The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? PDF

The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? Prachi Deuskar Department of Finance, College of Business, University of Illinois at Urbana-Champaign Joshua M. Pollet...

The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? Prachi Deuskar Department of Finance, College of Business, University of Illinois at Urbana-Champaign Joshua M. Pollet Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 Department of Finance, Eli Broad College of Business, Michigan State University Z. Jay Wang Department of Finance, College of Business, University of Illinois at Urbana-Champaign Lu Zheng Department of Finance, Paul Merage School of Business, University of California at Irvine and CAFR Does the mutual fund industry lose its best managers to hedge funds? We find that mutual funds are able to retain managers with good performance in the face of competition from a growing hedge fund industry. On the other hand, poor performers are more likely to leave the mutual fund industry. A small fraction of these poor performers find jobs with smaller and younger hedge fund companies, especially when the hedge fund industry is growing rapidly. Analogously, a small fraction of the better-performing mutual fund managers are retained by allowing them to manage a hedge fund side-by-side. (JEL G23, G29) Retention and promotion decisions are important in any employment relation- ship. Retaining skilled managers while firing incompetent ones is crucial to maintaining productivity. At the same time, retention and promotion policies We thank Heitor Almeida, Jonathan Berk, Sreedhar Bharath, Jeff Brown, John Campbell, Murillo Campello, Tim Johnson, Chris Malloy, George Pennacchi, Scott Weisbenner, Matt Spiegel (the editor), an anonymous ref- eree, and seminar participants at University of Illinois at Urbana-Champaign, University of Illinois at Chicago, University of California-Riverside, Financial Intermediation Research Society Conference, China International Conference in Finance, American Finance Association meetings, and UC Davis Symposium on Financial In- stitutions and Intermediaries for many helpful comments and suggestions. We would also like to thank Marek Jochec, Quoc Nguyen, and Liz Risik for excellent research support. Deuskar and Wang acknowledge the fi- nancial support of the Arnold O. Beckman Research Award by the University of Illinois at Urbana-Champaign Campus Research Board. Send correspondence to Z. Jay Wang, 340 Wohlers Hall, 1206 S. Sixth St., Champaign, IL 61820; telephone: (217) 265 6598. E-mail [email protected]. c The Author 2011. Published by Oxford University Press on behalf of The Society for Financial Studies. All rights reserved. For Permissions, please e-mail: [email protected]. doi:10.1093/rfs/hhr057 Advance Access publication June 21, 2011  The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? should take into account the dynamics of competition from the external labor market. Responding appropriately to the changing outside options of employ- ees is critical to the retention of skilled workers. During the last decade there has been a much greater expansion by hedge funds than by mutual funds.1 This rapid increase in the size of the hedge fund industry could be attributed to the relative absence of regulations regarding compensation contracts and trading strategies. Unlike mutual funds, hedge funds are able to charge an incentive fee that is a large proportion of the capital gain above a pre-specified hurdle rate because they are free from regulatory restrictions on the investment advisory contracts. This ability to use incentive Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 contracts with “option-like” features between the fund and the management company could allow hedge funds to lure talented managers away from mutual funds. However, mutual funds could respond to such competition by providing other opportunities to reward successful managers and overcome any potential handicap due to regulation. We analyze the ability of the mutual fund industry to retain managers with superior historical performance and jettison managers with poor performance in the midst of the rapid growth of hedge funds. Our sample consists of 287 mutual fund managers who joined hedge funds during the period from 1993 to 2006. Of these managers, 157 side-by-side managers retained their jobs in the mutual fund industry while simultaneously managing both mutual funds and hedge funds. The remaining 130 managers, complete switchers, severed all ties with the mutual fund industry to join hedge funds. We find that superior past performance as a mutual fund manager increases the likelihood of a side- by-side arrangement. Poor performers tend to completely exit the mutual fund industry, and some of them find jobs with hedge funds. Why would the hedge fund industry hire poorly performing mutual fund managers? One possibility is that these managers are better suited to hedge funds than mutual funds. However, our findings suggest that complete switch- ers continue to perform poorly relative to other managers in hedge funds. In addition, the bulk of the managers who left the mutual fund industry to join hedge funds did so during the boom period of the hedge fund industry (early 2000s).2 Given the scarce supply of money managers, some hedge funds might need to lower their hiring standards during periods of rapid hedge fund growth. Indeed, we find that poorly performing mutual fund managers tend to find jobs in hedge funds only when the hedge fund industry is growing extensively. Also, these complete switchers join significantly smaller and younger hedge 1 Between 1997 and 2007, assets under management in mutual funds grew by 167%, while those in hedge funds grew by 300%. Source: 2009 Investment Company Fact Book for statistics on mutual funds and www.hedgefundfacts.org, based on data from Hedge Fund Research Incorporated for statistics on hedge funds. 