Onshore and Offshore Hedge Funds: Are They ... - Semantic Scholar

Onshore and Offshore Hedge Funds: Are They Twins? George Aragon Arizona State University Bing Liang University of Massachusetts Amherst Hyuna Park Minnesota State University* First Draft: February 26, 2006 This Version: January 27, 2011

Abstract Contrary to offshore hedge funds, US-registered (“onshore”) funds are subject to strict marketing prohibitions, accredited investor requirements, limited number of investors, and tax disadvantage. We exploit this difference to test predictions about organizational design, capital flow, and fund performance. We find that onshore funds impose stronger share restrictions such as a lockup provision than offshore funds, but hold more liquid assets. Our results show that capital flows are less sensitive to past performance in onshore funds than in offshore funds due to regulation on advertising, and the flow sensitivity difference affects performance. Liquidityadjusted alpha is positive and significant (0.94% per month) only for stand-alone onshore funds that have not been affected by strong capital flows from offshore investors through a masterfeeder structure. Key words: offshore hedge funds, lock-up provision, liquidity risk, master-feeder structure                                                         * George Aragon is at W.P. Carey School of Business, Arizona State University, Tempe, AZ 85287-3406, phone: (480) 965-5810, e-mail: [email protected]. Bing Liang is at Isenberg School of Management, University of Massachusetts, 121 Presidents Drive, Amherst, MA 01003-9310, phone: (413) 545-3180, e-mail: [email protected]. Hyuna Park is at the College of Business, Minnesota State University, Mankato, 150 Morris Hall, Mankato, MN 56001, phone: (507) 389-5406, e-mail: [email protected]. Previous versions of this paper were circulated under the title of “Share Restrictions, Liquidity Premium, and Offshore Hedge Funds”. We are grateful for comments from Turan Bali, Arnoud Boot, Stephen Brown, Tom Fraser, Mila Getmansky, Hossein Kazemi, Bernard Morzuch, Joseph Reising, Tom Schneeweis, Paula Tkac, Mingming Zhou, and participants at the 2007 FMA annual meeting, the 2007 China International Conference in Finance (CICF), the Center for International Securities and Derivatives Markets (CISDM) 2007 annual research conference, the 2008 Financial Intermediation Research Society (FIRS) Conference, and the seminar participants at the University of Amsterdam, Binghamton University, Koc University, University of Massachusetts Amherst, and Minnesota State University Mankato. We are responsible for any error.

1

Electronic copy available at: http://ssrn.com/abstract=967788

JEL classification: G11, G12, G23, G32

2 Electronic copy available at: http://ssrn.com/abstract=967788

Onshore and Offshore Hedge Funds: Are They Twins?

Abstract Contrary to offshore hedge funds, US-registered (“onshore”) funds are subject to strict marketing prohibitions, accredited investor requirements, limited number of investors, and tax disadvantage. We exploit this difference to test predictions about organizational design, capital flow, and fund performance. We find that onshore funds impose stronger share restrictions such as a lockup provision than offshore funds, but hold more liquid assets. Our results show that capital flows are less sensitive to past performance in onshore funds than in offshore funds due to regulation on advertising, and the flow sensitivity difference affects performance. Liquidityadjusted alpha is positive and significant (0.94% per month) only for stand-alone onshore funds that have not been affected by strong capital flows from offshore investors through a masterfeeder structure. Key words: offshore hedge funds, lock-up provision, liquidity risk, master-feeder structure JEL classification: G11, G12, G23, G32

