Mutual Fund Managerial Working Experience, Career Concern, New Fund Opening and Fund Performance

Abstract

This thesis comprises three essays on mutual fund performance which provide new insights into different aspects of the mutual fund industry.

The first essay examines the relationship between the mutual fund manager’s past experience and mutual fund performance. The skills and knowledge acquired from the prior working experience may be transferred to the current working context, thereby influencing the current job performance (Schmidt et al., 1986). Using data on U.S. mutual fund managers’ work experience ranging from 1993 to 2012, we introduce a new method to evaluate mutual fund performance from the perspective of the manager’s lifetime working experience. Specifically, the method involves using the Principal Component Analysis to construct a Managerial Experience Index (MEI) based on 3 professional experience factors from the past career history of each manager: (i) investment objectives of the funds that s/he has managed (Zambrana and Zapatero, 2017), (ii) fund companies that s/he has worked for and (iii) industries of stocks in which s/he has invested (Kacperczyk et al., 2005). The MEI would increase along with the experience accumulation for each mutual fund manager. We group the sample based on the MEI into 5 quintiles from the lowest MEI score (most concentrated experience) to the highest MEI score (most diversified experience). The findings suggest that managers with more specialised experience outperform managers with more diversified experience. In addition, the “Specialist” managers tend to exhibit stock-picking ability while the “Generalist” managers tend to exhibit market-timing ability.

The second essay analyzes the performance patterns of new funds during the early stage after their creation, and provides potential explanations for their short-lived outperformance. Using a sample of incubation-free mutual fund data from 1996 to 2015, we address the questions of (i) whether new mutual funds outperform the market and (ii) if they do what may explain their superior performance. We find evidence of out performance for the new funds during their emerging period defined here as the first 6 months of their existence, both before and after fund expenses are taken into account. This outperformance, however, only lasts for a short term and disappears soon after the emerging period. This short-lived outperformance can be explained by the small size effect and IPO stock allocation, but is only weakly associated with managerial characteristics such as team managers and prior experience in equity fund management. Our analysis also provides evidence on a flow-performance relationship. The results suggest that IPO allocation is an effective strategy that enhances investment flows during the emerging period of a new fund. In addition, we find that funds created by team managers attract more flows than funds created by individual managers.

The third essay examines if fund managers would take into account turnover risk from a tournament when adjusting the risk of portfolios under their management, where the tournament is defined as the competition in a group with the purpose of being rewarded on their relative performance Conyon et al. (2001). In addition to exploring a statistical correlation between a manager’s discharge from a fund and the realized volatility of the fund that she had been managing, we use an instrumental variable (IV) approach to study whether one may infer causality from such a correlation. Using the instrumental variable (IV) measured as the peer flow pressure in the tournament following the “Rank-of-Ranks” approach in Kempf and Ruenzi (2008), we find that peer flow pressure is a highly statistically significant determinant of manager replacement. Further, the risk of replacement is significantly linked to the fund’s realized idiosyncratic volatility. The finding is robust to the use of an alternative instrumental variable (Segment Flow Rank), an alternative measure of realized risk (Carhart-adjusted Idiosyncratic Risk), and finite distributed lag specifications that incorporate one-period lags of explanatory factors.