Incentives behind Side-by-Side Management. of Mutual Funds and Hedge Funds *

Transcription

1 Incentives behind Side-by-Side Management of Mutual Funds and Hedge Funds * John Bae, Chengdong Yin, and Xiaoyan Zhang July 2017 Abstract We examine the incentives that motivate management firms to simultaneously manage mutual funds and hedge funds. By identifying side-by-side management at both the management firm level and the manager level, we find that mutual fund management firms use side-by-side practice to retain their talented managers and improve capital flows into their mutual funds. In contrast, hedge fund management firms engage side-by-side practice to collect more fees and smooth their compensation over time. The conflict of interest is more likely to exist when hedge fund managers establish mutual funds, and thus investors do not necessarily suffer from side-by-side management. Key Words: Side-by-Side Management, Mutual Fund, Hedge Fund, Incentive. JEL Classification: G23 * We would like to thank Lu Zheng, Yuehua Tang, Fan Chen, Pedro Matos, and participants at Krannert School Alcoa Workshop, FMA Asia/Pacific Conference 2017, and FMA European Conference 2017 for helpful comments and suggestions. We also want to thank Jinhee Kim and Chao Gao for excellent research assistance work. All remaining errors are ours. Martha and Spencer Love School of Business, Elon University, phone (336) , jbae@elon.edu. Krannert School of Management, Purdue University, phone (765) , yin80@purdue.edu. Krannert School of Management, Purdue University, phone (765) , zhang654@purdue.edu.

2 Incentives behind Side-by-Side Management of Mutual Funds and Hedge Funds July 2017 Abstract We examine the incentives that motivate management firms to simultaneously manage mutual funds and hedge funds. By identifying side-by-side management at both the management firm level and the manager level, we find that mutual fund management firms use side-by-side practice to retain their talented managers and improve capital flows into their mutual funds. In contrast, hedge fund management firms engage side-by-side practice to collect more fees and smooth their compensation over time. The conflict of interest is more likely to exist when hedge fund managers establish mutual funds, and thus investors do not necessarily suffer from side-by-side management. Key Words: Side-by-Side Management, Mutual Fund, Hedge Fund, Incentive. JEL Classification: G23

3 Side-by-side management refers to the practice in which a management firm, and sometimes the same portfolio manager, simultaneously manages multiple products, such as mutual funds, hedge funds, separately managed accounts, and so forth. Many financial institutions over the past two decades have adopted side-by-side management. For instance, Del Guercio, Genc, and Tran (2016) show that 22 of the top 30 largest mutual fund families were involved in certain types of side-by-side management, including Vanguard, Fidelity, and PIMCO. Under the side-by-side management setting, a conflict of interest may arise when management firms and portfolio managers favor one product over another because of different product characteristics such as their fee structures. One important and controversial example of this potential conflict of interest is the arrangement in which a management firm and/or a manager manages both mutual funds and hedge funds. While mutual funds normally charge the asset-based management fee, hedge funds commonly charge an additional performance-based incentive fee. The incentive fee allows hedge fund managers to collect a portion of fund profits as their compensation. Thus, management firms and portfolio managers likely put forth more effort to generate better performance for hedge funds so that they can collect more fees, and this additional effort put forth toward hedge fund performance may come at the expense of mutual fund performance. In practice, the side-by-side management of both mutual funds and hedge funds accounts for a growing slice of the fund universe. 1 For instance, during our sample period, average assets under management of side-by-side mutual funds and hedge funds are about $443 billion and $21 billion per year, respectively. Undoubtedly, investors are concerned about this practice and would like to know answers to questions such as how does side-by-side management affect my 1 See Eleanor Laise, Is your fund manager two-timing you?, Wall Street Journal, November 1, 2006 ( among others. 1

4 current investments?, should I invest in funds under side-by-side management?, and how can I pick among different types of side-by-side managed funds? 2 Therefore, we focus on the side-by-side management of mutual funds and hedge funds in this study, and we address investors concerns by examining the incentives that motivate management firms to engage the side-by-side practice. Naturally, different types of management firms switch from a single industry to the side-by-side practice for different reasons. Some previous studies, such as Chen and Chen (2009) and Nohel, Wang, and Zheng (2010), argue that mutual fund management firms establish hedge funds to keep their talented managers. 3 Because hedge funds allow managers to charge the performance-based incentive fee, portfolio managers with skills have incentives to join the hedge fund industry so that they can improve their compensation. To retain these good managers, mutual fund management firms may allow their good managers to establish hedge funds in the same family. If this argument holds, then side-byside mutual funds, whose management firms start with mutual funds and whose managers run hedge funds simultaneously, may have better performance because their managers are supposed to have better investment skills. Furthermore, because investors chase performance, those funds may also enjoy higher capital inflows. Therefore, our first hypothesis regarding management firms incentives is that mutual fund management firms engage in side-by-side practices to retain their skilled managers and to improve their capital flows. Meanwhile, motivations for hedge fund management firms to engage in side-by-side management might be different. Chen and Chen (2009) argue that hedge fund management firms 2 In this study, we identify funds under side-by-side management at both the management firm level and the portfolio manager level. In the former case, a fund is side-by-side managed if its management firm has mutual funds and hedge funds simultaneously. In the latter case, a fund is side-by-side managed if its portfolio manager has mutual funds and hedge funds concurrently in the same family. For the rest of the paper, side-by-side funds refer to the former case (i.e., side-by-side management at the firm level), unless otherwise specified. 3 In this study, mutual fund management firms refer to firms that only have mutual funds and firms that start with mutual funds and engage in side-by-side practices later. Hedge fund management firms are defined similarly. 2