3 Electronic copy available at: http://ssrn.com/abstract=967788

1. Introduction The hedge fund industry has grown rapidly in the last decade. In particular, offshore funds registered in low-tax jurisdictions such as the Cayman Islands and British Virgin Islands have grown much faster than US-registered (“onshore”) funds (27.2% vs. 15.2% per annum during 1994-2008 in terms of total assets under management (AUM) according to the Lipper TASS hedge fund data). As of December 2008, 64.3% of the total assets are managed by offshore funds, while onshore funds manage only 22.5% of the total assets. In contrast, the proportion was 20.3% offshore vs. 31.6% onshore in December 1993. The fast growth of offshore funds can be explained by capital flows from institutional investors who increased allocation to alternative investments after the U.S. equity market downturn in the early 2000s. Agarwal and Naik (2005) report that there has been a shift in the type of hedge fund investor: in the early 1990s, the typical hedge fund investor was a high networth U.S. individual investor (who needs onshore funds), but today the typical investor is an institutional investor (who prefers offshore funds due to tax reasons).1 With the rapid growth of the industry, hedge fund research has also proliferated. Researchers have examined whether this growth is accompanied by positive risk-adjusted performance. Some studies find that top-performing hedge funds can consistently deliver alpha, and their performance cannot be explained by luck (Fung and Hsieh (1997, 2004), Ackermann et

                                                        1

LePree (2008) shows that U.S. tax-exempt institutional investors such as endowments and pension funds prefer offshore funds to onshore funds in order to avoid unrelated business income tax (UBIT). As offshore hedge funds are corporations, income from leveraged strategies can be converted into dividends and avoid UBIT. However, onshore hedge funds cannot provide such benefits to institutional investors because they are pass-through entities such as a limited partnership in order to avoid double taxation in the United States (McCrary (2002)). See Section 3.4 and Table 1 for details on the legal structural difference between onshore and offshore hedge funds.

4

al. (1999), Brown et al. (1999), Agarwal and Naik (2004), Kosowski et al. (2007), Fung et al. (2008), and Jagannathan et al. (2010)).2 Despite the growing importance of offshore funds, most studies treat onshore and offshore hedge funds as a monolithic group.3 However, onshore hedge funds face regulatory constraints on capital formation that do not apply to offshore funds. Specifically, most onshore hedge funds rely upon a regulatory exemption pursuant to either Section 3(c)(1) or Section 3(c)(7) of the Investment Company Act of 1940.4 The exemptions impose restrictions on the number and type of investors, and also the manager’s ability to advertise the sale of securities to potential investors. In contrast, offshore funds are not constrained with respect to both the level of investor capital in the fund and the flow of new capital into the fund. Existing research suggests that constraints on raising capital could significantly affect fund performance, investor flows, and organizational design. In this paper, we examine these differences using a large sample of onshore and offshore hedge funds over the time period of 1994-2005. We report several new empirical findings. First, we find that onshore funds impose tighter restrictions on investor redemptions, including longer lockup periods, higher minimum investment, less frequent funding cycles, and longer redemption notice periods than offshore funds. All the differences are statistically significant at the 1 percent level. Second, we find that onshore funds manage assets with higher liquidity and lower liquidity risk than offshore funds on average. The evidence suggests that share restrictions are more likely when equity funding is restricted, because onshore funds cannot advertise their                                                         2

Recent research also finds positive alphas of hedge funds can be interpreted as a compensation for holding illiquid fund shares. See, e.g, Liang (1999), Aragon (2007), Bali, Gokcan and Liang (2007), Ding et al. (2009). 3 As commercial hedge fund databases such as TASS provide a combined data set of onshore and offshore hedge funds, most previous studies do not differentiate offshore funds from onshore funds in their research. One exception is Brown, Goetzmann, and Ibbotson (1999) who study offshore fund performance and attrition specifically by using data from the U.S. Offshore Funds Directory. 4 Among the 1,524 onshore hedge funds in our data, 1,400 funds (91.86%) are exempt from registration and only 124 funds are registered investment advisers.