2 The number of hedge funds increased from around 2,400 to 3,900 from 1995 to 2000. The number soared to about 8,700 in 2005. Source: www.hedgefundfacts.org, based on data from Hedge Fund Research Incorporated. 3009 The Review of Financial Studies / v 24 n 9 2011 fund management companies. It is possible that such hedge funds are unable to attract talented managers because they do not have sufficiently established reputations. We also find some evidence that successful mutual fund managers are more likely to begin a side-by-side arrangement when the hedge fund industry is growing rapidly. This is consistent with the conjecture that mutual fund com- panies respond to the improved outside options of their better managers by offering a side-by-side arrangement for retention purposes. This strategy par- tially alleviates the compensation constraints imposed on the mutual funds by the regulatory environment. Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 Even though our sample of switching managers is a small fraction of the universe of mutual fund managers and hedge fund managers, we are still able to address an important question regarding the competition for talented man- agers. Due to a supposed regulatory advantage, the hedge fund industry could potentially attract a large fraction of mutual fund managers. Thus, the fact that so few managers completely switch from mutual funds to hedge funds pro- vides preliminary evidence that hedge funds do not acquire talented managers directly from the mutual fund industry. In addition, of thousands of hedge fund managers, these managers are the only ones for whom prior performance data are available. Thus, for this group we are able to relate historical performance to the selection decision of hedge fund companies. Our article brings together two strands of research on managerial turnover. The first strand investigates internal signals affecting the retention and promo- tion of managers (e.g., Weisbach 1988; Fee, Hadlock, and Pierce 2006; Lehn and Zhao 2006; Cichello, Fee, Hadlock, and Sonti 2009). The asset manage- ment profession provides a unique context for labor market research with a well-tracked investment performance for individual fund managers (e.g., see Khorana 1996; Chevalier and Ellison 1999). Consistent with these studies, we find that better-performing managers are retained and promoted while poorly performing ones are fired. The second strand focuses on the impact of in- dustry/market conditions and the scarce supply of skilled labor on manage- rial turnover (e.g., Parrino 1997; Khanna, Noe, and Sonti 2008; Eisfeldt and Kuhnen 2011). We relate the turnover of money managers to the career op- portunities in the money management industry in an effort to understand the turnover of mutual fund managers during a special time period when the land- scape of the asset management industry is undergoing an extreme makeover due to the rapid growth of hedge funds. Our study finds that both good and bad managers get a better deal when the hedge fund industry is growing. There is a growing body of research on the impact of a surging hedge fund sector on various aspects of traditional asset management industry. Agarwal, Boyson, and Naik (2009) study the performance of “hedged mutual funds,” whereas Cici, Gibson, and Moussawi (2010) and Nohel, Wang, and Zheng (2010) investigate the potential conflicts of interest arising from the side-by- side management. Nohel, Wang, and Zheng (2010) find no conflicts of interest 3010  The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? because mutual funds managed by side-by-side managers consistently outper- form peer mutual funds after the manager enters the side-by-side arrangement. We focus primarily on the ex ante characteristics of mutual fund managers that enter the hedge fund industry either as complete switchers or as side-by-side managers. This approach allows us to analyze the competition for manage- rial talent across the two industries. Chen, Chen, and Cyree (2009) examine a smaller sample of mutual fund managers moving to hedge funds to investigate whether mutual fund managers can improve their own performance by manag- ing a hedge fund. They also find that complete switchers are poor performers relative to side-by-side managers. Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 Kostovetsky (2011) shows that, coinciding with the rapid growth period of hedge fund industry, the mutual fund industry experienced a widening per- formance gap between young and old managers. The study interprets these results as evidence of an implicit and explicit “brain drain” from the mutual fund industry caused by the superior ability of hedge funds to attract younger managers from mutual funds. While the paper raises a topic of great inter- est, the evidence is at best indirect because, unlike our study, there is no in- formation regarding actual career decisions. Our findings, that mutual funds offer the side-by-side arrangement to managers with better performance and sever their ties with managers performing poorly, suggest that mutual funds do not lose their existing talent to hedge funds. We acknowledge that our empiri- cal framework only examines departures of existing mutual fund managers to hedge funds (explicit brain drain) rather than the impact of competition from hedge funds on the quality or number of arrivals to mutual funds (implicit brain drain). However, if mutual funds are able to compete for existing talent, it is difficult to think of a mechanism that prevents mutual funds from competing for new arrivals as well. The rest of the article is organized as follows. The next section describes the data. Section 2 investigates what characteristics explain manager movement to hedge funds. In Section 3, we examine the performance of the switchers on the hedge fund side. Section 4 analyzes why some hedge funds hire poorly performing mutual fund managers. The final section concludes. 1. Data We construct the sample of switching managers by combining the Lipper TASS Hedge Fund database (TASS) and the Hedge Fund Research database (HFR) with the CRSP mutual fund database. The CRSP mutual fund database pro- vides information on fund complex, monthly total net assets (TNA), monthly returns, names and tenure of portfolio managers, and other characteristics such as expense ratio and turnover. The TASS and HFR databases track informa- tion such as monthly net asset value, fund inception date, investment objec- tives, and names of portfolio managers for the majority of the hedge fund population. 3011 The Review of Financial Studies / v 24 n 9 2011 Specifically, we compare mutual fund manager names with hedge fund man- ager names. For each manager name that appears in both mutual and hedge fund databases, we conduct an extensive cross-check on the employment his- tory with various sources (e.g., Morningstar, notes file in the hedge fund databases, and Internet searching) to make sure that the two names indeed refer to the same manager. Nohel, Wang, and Zheng (2010) provide details on the matching procedure. We restrict our attention to the set of managers that began as mutual fund managers and later joined the hedge fund industry. If there is an overlap between the tenure of the manager at mutual funds and at hedge funds, then we classify the manager as a “side-by-side manager,” i.e., Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 the manager simultaneously managing at least one mutual fund and at least one hedge fund for a certain period of time. If there is no overlap between the two tenure periods, we then classify the manager as a “complete switcher.” Fi- nally, we identify the mutual funds and the hedge funds the manager managed, either on her own or as part of a team. A limitation of our approach is that a comprehensive dataset does not exist for hedge funds. TASS and HFR each cover roughly 35%–40% of the universe of hedge funds, with relatively little overlap. Therefore, our sample has 70%– 80% of hedge funds but we acknowledge that we are not capturing the universe of switching managers that moved from the mutual fund industry to the hedge fund world. However, with comprehensive coverage of mutual funds, this lack of complete coverage of hedge funds only introduces noise and biases the sta- tistical tests against finding significant results. Using the procedure outlined above, we identify a total of 287 managers that switched from the mutual fund industry to the hedge fund world: 157 side-by- side managers and 130 complete switchers. Table 1 shows classification of the switching managers based on the styles of mutual funds they manage prior to the switch and the categories of hedge funds they join. It can be seen that the majority of the managers come from equity mutual funds and also join hedge funds with equity-driven strategies. This picture is reinforced by the fact that switching managers also have a larger fraction invested in common equity in their mutual fund portfolio (81% for a median mutual fund manager as opposed to around 90% for a median switching manager). Both groups of switchers are very small fractions of the 10,097 uniquely identified mutual fund managers and 9,616 uniquely identified hedge fund managers. This is an important finding in itself. However, it may be the case that even though the number is small, the most talented mutual fund managers move to hedge funds. We investigate this possibility in Section 2. To capture mutual fund performance, we calculate either style-adjusted av- erage return, style-adjusted MPPM, or 4-factor alpha, using returns before expenses. Each month style-adjusted return is calculated as the return of a mutual fund minus the average return of all the mutual funds with the same style. MPPM is the manipulation proof-performance measure suggested in Goetzmann, Ingersoll, Spiegel, and Welch (2007). For mutual fund i, for a 3012 The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? Table 1 Classification of switchers Panel A: Classification of Switchers by Mutual Fund Styles Side-by-Side Complete Total Style Managers Switchers Domestic equity 80 43 123 International equity 29 37 66 Domestic fixed income 28 18 46 International fixed income 4 6 10 Money market 4 1 5 Other 9 6 15 Unclassified 3 19 22 Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 Total 157 130 287 Panel B: Classification of Switchers by Hedge Fund Styles Side-by-side Complete Total Style Managers Switchers Convertible 3 3 6 Dedicated short selling 0 2 2 Emerging markets 8 6 14 Equity market neutral 11 16 27 Event driven 4 7 11 Fixed income 15 10 25 Fund of funds 13 5 18 Global macro 