5 have difficulty attracting capital flows because of information asymmetry and the lack of distribution and marketing channels. As a result, launching mutual funds gives hedge fund management firms a way to expand their clientele and thus improve fund capital flows. Another possible motivation for hedge fund management firms is to smooth their compensation over time. Hedge fund managers compensation might be volatile for several reasons. First, the performancebased incentive fee normally comes with a high-water mark provision, which requires hedge fund managers to make up any past losses before they can charge the incentive fee. Second, hedge fund investors are sophisticated and sensitive to fund performance, and they may withdraw money when funds perform poorly. In other words, when hedge fund performance declines, fund managers may not only lose the incentive fee, but also suffer capital outflows. In contrast, mutual funds charge the asset-based management fee, and the literature shows that mutual fund investors are insensitive to poor performance. 4 Therefore, by establishing mutual funds in the same family, hedge fund management firms might have a more stable source of income and thus smooth their compensation, that is, reduce the volatility of their fee incomes over time. To summarize, our second hypothesis is that hedge fund management firms engage in the side-by-side practice to attract higher flows for their hedge funds and smooth their compensation over time. The potential conflict of interest associated with the side-by-side management has attracted attention from regulators. To protect investors, the U.S. Securities and Exchange Commission (SEC) proposed several changes to fund filings over the years, which require management firms to disclose potential conflicts of interest to investors and how these management firms address 4 See Chevalier and Ellison (1997) and Sirri and Tufano (1998), among others, about mutual fund capital flows. See, for example, Naik, Ramadorai, and Stromqvist (2007), Fung et al. (2008), Getmansky et al. (2015), and Yin (2016) for a discussion of the flow-performance relationship in the hedge fund industry. 3

6 their potential conflicts of interest. 5 In addition, in its national examination program priorities, the SEC mentioned side-by-side management specifically in the years 2014 and 2016, and conflicts of interest have always been the examination priorities of the SEC over the past few years. 6 Thus, our third hypothesis is about the conflict of interest. To do so, we examine whether side-by-side mutual funds underperform their peers without side-by-side management and whether side-byside hedge funds outperform their peers without side-by-side management. This study examines the hypotheses above by analyzing the impact of side-by-side management on fund performance, fund flows, and managers compensation. Our results complement the current literature, which mostly focuses on fund performance, and provide a more comprehensive understanding about the incentives of different types of management firms. Understanding management firms motivations may provide better guidance for investors to choose among different funds and adjust their investments over time. Our study also sheds light upon the design of future regulations regarding side-by-side management. To test the hypotheses above, we begin by examining the performance of side-by-side funds. The impact of side-by-side management on fund performance has been studied in the literature. However, the results are somewhat mixed. Cici, Gibson, and Moussawi (2010) study the side-by-side practice at the management firm level and find that side-by-side mutual funds underperform their peers with similar characteristics. They argue that it is possible that management firms transfer performance from mutual funds to hedge funds to maximize their fee 5 For example, amendments to Form N-1A, Form N-2, and Form N-3 (see require mutual funds to identify each portfolio manager in their prospectuses, disclose other accounts managed by these portfolio managers, and describe the structure of and the method used to determine the compensation received by each manager starting from Amendments to Form ADV in 2010 (see require investment advisers to disclose side-by-side management and describe how they address conflict of interest. 6 Evolution of SEC Examination Priorities , Focus 1 Associates LLC, February 5, 2015 ( focus1associates.com/evolution-sec-examination-priorities /). 4

7 incomes. Del Guercio, Genc, and Tran (2016) look at mutual funds whose managers have three account types, that is, mutual funds, hedge funds, and separate accounts. They find that only mutual funds whose managers simultaneously manage hedge funds underperform their peers. In contrast, Chen and Chen (2009) examine side-by-side management at the manager level, and they argue that the conflict of interest exists when a hedge fund manager starts a mutual fund but not the other way around. Nohel, Wang, and Zheng (2010) study side-by-side management at the manager level and find that side-by-side mutual funds outperform their peers, but side-by-side hedge funds are at best on par with their peers. Thus, they do not find supporting evidence that mutual fund investors suffer from side-by-side management because of the conflict of interest. Following the literature, we identify management firms with side-by-side management by merging CRSP mutual fund database with TASS hedge fund database and matching management firms by name. 7 Using risk-adjusted returns, we find that side-by-side mutual funds outperform their peers without side-by-side management, and side-by-side hedge funds do not perform differently from their peers. To dig deeper, we double sort side-by-side funds into (two-by-two) four groups based on two types of information: whether their management firms hire the same manager to run both mutual funds and hedge funds (i.e., side-by-side management at portfolio manager level) and whether their management firms start with mutual funds or hedge funds. The results show that side-by-side mutual funds, whose management firms originate from the mutual fund industry and whose managers have hedge funds simultaneously, outperform their peers. This is consistent with Chen and Chen (2009), Nohel, Wang, and Zheng (2010), and Deuskar et al. 7 We believe that analyzing side-by-side management at the firm level is important and necessary. The literature shows that management firms may favor high value funds (e.g., high performance or high fee income) over low value funds. See Gaspar, Massa, and Matos (2006), among others, for a discussion. Thus, side-by-side management firms may favor hedge funds because of the high fee income, and this holds true even when different managers manage mutual funds and hedge funds. Besides, the amendments to Form ADV ( secg.htm) require disclosure of side-by-side management at both the management firm level and the manager level. 5