5

performance and must limit the number of accounts. Our findings also suggest that offshore funds can more efficiently manage illiquid assets, because they face lower equity funding risk than their onshore counterparts. Third, we find that onshore funds are associated with lower assets under management and a significantly lower sensitivity of investor flows to past performance. In addition, the reduced flow/performance sensitivity is more pronounced following lagged positive performance, where advertising constraints are more likely to bind. Overall the evidence supports our key identifying assumption that the exemption requirements faced by onshore funds constrain the amount and flow of investor capital under management. Finally, we find that alpha is positive and significant only for stand-alone onshore funds (0.94% per month). In contrast, the average alpha of either offshore funds or onshore funds that are part of a master-feeder (MF) structure is insignificant. The offshore MF vehicle is a means to raise additional capital from non-U.S. investors and U.S. institutional investors.5 We interpret this finding as consistent with the predictions of Berk and Green (2004, hereafter BG), Pastor and Stambaugh (2010, hereafter PS), and Goetzmann, Ingersoll, and Ross (2003). Specifically, hedge fund managers who set up the MF structure to accept capital flows from both onshore and offshore investors lose their ability to deliver alpha due to decreasing returns to scale, while those who operate within the capacity limit of their strategy as a stand-alone (SA) onshore fund could have delivered high risk-adjusted performance. Previous research finds that mutual fund investors chase performance but performance is not persistent. To explain this seemingly puzzling flow-performance relationship, BG (2004) builds a theoretical model in which investor capital is supplied competitively as investors update                                                         5

See Section 3.1 for the details on how a hedge fund manager can use the MF structure to attract capital from both onshore and offshore investors.

6

their beliefs about manager ability. A central prediction of their model is that investors are sensitive to past returns and supply capital such that excess returns are zero on average. Consistent with BG (2004), Fung et al. (2008) find that the inflow of new capital to better performing funds-of-funds leads to the erosion of superior performance over time. Recently, PS (2010) extends the BG model by assuming that the decreasing returns to scale happens at the entire industry level instead of the individual fund level while investors face endogeneity and disinvest less in the poor performing industry. An important difference between these two models is, alpha can be positive in the PS model, but zero alpha is a necessary condition for equilibrium in the BG model. One caveat is that these models assume no incentive fee because they are developed to explain the flow-performance relation in mutual funds. However, incentive fee is an important part of the compensation package for hedge fund managers.6 Our results are supportive of the competitive markets view for our sample of hedge funds. In particular, we find no evidence of positive alpha on average among offshore hedge funds, where the premise of competitive supply of capital is more likely to be satisfied. In addition, we reach a similar finding for US-registered managers that also manage offshore accounts as part of a MF structure. Apparently, capital flows from offshore investors reduce a manager’s ability to deliver alpha due to decreasing returns to scale. In fact, we find positive alpha only among stand-alone onshore funds, precisely where capital levels and flows are most constrained. We interpret the positive alpha observed only in stand-alone onshore funds as evidence of successful funds’ unwillingness to accept new money due to a fixed amount of arbitrage profits in capital markets and the fee structure of hedge funds. Goetzmann, Ingersoll, and Ross (2003)                                                         6

See Table 1 Panel C for the summary statistics of hedge fund incentive fee. 

7

find that hedge funds with superior performance do not tend to sell new shares and experience net share repurchases in contrast to similar studies in the mutual fund industry. They conjecture that the option-like incentive fee structure of hedge funds exist because managers cannot trade on superior performance to increase compensation through growth.7 Consistent with this interpretation, we find that stand-alone hedge funds are closed to new investments more often than MF hedge funds (13.04% vs. 9.28%). To our knowledge, this is the first paper that analyzes the impact of regulatory environment on share restrictions, capital flow, and performance of onshore and offshore hedge funds by separating stand-alone funds from master-feeder funds. We are also the first to analyze the risk-adjusted performance of hedge funds after controlling all three types of liquidity risk: market liquidity (a systematic risk factor as in Sadka (2010)), asset liquidity (idiosyncratic illiquidity that causes serial correlation as in Getmansky, Lo, and Makarov (2004)), and share liquidity (share restrictions as in Aragon (2007)).8 The rest of this paper is organized as follows: Section 2 develops testable hypotheses, and Section 3 describes data and summary statistics. Section 4 presents empirical results and Section 5 provides robustness check. Finally, Section 6 concludes.

