Performance persistency of retirement benefit fund managers in Kenya

Abstract

This study sought to study performance persistence, over time, during a study period of between 2006 through 2011 in the performance of fund managers’ responsible for investments of Kenya Retirement Benefit Funds on behalf of the Trustees and the Scheme members. The fund managers make investments decisions and invest the schemes funds in an array of investment vehicles ranging from property, government securities, quoted shares, unquoted shares, corporate bonds, offshore investments, guaranteed funds among many others and declare income rate at the close of specific period, usually per annum; this income rate being the proportion of the accumulated incomes over the fund value under a particular fund manager. The study adopted a descriptive a Research Design. It studied 10 fund managers with a sample of 265 segregated retirement benefit schemes, data secondary in nature, was drawn from the RBA systems Analysis involved averaging abnormal returns across the funds under a fund manager which was both equally weighted by fund size. Asset pricing models, as the Fama- French three-factor model and the single factor CAPM, were used in iterative regression models, and the factor loadings were estimated over the whole study period. Performance of a fund manager was computed by averaging the abnormal returns of the pension funds managed by a fund manager for each year. Performance ranked portfolio tests were used to sort fund manager each period into 2 portfolios with equal numbers of fund managers based on past performance with an annual ranking period where ranking was on basis of the average return on the funds under management in the ranking period. Top portfolio consisted fund managers with the highest average abnormal returns in the ranking period while bottom portfolio with those fund managers with the lowest average abnormal returns. After which, equally weighted average portfolio abnormal return of the top and bottom portfolios over a subsequent evaluation period were calculated which were further averaged over all evaluation periods and a difference was computed which indicated persistency as it was not centered at zero.