8 (2011) and supports the argument that mutual fund management firms retain their talented managers by allowing them to establish hedge funds. Interestingly, we also find that side-by-side mutual funds established by hedge fund management firms underperform their peers when their managers concurrently have hedge fund. This finding is consistent with Chen and Chen (2009) and implies that a conflict of interest exists when hedge fund managers launch mutual funds. In contrast, side-by-side hedge funds with various backgrounds do not perform differently from their peers. In the second step, we examine the impact of side-by-side management on capital flows. Capital flows are important because they influence fund size and fund performance, which then affects managers compensation. We find that side-by-side mutual funds attract higher flows than their peers without side-by-side management. Further analysis shows that this result is mainly driven by side-by-side mutual funds established by mutual fund management firms but not limited to those funds whose managers run hedge funds at the same time. This indicates that having hedge funds in the same family can be a good signal for mutual fund management firms and enables them to attract higher flows from investors. However, side-by-side mutual funds established by hedge fund management firms have lower flows than their peers, especially when their managers simultaneously manage hedge funds. This is consistent with the poor performance of those funds and implies that investors are smart. In contrast, we do not find significant differences in capital flows between side-by-side hedge funds and their peers without side-by-side management. Thus, our results do not support the hypothesis that hedge fund management firms can use side-by-side practice to improve capital flows into their funds. This is not surprising given that side-by-side hedge funds do not perform differently from their peers and that only qualified investors can invest in hedge funds. 6

9 The findings above provide some insight into the incentives of mutual fund management firms, which include retaining their skilled managers and improving fund capital flows. However, the motivations of hedge fund management firms engaging in side-by-side management are still not clear. So far, we do not find evidence that side-by-side management improves performance or capital flows of funds established by hedge fund management firms. Therefore, in the third step, we look at management firms compensation. As discussed earlier, mutual funds provide a more stable source of income to side-by-side management firms, and hedge fund management firms may engage in side-by-side practices to increase their compensation and smooth their fee incomes, that is, reduce the volatility of their compensation. Consistent with this hypothesis, we find that hedge fund management firms collect a significant portion of family assets and family total compensation from mutual funds during the side-by-side periods. In addition, after hedge fund management firms start side-by-side management, the volatility of their fee incomes decreases significantly. This study contributes to the literature in several ways. First, our study is one of the first few to examine the incentives of side-by-side management firms, and we show that different types of management firms engage in side-by-side management for different reasons. Mutual fund management firms benefit from side-by-side management by retaining their skilled managers and improving capital flows into their mutual funds. More importantly, we provide some novel evidence that hedge fund management firms use side-by-side management to collect more fees and smooth their compensation over time. Second, this study complements the literature regarding the potential conflict of interest in side-by-side management. We reconcile previous research by identifying side-by-side management at both the management firm and the portfolio manager level. We find that side-by-side mutual funds, whose management firms originate from the mutual fund 7

10 industry and whose managers simultaneously manage hedge funds, outperform their peers. Meanwhile, side-by-side mutual funds established by hedge fund management firms underperform their peers when their managers concurrently manage hedge fund. In other words, the conflict of interest is more likely to exist when hedge fund managers start mutual funds. Thus, mutual fund investors do not necessarily suffer from the side-by-side practice. I. Data and Methodology We collect mutual fund data from the CRSP survivorship-bias free mutual fund database, which provides information regarding fund performance, fund size, and fund characteristics such as expenses ratios, total loads, and turnover ratios. Following the literature, we aggregate multiple share classes into a single fund using fund ID information from the MFLINK database and fund assets as the weight. Hedge fund data are from the Lipper TASS hedge fund database. Following the literature, we only keep funds that report monthly net-of-fee returns in US dollars (USD). Fund-month observations with missing information about returns, assets under management, or investment styles are removed. To mitigate survivorship bias, we include defunct hedge funds in our sample. Because the TASS database provides data of defunct funds since 1994, our initial sample starts from January To mitigate backfill bias, we exclude observations before the dates when funds were added to the TASS database. 8 If the add-dates are not available, we exclude the first 18 months of data. 8 This step can also mitigate a potential survivorship bias in the TASS database. See Aggarwal and Jorion (2010), among others, for a discussion. 8