4 5 9 Long/short equity 90 68 158 Managed futures 4 3 7 Market timing 1 0 1 Multi-strategy 4 5 9 Total 157 130 287 This table provides a stylistic pattern of mutual fund managers joining the hedge fund industry. If a manager manages more than one fund, we choose the style of the largest fund. time period ending at t, it is calculated over prior T months as   X t  1−ρ 1 1 1 + r i,k M P P Mi,t = ln  . (1) 1−ρ T 1 + r f,k k=t−T +1 For month k, ri,k is the return for a mutual fund i and r f,k is the risk-free rate. The measure looks like the average of a power utility function with relative risk-aversion coefficient ρ. The choice of ρ depends on the benchmark port- folio against which the mutual fund is evaluated. As reported in Goetzmann, Ingersoll, Spiegel, and Welch (2007), using CRSP value-weighted market re- turn as the benchmark gives a value for ρ between 2 and 4. For our calcula- tions, we use ρ equal to 2, 3, or 4. Style-adjusted MPPM is the MPPM for a fund minus the average MPPM of all the funds with the same style divided by standard deviation of MPPM across those funds. We use the Wiesenberger Fund Type Code before 1993 and the Standard & Poor’s Detailed Objective Code from 1993 onward to determine fund style. The 4-factor alphas are based 3013  The Review of Financial Studies / v 24 n 9 2011 on market, size, value, and momentum factors (see Carhart 1997). We obtain the data for the factors and the risk-free rate from Kenneth French’s website. We use monthly data for three or five years to measure the performance. 2. What Explains Entry of Mutual Fund Managers into Hedge Funds? In this section, we analyze the entry of mutual fund managers into the hedge fund industry as a function of their past performance, trading behavior, ex- perience, and assets under management among other characteristics. We use the panel data of mutual fund managers described before. We consider three possible career changes: completely dropping out of the money management Downloaded from http://rfs.oxfordjournals.org/ at Michigan State University on November 2, 2011 industry, completely switching to hedge funds, and a side-by-side arrangement. We use multinomial logistic regression to jointly model the probability of each of these career moves against a reference category of managers continuing only in the mutual fund industry. If a mutual fund manager manages multiple funds, we use the average of the fund variables weighted by the assets under management of each fund except when described otherwise. Past performance is a measure (although a noisy one) of the skill of a man- ager. It is also a measure of her visibility since better-performing mutual fund managers enjoy the limelight and are able to attract fund flows. We use perfor- mance before expenses so as to better capture managerial ability. Hedge funds are likely to search for managers who have their own active strategies. Low turnover can be taken as a sign of passive strategy. High track- ing error (calculated either as the standard deviation of residual from a 4-factor model or as the standard deviation of style-adjusted return) would indicate a strategy that is different from the standard 4-factor strategy or from the usual strategy within that style. We also include proportion invested by the manager in common stocks. If hedge funds are looking to invest primarily in equities, they would want mu- tual fund managers with that experience. Total net assets under management of the manager (log of sum of assets across funds) would capture some character- istics attractive to hedge funds, such as reputation, visibility, and ability of the manager to attract funds. We also include mutual fund expense ratio because expenses might reflect the ability of the manager to raise money or some other quality of the manager not captured by performance. We also include experi- ence and experience-squared in our analysis as additional controls related to age and ability to adapt to hedge funds. Table 2 presents the results for multinomial logistic regression using 5-year or 3-year style-adjusted return as a measure of performance. Table 3 presents the results using style-adjusted MPPM with ρ equal to 2.3 All the specifica- tions include year fixed effects, and we cluster the standard errors at the man- ager level. As the tables show, better past performance predicts a side-by-side arrangement, whereas poor past performance predicts an exit from the mutual 3 The results are similar if we use 4-factor alpha or style-adjusted MPPM with ρ equal to 3 or 4. 3014  Table 2 Career moves of mutual fund managers 5-Year Performance 3-Year Performance Complete Complete Side-by-Side Complete Complete Side-by-Side Drop Outs Switchers Managers Drop Outs Switchers Managers Performance −37.490*** −69.161*** 57.746** −27.166*** −40.176** 10.707 Proportion invested in equity 0.078 0.660* −0.291 0.066 0.717** −0.163 Turnover 0.086*** 0.086 0.163* 0.085*** 0.052 0.160** Tracking error −8.204*** −6.694 6.112 −8.564*** −8.138 4.142 Assets under management −0.286*** 0.146* 0.150** −0.283*** 0.124* 0.202*** Experience 0.015* −0.079 −0.025 0.022*** −0.079 −0.009 Experience-squared −0.001* −0.001 −0.002 −0.001** −0.001 −0.003 Expenses −1.050 60.181** 103.100*** −0.803 65.709*** 98.825*** Pseudo R 2 0.107 0.103 Total number of observations 32911 37876 Number in the category 3255 81 102 3825 98 114 p -value for the test that the The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds? coefficients for performance are equal Complete switchers v. side-by-side managers