                                                        7

Strasburg and Gongloff (2010) report an example of a successful hedge fund manager who has been closed to new investments for more than a decade, and returns capital to investors because he believes that an “enormous amount of capital” has a negative impact on his long-term track record. 8 We distinguish asset liquidity from share liquidity, while previous research assumes a positive relation between share restrictions and asset illiquidity and uses share restrictions such as a lockup provision as a proxy for asset illiquidity. See Section 3.2, 3.3, and 4.4 for details on how we control market liquidity, asset liquidity, and share liquidity when measuring risk-adjusted performance of hedge funds.

8

2. Background and Testable Hypotheses

In this section we discuss the regulatory framework motivating our empirical work. We also discuss related literature to develop the specific testable hypotheses that are tested later in the paper.

2.1. Regulatory Framework Although a typical public investment company in the United States is required to be registered with the US Securities and Exchange Commission (SEC), hedge funds are largely exempted from the Investment Company Act of 1940 mainly due to the legal structure and the special types of investors. Normally, US-registered hedge funds are organized as partnerships with limited and qualified investors from wealthy individuals or institutions, who are deemed to be financially sophisticated and have little need for protection by government regulations. In particular, under Sections 3(c)1 and 3(c)7 of the Investment Company Act of 1940, 3(c)1 hedge funds with no more than 100 accredited investors and 3(c)7 hedge funds with unlimited number of “qualified purchasers” are exempted, respectively9. Hedge funds are also exempted from registering their securities by the Securities Act of 1933 if they do not seek funding from the general public. Finally, the Securities Exchange Act of 1934 requires hedge funds with more than 499 investors to report on a quarterly basis so a 3(c)7 fund can effectively avoid quarterly reporting by having 499 investors or less.

                                                        9

An accredited investor must meet at least one of the following requirements: 1) earn an individual income of more than $200,000 per year, or a joint income with spouse of $300,000, in each of the last two years and expect to reasonably maintain the same level of income, 2) have a net worth exceeding $1 million, either individually or jointly with spouse, and 3) be a general partner, executive officer, director or a related combination thereof for the issuer of a security being offered. A qualified investor is an individual or institution with at least $5 million in assets to invest with.

9

Therefore, the US-registered onshore funds in general do not seek funds from the general public and have less than 100 or 499 investors depending on whether they are a 3(c)1 or 3(c)7 fund. Under these restrictions, managers will impose stricter share restrictions than their offshore counterparts in order to keep investor capital and deter redemptions. Based on the above observations, we develop the following testable hypotheses to capture the fundamental difference between onshore funds and offshore funds.

2.2. Testable Hypotheses

The first hypothesis relies on the idea that equity capital represents an important source of funding for hedge fund managers, and their capital withdrawals can reduce fund performance when investment opportunities are favorable. For example, Shleifer and Vishny (1997) argue that when arbitrage requires capital, arbitrageurs can become most constrained when they have the best opportunities, because fund investors are informationally disadvantaged about the fund’s strategy. Hedge funds can manage their risk by limiting the size of redemptions and/or replacing exiting capital. However, since onshore funds face regulatory constraints on the number of accounts and the ability to issue new shares, we expect these funds to more often impose redemption restrictions, like lockups and notice periods, to reduce the size of redemptions.10 H1 [Share Restriction Hypothesis]: Due to regulatory requirements on the number of investors and the way to offer fund shares, onshore funds impose greater share restrictions than their offshore counterparts: they impose longer lockup periods, less frequent redemption and                                                         10

The role of lockups in managing redemptions in open-ended funds has been examined by Chordia (1996), Nanda, Narayanan, and Warther (2000), Lerner and Schoar (2004), and Aragon (2007). The prevailing mechanism is that lockups allow funds to attract investors that are less likely to experience a liquidity shock, and therefore have a longer investment horizon. A similar mechanism drives the equilibrium described by Amihud and Mendelson (1986).

10

subscription periods, larger initial investment requirements, in order to attract long-term investors, keep investor capital and deter redemptions.