11 In addition, we require both mutual funds and hedge funds in our sample to have at least $5 million under management and 24 months of observations. To eliminate reporting errors and outliers, we winsorize fund returns and capital flows at 1% and 99% level. A. Matching Procedure To identify management firms that engage in side-by-side management, we merge the two databases by firm names. For matched firms, we manually cross-check their information with several sources, such as S&P Capital IQ and internet search, to make sure that the two names refer to the same firm. 9 We then identify the time periods during which a management firm has at least one mutual fund and at least one hedge fund, respectively. If there is an overlap between the two periods, then the management firm is considered a side-by-side firm during the overlapped period. Eventually, we have 88 management firms that engage in side-by-side management from January 1994 through December Table I shows the number of mutual funds and hedge funds that are under side-by-side management by year. Consistent with the literature, such as Nohel, Wang, and Zheng (2010), most funds under side-by-side management are launched after Therefore, to make our analysis more meaningful, we focus on the sample period from January 2001 through December 2014 in the following analysis. During this period, side-by-side management firms manage a total of 853 mutual fund and 192 hedge funds. 10 Panels A and B of 9 In our sample, we exclude different branches of the same firm in this study because they commonly have different ID numbers and are treated as different firms in the databases. We do not use mutual fund filings to identify side-byside management, as in Del Guercio, Genc, and Tran (2016), for several reasons. First, mutual fund filings do not provide any information regarding the hedge funds managed by the same manager. Thus, we cannot analyze the performance, capital flows, or compensation of those hedge funds. Second, required disclosure of side-by-side management started after The number of side-by-side hedge funds in our sample is relatively smaller than in other studies in the literature for several reasons. First, we require hedge funds to have at least $5 million assets. Many small funds are dropped from the sample. Second, we exclude backfilled data of hedge funds from the sample. This may also exclude some side-byside periods. 9

12 Internet Appendix Table IAI report the number of side-by-side mutual funds and hedge funds by style categories, respectively. While there are variations in styles of mutual fund under side-byside management, equity mutual funds account for the majority in our sample. Thus, following Nohel, Wang, and Zheng (2010) and Del Guercio, Genc, and Tran (2016), we also conduct the analysis separately for equity mutual funds. 11 For side-by-side hedge funds, the first and second largest styles are Long/Short Equity Hedge and Fund of Hedge Funds, respectively. Because Fund of Hedge Funds style invests in other hedge funds rather than directly in securities, we perform all following analyses on Fund of Hedge Funds separately, and we refer to hedge funds in other styles as regular hedge funds. [Insert Table I about here] Note that mutual funds and hedge funds that belong to the same management firms may have different managers. This may lead to the different findings documented in the prior literature. Therefore, we further match portfolio managers by their names within the identified side-by-side firms defined above. For matched managers, we compare their tenure periods reported in the CRSP database with those in the TASS database. If there is an overlap, then they are considered as sideby-side portfolio managers. Within side-by-side management firms, we are able to identify 83 managers who have simultaneously managed at least one mutual fund and one hedge fund during our sample period. In the following analysis, we will use side-by-side information at both the management firm level and the portfolio manager level. 11 In this study, we focus on well-diversified domestic equity mutual funds that belong to the following CRSP styles: EDCM, EDCS, EDCI, EDYG, EDYB, and EDYI. We exclude index funds and ETFs using identifiers provided by CRSP database. 10

13 B. Performance and Capital Flow Measures As in Nohel, Wang, and Zheng (2010) and Del Guercio, Genc, and Tran (2016), we use risk-adjusted returns to measure fund performance. 12 To be more specific, we use Carhart (1997) four-factor model for mutual funds, and the risk-adjusted return of mutual fund i in month t would be: AAAAAAhaa ii,tt = RR ii,tt rr ff,tt ββ 1 RR MM,tt rr ff,tt + ββ 2 SSSSSS tt + ββ 3 HHHHHH tt + ββ 4 MMMMMM tt, (1) where RR ii,tt is the after-fee return of fund i in month t, rr ff,tt is the risk-free rate, RR MM,tt is the market return, SMB is the size factor, HML is the value factor, and MOM is the momentum factor. We estimate the factor loadings using a three-year rolling window with at least 24 observations. Because hedge funds can invest in multiple asset classes, the four-factor model above may not fully capture the risk-taking of hedge funds. Thus, following Nohel, Wang, and Zheng (2010), we calculate risk-adjusted returns of hedge funds based on the Fung and Hsieh (2004) seven-factor model. To be more specific, the alpha of hedge fund i in month t is defined as follows: AAAAAAhaa ii,tt = RR ii,tt rr ff,tt (ββ 1 SS&PP tt + ββ 2 SSSSSS tt + ββ 3 BBBB10YY tt + ββ 4 CCCCCCCCCCCCCC tt + ββ 5 PPPPPPPPPPPP tt + ββ 6 PPPPPPPPPPPPPP tt + ββ 7 PPPPPPPPPPPP tt ), (2) where the set of factors comprises the equity market factor (S&P), measured as Standard & Poor s 500 index monthly total return, the size spread factor (SML), constructed as the difference between the Russell 2000 index monthly total return and S&P s 500 monthly total return, the bond market factor (BD10Y), which is the monthly change in the 10-year Treasury constant-maturity yield, the credit spread factor (CredSpr), calculated as the monthly change in Moody s Baa yield less the 10- year Treasury constant-maturity yield, and three trend-following risk factors, namely, the excess 12 We present results using alternative performance measures, such as return gap in Kacperczyk, Sialm, and Zheng (2008) and the CS measure in Daniel et al. (1997), in Section V Robustness Tests. 11

