MARKET TIMING ABILITY OF FUND MANAGERS : AN EMPIRICAL STUDY OF THE SINGAPORE FUND MANAGEMENT INDUSTRY

Abstract

The performance of fund managers determines the manner in which investors allocate their wealth among various professionally managed portfolios and directly or indirectly measure affects the compensation of the managers. The ability to accurately investment performance is therefore an important topic in finance. A useful evaluative framework must be able to draw the distinction between the ability to time the market and the ability to forecast the return on the individual assets. An ability to differentiate between these two sources of superior performance will allow a managers greater measure of the services provided by the fund managers. This study aims to provide empirical evidence on the timing abilities of Singapore fund managers, that is, it seeks to establish whether such managers have consistently varied the risk levels of their portfolios through recomposition, in anticipation of broad movements in the returns of both risky and riskfree assets. Based on the theoretical framework of market timing developed by Merton [1981], studies by Henriksson [1984) and Chang and Lewellen [1984] have found little evidence on the part of fund managers in anticipating stock market movements. Our conclusion on Singapore fund managers is more ambiguous. Tests using the non-parametric approach indicate that Singapore fund managers have positive timing ability. However, when the parametric approach is used, Singapore fund managers do not demonstrate significant timing ability over the sample period. In fact, a few funds actually demonstrate significantly negative timing ability. These funds recompose in the direction of lower portfolio systematic risk in rising markets and higher systematic risk in falling markets. The results for the Singapore fund managers are obtained after taking into consideration the presence of heteroscedasticity in the stock returns data. Breen, Jagannathan and Ofer [1986] produced evidence using simulation techniques that wrong statistical inferences can arise if heteroscedasticity is ignored. The study is also extended to consider the information market economic agents use in forming their expectations of stock return movements. In this paper, a predictive model for the nominal market excess return using the lagged riskfree rate is evaluated. A negative correlation coefficient is found between the nominal market excess return and the lagged riskfree rate for the period January 1980 to December 1987. The forecasting ability of the model is found to be positive but not highly significant. Another equally interesting strong negative correlation result of this study is the between selectivity and the measures of timing. The evidence is consistent with that of Kon [1983], Henriksson [ 1984] and Chang and Lewellen [1984]. An explanation to this anomaly Jagannathan and is given by Korajczyk [1986]. They demonstrate is possible to create that it artificial market forecasting or timing ability even when no market forecasting or security-specific forecasting done. This can happen when is being a manager investing in option-like securities evaluated using the parametric approach. The negative correlation between selectivity means that and timing investing timing in portfolios ability can be created (or suppressed) by with extremely low (or high) selectivity measures. The implication the results from to this study is that the parametric test may be confounded by the juxtaposition of selectivity and timing measures single specification, in a for the conflicting thus providing a plausible explanation inferences between the parametric and non-parametric procedures. In conclusion, among fund managers the evidence of market timing ability in Singapore negative relation between market is not conclusive. A timing and selectivity is found for the sample period. Similarly, riskfree rate the lagged nominal is found to be negatively correlated nominal market excess to the return.