The second hypothesis relies on the idea that a manager’s portfolio choice depends on the fund’s exposure to investor liquidity shock. Edelen (1999) shows that investor liquidity shocks can increase non-discretionary trading costs and reduce fund profitability. These nondiscretionary trading costs are likely to be larger for funds holding less liquid assets. Hedge funds can manage this risk by holding more liquid assets and/or reversing these shocks by raising new capital. However, since offshore funds face regulatory constraints on the number of accounts and the ability to issue new shares, we expect these funds to hold more liquid assets, so that redemptions are met at lower cost. Hence, we have the following testable hypothesis: H2 [Asset Illiquidity Hypothesis]: Due to regulatory requirements on the number of investors and the way to offer fund shares, onshore funds manage assets with greater liquidity and lower liquidity risk than offshore funds.

The third hypothesis is based on the idea that investor search costs are an important determinant of fund flow. In particular, Sirri and Tufano (1998) and Huang, Wei, and Yan (2007) report empirical evidence suggesting that greater marketing efforts by mutual funds deliver strong performance-flow sensitivity, and especially for high performing funds. Likewise, we expect that the regulatory constraints on onshore funds’ marketing efforts will lead to lower performance sensitivity as compared to offshore funds. This leads to our third hypothesis: H3 [Fund Flow Hypothesis]: Due to restrictions on both the number of investors and the public issuance of fund shares, investors face greater search costs when investing with onshore funds.

11

As a result, the flow-performance relation is less sensitive for onshore fund than offshore funds, especially for high-performing funds.

Our final hypothesis relies on the idea that fund performance is unlikely to persist when investor capital is sensitive to past performance. For example, BG (2004) and PS (2010) assume that managers face decreasing returns to scale from asset management, and that investors will supply capital competitively as they learn from past performance. In the Berk and Green equilibrium, therefore, abnormal fund returns are eliminated by investors as they direct more capital to superior managers. A key assumption in their model is that there is perfect competition among investors for manager skill. However, as we argue above, the regulatory constraints faced by onshore funds make it difficult to advertise performance to investors, leading to a reduced flow/performance relation. Therefore, we expect greater performance among onshore funds, where investor capital does not chase away performance. Meanwhile, offshore funds (who presumably are not subject to the same regulatory requirements) will show lower risk-adjusted performance. Therefore, we have the following testable hypothesis for fund performance: H4 [Fund Performance Hypothesis]: Offshore funds conform more closely to the BG (2004) assumption that investor capital is supplied competitively and without search costs, due to the absence of regulation on the number of investors, accredited investor requirement, and strict marketing prohibitions. As a result, we expect lower risk-adjusted performance among offshore funds because fund profits are chased away by unrestricted capital flows.

12

3. Data and Summary Statistics In this section we discuss the hedge fund data used in the empirical analysis. We also discuss the variables used to benchmark fund returns and how to measure market liquidity risk and asset illiquidity. Finally, we present summary statistics for the key variables of the sample.

3.1. TASS Database

We obtain individual hedge fund data from Lipper TASS, which is one of the largest used in academic research. The database provides monthly net-of-fee returns, assets under management (AUM), and other fund characteristics such as investment style, legal structure, domicile country where a fund is registered, management company, fee structure, and share restriction provisions. The investment styles of hedge funds we analyze are convertible arbitrage, dedicated short bias, equity market neutral, emerging markets, event driven, fixed income arbitrage, global macro, long-short equity hedge, and multi-strategy funds. We exclude managed futures as this style focuses specifically on futures which are different from hedge funds or funds of hedge funds.11 We do not include funds that report i) returns in a foreign currency, instead of US dollars, ii) quarterly (instead of monthly) returns, or iii) gross return (instead of net-of-fee returns). We include both live funds and defunct funds to avoid survivorship bias. As TASS does not retain data on defunct funds before 1994, our sample period starts in January 1994 and ends in November 2005. To mitigate backfill bias, we delete the first two years of return data.12 Based on the above criteria, we have 3,573 funds left in our sample for our portfolio level analysis,                                                         11

Liang (2004) indicates that there are differences between managed futures and hedge funds in terms of performance, risk, attrition, and correlation structures with major market indices. 12 Another method to reduce backfill bias is to use the date on which each fund was added to the database and to delete the returns before the date if the date is available. As a robustness check, we repeat our tests using both methods of adjusting backfill bias and find that our main results do not change.