14 returns on portfolios of look-back straddle options on currencies (PTFSFX), commodities (PTFSCOM), and bonds (PTFSBD). 13 Again, we use a rolling-window to estimate factor loadings. Following Sirri and Tufano (1998), we assume that capital flows happen at the end of each month. Capital flows of fund i in month t are defined as follows: where AUM is assets under management. FFFFFFFF ii,tt = AAAAAA ii,tt AAAAAA ii,tt 1 (1+RRRRRRRRRRRR ii,tt ) AAAAAA ii,tt 1, (3) C. Summary Statistics Panel A of Table II reports the summary statistics of mutual funds. We divide funds into two groups, that is, funds that have experienced side-by-side management (the SBS group) and funds that have never been under side-by-side management (the Non-SBS group) during our sample period. We compare mutual funds in these two groups using t-tests. The results indicate that mutual funds in the SBS group deliver better performance than funds in the Non-SBS group. The difference in net-of-fee return is 7 bps per month and is significant at the 5% level. These funds also generate higher risk-adjusted returns by 1 bps per month, but the difference is not significant. Better performance leads to higher flows as well. Mutual funds in the SBS group attract 0.31% more capital flows per month. In terms of fund characteristics, mutual funds with side-byside experience tend to be younger, and smaller in size, but exhibit higher expense ratios and management fees than their peers. [Insert Table II about here] 13 Data of the three trend-following factors are available at FAC.xls. 12

15 Panel B of Table II reports the summary statistics of hedge funds. Similarly, dependent on whether hedge funds have been under side-by-side management during our sample period, we divide hedge funds into two groups. On average, hedge funds in the SBS group have higher netof-fee returns (by 7 bps per month) but lower risk-adjusted returns (by 7 bps per month), and both differences are significant at 5% level. The mixed performance may lead to the insignificant difference in capital flows between the two groups. In terms of fund characteristics, hedge funds in the SBS group have smaller size but older age. They charge lower management fees but higher incentive fees and are less likely to use leverage than their peers. II. Performance of Side-by-side Funds In this section, we examine the incentives of management firms by analyzing the impact of side-by-side management on fund performance. We start with an analysis of side-by-side funds at the management firm level in section II.A. We then differentiate side-by-side funds based on whether their portfolio managers have both mutual funds and hedge funds simultaneously in section II.B. The results would help us reconcile different findings in the literature. In section II.C, we test whether mutual fund management firms engage in side-by-side practices to retain their talented managers and whether mutual fund investors suffer from potential conflicts of interest. A. Side-by-side Management at Management Firm Level To examine the impact of side-by-side management on fund performance while controlling for various fund characteristics, we use a pooled regression as in equation (4). The dependent variable is risk-adjusted returns. The key independent variable is the SBS indicator, which equals one if fund i is under side-by-side management in month t and zero otherwise. Common 13

16 independent variables for both mutual funds and hedge funds include fund size, fund family size, capital flows, and fund age. Fund family size is the total assets of other funds in the same family. We also include total loads, expense ratios, and turnover ratios as control variables for mutual funds. For hedge funds, additional control variables are the fee structure (i.e., the management fee percentage, the incentive fee percentage, and the high-water mark) and share restrictions (e.g., redemption frequency, notice periods, and lockup periods). Style fixed effects and year fixed effects are also included. Following Petersen (2009), we cluster standard errors by fund. PPPPPPPPPPPPPPPPPPPPPP ii,tt = ββ 0 + ββ 1 SSSSSS ii,tt + ββ 2 FFFFFFFF SSSSSSSS ii,tt 1 + ββ 3 FFFFFFFF FFFFFFFFFFFF SSSSSSSS ii,tt 1 +ββ 4 FFFFFFFF ii,tt 1 + ββ 5 FFFFFFFF AAAAAA ii,tt 1 + CCCCCCCCCCCCCC VVVVVVVVVVVVVVVVVV + εε ii,tt (4) The regression results for mutual funds are summarized in Panel A of Table III. In the first column, the coefficient on the SBS indicator is positive and marginally significant for all mutual funds. Thus, after controlling for fund characteristics, side-by-side mutual funds outperform their peers without side-by-side management by 4 bps per month (or 0.5% per year). The results for well-diversified domestic mutual funds in the second column are much stronger. The coefficient of the SBS indicator is positive and significant at the 1% level, and side-by-side equity mutual funds outperform their peers by 8 bps per month (or about 1% per year). In terms of fund characteristics, we find that mutual funds with older age, lower expense ratio, and higher flows experience better performance. [Insert Table III about here] Table III, Panel B presents the regression results of hedge funds. The negative coefficients on the SBS indicator suggest that both regular hedge funds and funds of hedge funds that are under side-by-side management underperform their peers without side-by-side management, but the difference is not significant. In other words, side-by-side hedge funds are at best on par with their 14

