No. 1040

Title:

Dynamic Modeling and Estimation for Analyzing Investment Style of Japanese Equity Mutual Funds Considering individual stock selection ability

Author(s):

Yamashita, Takashi (Department of Statistical Science The graduate University for Advanced Studies 4-6-7 Minami-Azabu, Minato-ku, Tokyo 106-8569, Japan)

Key words:

Dynamic modeling; Mutual fund; Investment style; Kalman filter

Abstract:

    The purpose of this paper is to report my work building dynamic models for estimating timevaring investment style drift and identification using actual market data. For the purpose of this paper, three steps of research were taken. First, I verified method for estimating investment style using the W.\protect \unhbox \voidb@x \penalty \@M \ {}Sharpe's method and exhibited some problems with it. Next, I modeled some static and dynamic systems for estimating investment style, one of the dynamic one is considering the factor of individual stock selection ability of the fund.Then I identified models by actual mutual fund data using the innovation approach with the Kalman filter in the final step. We can estimate timevarying style changes of an equity fund manager by the dynamic modeling in this work. It was found that the simple Gaussian noise random walk models showed better fitting performance than static models. I conclude that the dynamical systems are essential for style analysis. It is recommended that the modeling of dynamic system corresponding to style jump and the identification method be improved.