchmark here to evaluate the funds performance. The empirical results showed that the extending four factors benchmark can interpret the funds performance more comprehensively and unbiasedly. Moreover, funds managers obviously had their own investment styles, i.e. so-called style investors. These funds managers preferred investing in big-cap growth stocks with a strong momentum deliver during the bull markets. On the contrary, they preferred investing in big-cap value stocks during the bear markets. As a whole, the funds managers demonstrated significant stock selection ability and little timing ability throughout the evaluation period.In addition we adopted H-M, T-M and GⅡtiming model based on three benchmarks, i.e. single factor CAPM benchmark, three factors Fama-French benchmark and four factors Carhart benchmark with both the samples' monthly yields data and daily yields data to make a comparative synthetic research. The correlation tests showed that model selection made no difference to the empirical results, however, the data frequencies and benchmark selection played an important role in interpreting the performance.