17 peers. In terms of fund characteristics, regular hedge funds with high-water mark provisions and longer redemption notice periods are more likely to have better performance. Funds of hedge funds that have longer redemption frequencies, longer redemption notice periods, and longer lockup periods generate better performance. The results in Table III suggests that side-by-side mutual funds generate better performance than their non-side-by-side peers, but side-by-side hedge funds do not perform differently from their peers. In Internet Appendix Section B, we examine whether there is any significant change in performance after side-by-side management starts. To be more specific, we add to equation (4) a Pre-SBS indicator, which equals one if a fund is not under side-by-side management in month t but will be side-by-side managed later and zero otherwise. Panel A of Table IAII presents regression results of mutual funds. For all mutual funds, the coefficients of Pre-SBS and SBS are both positive but not significant at conventional levels. The F-test indicates that the two coefficients are not significantly different. For equity mutual funds, the coefficient of the Pre-SBS indicator is positive and insignificant, but the coefficient on SBS is positive and significant. Interestingly, the F-test shows that the coefficients on Pre-SBS and SBS are not significantly different. Thus, we do not find evidence that the performance of mutual funds deteriorates after the side-by-side practice starts. The results for hedge funds are shown in Panel B of Table IAII. Consistent with Table III, side-by-side hedge funds do not perform differently from their peers, given the insignificant coefficients on the SBS indicator. Based on our F-test values, the difference between the coefficients on Pre-SBS and SBS are not significant in either regression. In other words, hedge fund performance does not significantly change after side-by-side management starts. Tables IAII 15

18 complements Table III and suggests that investors do not necessarily suffer from side-by-side management. B. Side-by-side Management at Manager Level The previous section examines the impact of side-by-side management on fund performance at the firm level. However, the literature (e.g., Nohel, Wang, and Zheng (2010) and Del Guercio, Genc, and Tran (2016)) argues that the conflict of interest is more likely to influence fund performance when the same manager has both mutual funds and hedge funds. Because of the performance-based incentive fee, fund managers might be motivated to favor hedge funds over mutual funds. As a result, side-by-side mutual funds at the portfolio manager level are expected to have poor performance, and side-by-side hedge funds at the portfolio manager level are expected to have good performance. To examine the impact of side-by-side management at the portfolio manager level on fund performance, we add Same Mgr to equation (4) as below. Same Mgr is an indicator that equals one if the portfolio manager of fund i has at least one mutual fund and one hedge fund under management in month t and zero otherwise. 14 Other independent variables and control variables are similar as before. PPPPPPPPPPPPPPPPPPPPPP iiii = ββ 0 + ββ 1 SSSSSS iiii + ββ 2 SSSSSS iiii SSSSSSSS MMMMMM iiii + ββ 3 FFFFFFFF SSSSSSSS ii,tt 1 +ββ 4 FFFFFFFF FFFFFFFFFFFF SSSSSSSS ii,tt 1 + ββ 5 FFFFFFFF ii,tt 1 + ββ 6 FFFFFFFF AAAAAA ii,tt 1 +CCCCCCCCCCCCCC VVVVVVVVVVVVVVVVVV + εε iiii. (5) 14 Notice that, because we match portfolio managers within side-by-side management firms, the SBS indicator always equals one when the Same Mgr indicator equals one. Thus, we do not include the individual effect of Same Mgr in the regression. 16

19 Panel A of Table IV shows results for mutual funds. In the first regression, the coefficient of the interaction term between the SBS indicator and the Same Mgr indicator is positive and significant at the 5% level. Thus, mutual funds whose managers have hedge funds at the same time outperform other side-by-side mutual funds by 0.13% per month (or 1.57% per year). They outperform their peers without side-by-side management by (i.e., SBS+SBS Same Mgr) by bps per month (or 1.75% per year). The difference is economically large and statistically significant (F=9.00, p-value=0.0027). The results for equity mutual funds in the second column are stronger. Side-by-side equity mutual funds at the portfolio manager level outperform other side-by-side equity mutual funds by 0.19% per month (or 2.28% per year). They outperform their peers without side-by-side management by 0.21% per month (or 2.52% per year). In contrast, at the manager level, side-by-side regular hedge funds and side-by-side funds of hedge funds do not perform differently from other side-by-side funds. In both regressions, the interaction terms in Panel B of Table IV are insignificant. The difference in performance between side-by-side hedge funds and their peers without side-by-side management is not significant either (for regular hedge funds, F=0.71, p-value=0.3980; for funds of hedge funds, F=0.20, p- value=0.6551). In summary, the results in Table IV are consistent with the findings in Nohel, Wang, and Zheng (2010). That is to say, side-by-side mutual funds at the portfolio manager level outperform mutual funds without side-by-side management, and side-by-side hedge funds at the portfolio manager level do not perform differently from their peers. [Insert Table IV about here] 17

20 C. Joint Impact: Same Manager and Management Firm Type So far, we have examined performance of side-by-side funds at both the management firm level and the portfolio manager level. However, to test our hypotheses and better understand the incentives that motivate different management firms to engage in side-by-side practices, we need to dig deeper. To be more specific, in addition to the portfolio manager information, we also distinguish side-by-side funds by the background of their management firms. A management firm is considered as a mutual fund management firm if it only has mutual funds or if it starts with mutual funds and switches to side-by-side management later. Hedge fund management firms are defined similarly. As discussed earlier, different management firms engage in side-by-side management for different reasons and thus have a different impact on fund performance. Furthermore, management firms may have advantages in their original business. For instance, mutual fund management firms are more familiar with how to comply with mutual fund regulations, while hedge fund management firms have more experience with strategies involving short selling and derivatives, which are not allowed for mutual funds. Using portfolio manager information and management firms background information, we can double sort each side-by-side fund into one of the following four groups: (1) mutual fund management firm and same manager, (2) mutual fund management firm and different managers, (3) hedge fund management firm and same manager, and (4) hedge fund management firm and different managers. Here, same manager means fund i is under side-by-side management at the portfolio manager level, and different manager means fund i is under side-by-side management only at the management firm level. In this way, we can examine the joint impact and see whether side-by-side funds in different groups have different performances. For example, if our first hypothesis (i.e., mutual fund management firms establish hedge funds to retain their talented 18

