1. 1 Do Portfolio Managers Moonlighting between Mutual Funds and Hedge Funds Create Conflicts of Interest? Li-Wen Chen National Chi-Nan University Fan Chen Louisiana State University For presentation at NTU International Conference on finance Dec 13th, 2006

2. 2 Background Information Main differences between mutual funds and hedge funds Portfolio holding Compensation structure Regulatory issues

3. 3 Portfolio Holding Mutual fund Mutual fund managers can only trade and hold publicly traded securities. Hedge fund Hedge fund managers can hold a wide spectrum of financial instruments and have the flexibility to pursue whatever investment or trading strategies they choose.

4. 4 Compensation Structures Mutual fund Mutual fund companies charge fees of from 1% to 2% on the amount of assets under management, irrespective of performance. Hedge fund Most hedge fund companies charge a flat management fee of 1% to 2% based on the total assets under management plus 20% to 25% incentive fees.

5. 5 Regulatory Issues Prior to 2004 Mutual funds are subject to extensive disclosure under 1940 Investment Company Act, but hedge funds are not. In May 2005 The SEC adopts 203 (b) (3)-2 which required hedge funds with with 30 million in assets and 15 or more U.S clients during the preceding 12 months to register with SEC under the Advisors Act. In June 2006 A federal appeals court overruled the SEC for too broadly interpreting legal definitions to bring hedge fund under scrutiny.

6. 6 Benefit for Moonlighting Moonlighting provides a valuable vehicle for portfolio managers to target different sets of audiences and yet achieve economies of scale. Reputational capital and marking to market device discipline MF managers and reduce potential agency problems. HF managers increase their recognitions through MF operations.

7. 7 Cost for Moonlighting Potential conflicts of interests on where PMs send the trades. The stealth trading hypothesis (Barclay and Warner (1993)) Moonlighting managers can place hedge fund orders prior to those of a mutual fund to capture higher potential gains from informed trading for hedge fund clients at the expense of mutual fund clients.

8. 8 Why Studying Conflicts of Interests among Moonlighting Fund Managers? U.S. mutual funds are corporations or business trusts, overseen by the board of directors or trustees, and should be organized and operated in the best interest of shareholders. Funds typically have no direct employees but hire management firms and other service providers. This design clearly aims to put investors in the first priority. Fund returns, fund flows and managers’ risk incentives provide direct measurements for conflicts of interests Conflicts of interest between mutual fund investors and fund families arise because investors demand high risk-adjusted performance at low cost while fund families intent to maximize assets under management (i.e., their market share) and the resulting fees.

9. 9 Contributions This paper makes the first attempt in the literature to analyze this issue. Provide empirical evidence on whether moonlighting between different fund portfolios creates conflicts of interests or strengthens incentives.

10. 10 Research Questions First, whether there is a difference of initiative motivations for moonlighting portfolio managers? Second, whether moonlighting among mutual fund and hedge fund create conflicts of interest or strengthen incentives for managers and their shareholders? Third, whether portfolio managers change their risk incentives after moonlighting decisions and whether this change aligns with performance?

11. 11 Major Findings Prior to moonlighting, hedge fund managers experience worse performance, while mutual fund managers achieve better performance, relative to their full-time peers. Hedge fund managers that choose to moonlight are disproportionately those with less experience and poorer performance. Although recognition increases with moonlighting, the asymmetry of performance, asset flows, and risk incentives between portfolios suggest potential conflicts of interest. Reputational capital, marking to market, and option-like incentive contracts induce mutual fund managers that choose to moonlight to take on additional risk, resulting in outperformance relative to their full-time peers.

12. 12 Sample Selection and Methodology Sample period Jan 1980-May 2005 Database HedgeFund.net and CISDM database for hedge funds. Morningstar and CRSP database for mutual funds. Survival bias issue We modify our annual return for each hedge fund in the sample by an annual rate of 3.6% to mitigate upward bias based on Fung and Hsieh (2000) estimate.

13. 13 Sample Selection and Methodology Initial Sample Initial samples contain 109 open-ended mutual funds and 80 hedge funds with 63 classified as 1M2H and 42 classified as 1H2M events, a total of 73 portfolio managers. We drop samples of total assets less than $5 million, funds without 24 consecutive reported monthly returns, unidentified fund managers, with identical starting dates (1H1M or 1M1H), and international funds and bond funds to focus this study of U.S. equity funds. Final Sample 1M2H contains 30 mutual funds and 25 hedge funds with 25 moonlighting managers while 1H2M contains 36 hedge funds and 30 mutual funds with 21 moonlighting managers. We cross check fund managers inception date by direct telephone contact.