13

among which 2,111 are live funds and 1,462 are defunct funds. There are 2,049 (1,200 live and 849 defunct) offshore funds and 1,524 (911 live and 613 defunct) onshore funds.13 Later, when we estimate alpha to evaluate the risk-adjusted performance at the individual fund level, we further require a minimum return history of twenty-four months. There are 2,233 funds (1,230 offshore and 1,003 onshore) that meet these criteria. We further partition the sample of offshore and onshore funds depending on whether they are part of a master-feeder (MF) structure. A MF structure is devised for hedge fund managers who wish to market a fund to both onshore and offshore investors. Instead of managing two different portfolios side-by-side, a MF manager usually sets up one “master” company and two “feeders”: one feeder is a limited partnership for onshore investors and the other feeder is an offshore corporation for offshore investors. The sole investment of these two feeders is an ownership interest in the master, which is typically an offshore limited liability company. The actual portfolio investment is made at the master company level.14 We use the management company information provided by TASS to define master-feeder (MF) funds and stand-alone (SA) funds. If there are both onshore and offshore funds with the same investment style managed by the same company, we classify them as MF funds and the other funds are SA funds. Among the 1,230 offshore funds with at least twenty four monthly returns, 341 are MF funds and 889 are SA funds. Among the 1,003 onshore funds, 389 are MF funds and 614 are SA funds. We recognize that there can be cases where a MF onshore (offshore) fund is misclassified as a SA fund if the counterpart offshore (onshore) fund in the MF

                                                        13

We classify a hedge fund as an onshore fund if it is registered in the United States. Offshore funds are registered in the Cayman Islands, British Virgin Islands, Bermuda, Bahamas, Guernsey, Netherlands Antilles, Mauritius, Liechtenstein, or Saint Kitts and Nevis. 14 See Buscema (1996) and McCrary (2002) for detailed description on the master-feeder structure of hedge funds.

14

structure does not report to the database. However, the bias is against finding any difference between MF funds and SA funds. Finally, when analyzing the relation between share illiquidity and asset illiquidity, we use five share restriction variables available in TASS: lockup period, redemption frequency (RF), redemption notice period (RNP), subscription frequency (SF), and the minimum investment amount (MinInv). The lockup period specifies a time interval during which a new investor is not allowed to redeem the shares of a fund without a penalty. As in previous research, we use a lockup dummy variable instead of the lockup period because the lockup period is clustered around zero (for 68% of the funds) and twelve month (for 26% of the funds), and have little variability. RF and SF show how often the fund processes the redemption and subscription requests from investors. RNP is the amount of advance notice that investors should give a fund manager before cashing in the fund shares. Note that lockup is a one-time restriction applied only to new investors while other variables are rolling restrictions applied to all investors.

3.2. Benchmarking Hedge Fund Returns To estimate alpha, we use the seven risk factors of Fung and Hsieh (2004). That is, i) the excess return on the S&P 500 index, ii) the size factor as in Fama and French (1993), iii) the monthly change in the 10-year treasury constant maturity yield, iv) the change in the credit spread of the Moody’s Baa bond over the 10-year treasury bond, v), vi), and vii) the excess returns on portfolios of look back straddle options on currencies, commodities, and bonds as in Fung and Hsieh (2001).15

                                                        15

We thank Kenneth French and David Hsieh for providing downloadable data on their websites. The size factor was obtained from http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html, The trend-following factors were downloaded from http://faculty.fuqua.duke.edu/~dah7/HFRFData.htm.