21 managers) is true, then side-by-side mutual funds in group (1) should have better performance, because their managers are supposed to have skills. To be more specific, we add MFtoSBS, which equals one if a side-by-side fund s management firm starts with mutual funds (we use HFtoSBS, which is defined similarly, for hedge funds) and zero otherwise, to equation (5): 15 PPPPPPPPPPPPPPPPPPPPPP iiii = ββ 0 + ββ 1 SSSSSS iiii + ββ 2 SSSSSS iiii SSaammmm MMMMMM iiii + ββ 3 MMMMMMMMMMMMMM ii +ββ 4 SSSSSS iiii MMMMMMMMMMMMMM ii + ββ 5 SSSSSS iiii SSSSSSSS MMMMMM iiii MMMMMMMMMMMMMM ii +CCCCCCCCCCCCCC VVVVVVVVVVVVVVVVVV + εε iiii. (6) Panel A of Table V reports the regression results of equation (6) for mutual funds. To facilitate the interpretation of the results, we summarize the total effects, that is, the difference in performance between different types of side-by-side funds and their peers without side-by-side management, in Panel B. The positive and significant coefficient of the SBS indicator suggests that mutual funds, which hedge fund management firms establish and whose managers do not have hedge funds, outperform their peers without side-by-side management by 12 bps per month (or 1.44% per year). In contrast, mutual funds, which hedge fund management firms establish and whose managers simultaneously run hedge funds, underperform their peers by (i.e., SBS+SBS Same Mgr) bps per month (or 3.2% per year; F=3.61, p-value=0.0576). For sideby-side mutual funds established by mutual fund management firms, if their managers do not have hedge funds, they do not perform differently from their peers (SBS+SBS MFtoSBS= , F=0.00, p-value=0.9679). However, if their managers simultaneously run hedge funds, they outperform their peers (i.e., SBS + SBS Same Mgr + SBS MFtoSBS + SBS MFtoSBS Same Mgr) by bps per month (or 1.76% per year) and the difference is highly significant (F=8.36, p-value=0.0039). 15 Internet Appendix Section B.2 presents results with only the SBS and MFtoSBS indicators. 19

22 [Insert Table V about here] The results for equity mutual funds are qualitatively the same. For example, mutual funds whose management firms have a mutual fund background and whose managers simultaneously run hedge funds outperform their peers without side-by-side management by 0.25% per month (or 3% per year). Meanwhile, mutual funds established by hedge fund management firms underperform their peers by 0.23% per month (or 2.76% per year) if their managers have hedge funds at the same time. Panel C of Table V presents the regression results for hedge funds, and Panel D summarizes the total difference in performance for different types of side-by-side hedge funds. In general, there is no significant difference in performance between various types of side-by-side hedge funds and their peers without side-by-side management. In summary, Table V provides several interesting findings regarding the impact of sideby-side management on fund performance. First, mutual funds whose management firms originate from the mutual fund industry and whose managers simultaneously run hedge funds outperform their peers. This result suggests that those managers have skills and is consistent with the argument that mutual fund management firms engage in side-by-side practices to retain their talented managers. Second, side-by-side mutual funds established by hedge fund management firms underperform their peers when their managers concurrently manage hedge funds. This result implies that conflicts of interest may exist when hedge fund managers engage in side-by-side practices. This result is consistent with the findings in Chen and Chen (2009). Third, consistent with Nohel, Wang, and Zheng (2010), we find that side-by-side hedge funds with various background do not perform much differently from their peers. 20

23 III. Capital Flows of Side-by-side Funds How side-by-side management influences capital flows has not been examined thoroughly in the literature. This is an important question for fund management firms because capital flows influence fund size and future fund performance and in turn affect managers compensation. Thus, examining the impact on capital flows can help us understand the incentives that motivate firms to engage in side-by-side management. For instance, mutual fund management firms can retain their best managers by allowing them to simultaneously manage hedge funds. Thus engaging in sideby-side practices might be a good signal of mutual fund managers quality, and it may help attract more capital inflows from mutual fund investors. However, mutual fund investors might be concerned that fund managers put more efforts into hedge funds because of the additional incentive fee. If this concern dominates, side-by-side mutual funds might suffer from capital outflows. For hedge fund management firms, one potential benefit of establishing mutual funds is to help them expand their clientele base and attract more investors. However, not all mutual fund investors are qualified to invest in hedge funds. In this section, we would like to examine the possibilities above. Section III.A provides a preliminary analysis at the management firm level. Section III.B examines flows of side-by-side funds at portfolio manager level. Section III.C examines whether different types of side-by-side funds attract different flows. A. Side-by-side Management at Management Firm Level To get a heuristic understanding of the impact on capital flows, we regress fund flows on the SBS indicator and control for various fund characteristics as follows, FFFFFFFF iiii = ββ 0 + ββ 1 SSSSSS iiii + ββ 2 PPPPPPPPPPPPPPPPPPPPPP ii,tt 1 + ββ 3 FFFFFFFF SSSSSSSS ii,tt 1 +ββ 4 FFFFFFFF FFFFFFFFFFFF SSSSSSSS ii,tt 1 + ββ 5 FFFFFFFF ii,tt 1 21