14. 14 Sample Selection and Methodology (cont.) Performance measure Monthly raw returns obtained from the Morningstar mutual fund database and HedgeFund.net database. Monthly market-adjusted returns is S&P 500 index for mutual fund and Morningstar/Tremont Hedge Fund Index for hedge fund. Risk-adjusted returns (i.e., alphas) from Shapre’s CAPM for mutual fund and excess-of-risk-free-rate for hedge funds (Fung and Hsieh (2000, 2005) and Boyson (2005). Objective-adjusted returns (Morck, Shleifer, and Vishny (1989) and Khorana (2001)) adopt benchmark index from Morningstar for mutual fund and primary or secondary strategy from HedgeFund.net for hedge fund. Matched sample returns by controlling for style, size and past returns for robustness check purposes.

15. 15 Sample Selection and Methodology (cont.) Flows measure Sirri and Tufano’s (1998) measurement [TNA i, t - TNA i, t-1 * (1+R i,t )]/ (TNA i,t-1 ), where TNA i, t is the total net asset for fund i at time t and R i, t is the raw return at time t. Risk measure Brown, Harlow and Starks (1996) risk adjustment ratio (RAR). We measure RAR by using post risk ratio (variance of raw returns) divide prior risk ratio (variance of raw returns)., t 1 =0 or 1, t 2 =12 or 24 RAR=

16. 16 Sample Selection and Methodology (cont.) Robustness test We create matched samples based on total net assets for mutual funds and the Sharp ratio for hedge funds. The funds are divided into 5 groups after controlling for investment objective. We also develop one-on-one matched samples by matching on the basis of investment objective as well as total net assets and past returns.

17. 17 Empirical Result (prior to moonlighting) (based on performance and flows analysis) 1M2H Mutual fund managers outperform their full-time peers. Mutual fund managers attract less asset flows. 1H2M Hedge fund managers underperform their full-time peers while no statistical difference on asset flows.

18. 18 Empirical Result (prior to moonlighting) (based on performance and flows analysis) Category Returns (-12,-1) Flows (-12,-1) Panel A ： Mutual Fund (1M2H) Sample (N=30)7.75 ** 0.28 *** Matched2.350.86 *** Difference (Sample - Matched) 5.83 ** -0.67 ** Panel B ： Hedge Fund (1H2M) Sample (N=36)2.310.23 ** Matched11.97 *** 0.26 *** Difference (Sample - Matched) -6.87 *** -0.11

19. 19 Empirical Result (after moonlighting) (based on performance and flows analysis) 1M2H Mutual fund performance and asset flows both increase. Hedge fund performance and asset flows both increase. 1H2M Mutual fund performance and asset flows increase. Hedge fund performance decreases but asset flows increase.

20. 20 Empirical Result (based on performance analysis)

21. 21 Empirical result (based on RAR analysis) 1M2H Mutual fund portfolio variance ratio decreases on both event date measure (0, 12) and calendar year measure (-1, +1). 1H2M Mutual fund portfolio variance ratio decreases on both event date measure (0, 12) and calendar year measure (-1, +1). Hedge fund portfolio variance ratio decreases on both event date measure (0, 12) and calendar year measure (-1, +1) although not statistic significant.

22. 22 Empirical result (based on RAR analysis)

23. 23 Regression Analysis We use OLS regression to examine the differences of monthly return on both mutual funds on 1M2H and hedge funds on 1H2M to matched samples in the events of moonlighting from twelve months prior to twenty-four months after the moonlighting events (-12,+24). After control for past flows, expense ratio, turnover, and size, we see the statistically significant for the group of moonlighting to the group without.

24. 24 Regression Analysis

25. 25 Conclusions Our results are consistent with Merton’s (1987) investor recognition hypothesis. Managing publicly disclosed mutual funds increases public recognition for hedge fund managers. However, the asymmetric relationships for the performance, asset flows, and risk incentives of moonlighting hedge fund managers’ existing and newly- managed portfolios suggest that there is a potential for conflicts of interest.

26. 26 Conclusions Option-like incentive compensation contracts provide sufficiently strong incentives for mutual fund managers that choose to moonlight to take on the appropriate risk and to outperform their full-time peers. Even though agency problems exist, reputational capital can discipline agents, helping to solve potential conflicts of interest.