Predicting Tunisian mutual fund performance using dynamic panel data model

Abstract

Purpose

The purpose of this paper is to examine the ability of well known fund characteristics such as the recent past performance, fund size, management fees, fund age, net asset value and fund growth so as to explain Tunisian equity mutual fund performance.

Design/methodology/approach

The sample was split according to investment objectives, and the advanced dynamic panel data approach was used over the period 1999‐2006.

Findings

The authors find that past performance and fund size have a positive and significant influence on future performance for all fund categories, irrespective of what performance measure was used. This may indicate the existence of scale economies in the Tunisian equity mutual fund industry. The author also find that the other fund characteristics play an important role in explaining performance, but their impact varies among the fund categories. In all, regression results support the dynamic links between fund characteristics and future performance.

Research limitations/implications

The findings do not take into account the behaviour of fund managers and their ability to extend the investment opportunities set. It seems that there are more complex factors related to the strategic behaviour of the manager and driving differences in performance across funds than previous studies have indicated.

Practical implications

The authors confirm the empirical evidence that historical performance contains some information about future performance and such information may be important to mutual fund investors. It was also found that fund size is positively related to future performance of small fund category as well as of large fund category. This may indicate the existence of scale economies in the Tunisian equity mutual fund industry. In addition, the influence of the other control variables varies among the fund categories, but often is the same as in earlier studies.

Social implications

The paper provides information to foreign investors for investing in Tunisian capital market.

Originality/value

In this regard, the study of literature revealed that the explanation of performance, based on quantitative factors, is often limited to a static approach that involves making estimates resting on multiple regression, regression in cross section and principal component analysis for short periods. However, several empirical studies highlight the impact of past performance on future performance. It seemed essential to enrich the analysis by using a dynamic approach.

Citation

Ben Belgacem, S. and Hellara, S. (2011), "Predicting Tunisian mutual fund performance using dynamic panel data model", Journal of Risk Finance, Vol. 12 No. 3, pp. 208-225. https://doi.org/10.1108/15265941111136950