24 +ββ 6 FFFFFFFF AAAAAA ii,tt 1 + CCCCCCCCCCCCCC VVVVVVVVVVVVVVVVVV + εε iiii. (7) The dependent variable is fund flows as defined in equation (3). Following the literature, we include past fund performance in the regression. For mutual funds, we use risk-adjusted return in the last month. For hedge funds, share restrictions are commonly used. In our sample, hedge funds, on average, have a redemption frequency of about 90 days and a notice period of about 30 days. Therefore, we include hedge fund performance over the past three months in the regression. We also include the standard deviation of fund performance over the past quarter to measure risk. The SBS indicator and other variables are similar as before. In Panel A of Table VI, the positive and significant coefficient of the SBS indicator suggests that mutual funds under side-by-side management attract capital inflows bps per month higher than their peers without side-by-side management. This is consistent with the hypothesis that having hedge funds in the same family might be a good signal to attract mutual fund investors. We find similar results for equity mutual funds, and side-by-side equity mutual funds attract inflows bps per month higher than their peers. As for control variables, mutual funds with better past performance, younger age, lower expenses, and higher past flows attract higher capital flows. [Insert Table VI about here] In contrast, capital flows of side-by-side hedge funds are not different from their non-sideby-side peers, given the insignificant coefficient of the SBS indicator in Panel B of Table VI. The results are consistent with the fact that most mutual fund investors are not qualified to invest in hedge funds, and hedge funds commonly have high minimum investment requirements and are not open to public. This is also consistent with the results in Table III that side-by-side hedge funds 22

25 do not perform differently from their peers. In terms of control variables, hedge funds with better past performance, lower return volatility, higher past flows attract higher flows. In Internet Appendix Section C, we examine whether fund flows change significantly after side-by-side management begins. We add a Pre-SBS indicator, which equals one if a fund is not under side-by-side management in month t but will be side-by-side managed later, to equation (7). Panel A of Table IAIV shows that, for all mutual funds, the difference between Pre-SBS and SBS is significant (F=10.62, p-value=0.0011), and thus there is a significant increase in fund flows after side-by-side management starts. This is consistent with the possibility that investors view the sideby-side practice of mutual fund management firms as a good signal of managers quality. The results are similar for equity mutual funds, but the difference between the two coefficients is not significant (F=2.40, p-value=0.1217). In Panel B of Table IAIV, the difference between coefficients on Pre-SBS and SBS is not significant (F=0.01, p-value=0.9301) for regular hedge funds. Thus, there is no significant change in capital flows after the side-by-side practice begins. In contrast, funds of hedge funds attract higher flows than their peers before side-by-side practice starts, and based on the F-test (F=6.66, p-value=0.0100), the change in flows after side-by-side management begins is significant. The results in Table IAIV, in addition to the results in Table VI, provide evidence that mutual funds benefit from side-by-side practice by attracting more capital inflows, but side-by-side management does not improve capital flows of hedge funds. B. Side-by-side Management at Manager Level In this section, we examine whether side-by-side funds at the portfolio manager level attract different flows. On the one hand, Table IV shows that side-by-side mutual funds at manager 23

26 level generate better performance than their peers without side-by-side management. If investors chase performance, then those funds might attract higher inflows. On the other hand, if investors worry that portfolio managers favor hedge funds, then they may avoid side-by-side mutual funds at the manager level. To see which effect dominates, we add the Same Mgr indicator to equation (7). Table VII, Panel A shows regression results for mutual funds. In the first column, the coefficient of the SBS indicator is positive and significant. Thus, side-by-side mutual funds at the management firm level attract higher flows than their peers without side-by-side management by bps per month. The coefficient on the interaction term is insignificant. Thus, side-by-side mutual funds at manager level do not attract additional inflows. However, mutual funds with sideby-side management at the manager level still attract higher capital flows (i.e., SBS+SBS Same Mgr) by bps per month (F=5.41, p-value=0.0201) than their peers without side-by-side management. In the second column, the coefficients on the SBS indicator and on the interaction term are insignificant. Thus, we do not find significant results for equity mutual funds. [Insert Table VII about here] Panel B of Table VII reports the regression results for hedge funds. For regular hedge funds in the first column, the coefficient of the interaction term is insignificant, which indicates that capital flows of side-by-side hedge funds at the manager level are not different from those at the management firm level. Furthermore, there is no significant difference in capital flows between side-by-side hedge funds at the manager level and their peers without side-by-side management (F=0.22, p-value=0.6382). In the second column, the negative and significant coefficient of the SBS indicator suggests that side-by-side funds of hedge funds at the management firm level attract lower flows than their peers without side-by-side management by 0.57% per month. The 24