15

In addition to the seven risk factors, we include another systematic risk factor in time series regressions to obtain market-liquidity adjusted alpha. The added factor is the liquidity risk factor as in Pastor and Stambaugh (2003, hereafter PS) and Sadka (2006). Sadka (2010) examines the cross-section of hedge fund returns and finds that funds that significantly load on market liquidity risk outperform low-loading funds and this outperformance is independent of share restrictions.16

3.3. Measuring Asset Illiquidity and Market Liquidity Risk

Lo (2001) and Getmansky, Lo, and Makarov (2004, hereafter GLM) show that hedge fund returns are often serially correlated, and the most likely explanation is illiquid exposure although return smoothing is also a possibility. Therefore, Lo (2001), GLM (2004), and Khandani and Lo (2007) suggest using the first-order serial correlation coefficient (ρ) of a fund’s returns as a measure of asset illiquidity.

GLM (2004) also suggests another measure of asset illiquidity by distinguishing between a fund’s reported returns and economic returns. The idea is that the reported returns of illiquid portfolios only partially reflect the true economic returns contemporaneously but the economic returns are incorporated to reported returns eventually. That is, the reported return in period t ( Rt0 ) satisfies the following equations: Rt0   0 Rt  1 Rt 1   2 Rt  2     k Rt  k

(1)

0  i  1 for all i  0,1, 2,  k , and

(2)

 0  1   2     k 1

(3)

                                                        16

The results reported in this paper are based on Sadka’s (2006) liquidity factor, but we also tested the PS (2003) liquidity factor as a robustness check and found similar results.

16

where Rt is the fund’s true economic return in period t. As in GLM (2004), we assume that the demeaned economic returns are mean-zero, normal random variables, and use the previous sixty-month return history of a fund to estimate the parameters in Equation (1) by maximum likelihood estimation. 0 represents the fraction of a fund’s economic return that is simultaneously incorporated in its reported return. Hence low 0 means a more illiquid portfolio. Therefore, we call  = 1- 0 the GLM (2004) measure of asset illiquidity. We test both the GLM (2004) illiquidity measure () and the first-order serial correlation coefficient () as a measure of asset illiquidity, and find similar results.17 Note that both higher  and higher  indicate higher asset illiquidity. In addition to the idiosyncratic GLM illiquidity measure, we estimate the systematic market-liquidity-risk beta for each fund where market liquidity is defined as in PS (2003) and Sadka (2006). They show that market liquidity is a priced risk factor in the cross-section of stock returns. The PS (2003) liquidity factor is based on the principle that order flows induce greater return reversal when liquidity is low, while Sadka’s (2006) liquidity factor is based on the permanent variable component of the intraday price impact of stock trades. The impact of market liquidity on hedge funds has also been confirmed by the Long-term Capital Management debacle in 1998 and the recent liquidity crisis happened from 2007 to 2009.

                                                        17

The only difference is that the requirement of sixty-month return history to estimate parameters in Equation (1) reduces our sample size from 2,233 (1,230 offshore and 1,003 onshore) to 1,400 (740 offshore and 660 onshore). Therefore, the results reported in this paper are using the first-order serial correlation coefficient () as a measure of asset illiquidity to maintain a large sample size.

 

17

Recently, Sadka (2010) applies the market liquidity factor to the cross-section of hedge fund returns, and reports that funds with significant loading on market liquidity risk outperform the low-loading funds by 8 percent on an annual basis over the period 1994-2007. Using a twoway sorting approach based on market liquidity loading and share liquidity, he also shows that this outperformance is independent of share restrictions. To build on these findings, we include a market liquidity factor in the time-series regression to estimate alpha in addition to the seven risk factors of Fung and Hsieh (2004).

3.4 Summary Statistics

Table 1 presents summary statistics for our main sample of hedge funds. The majority (2,049) of the 3,573 hedge funds in our sample are offshore funds. We also tabulate the number of offshore funds by domicile country. The vast majority of offshore funds are domiciled in the Cayman Islands (1,193), followed by the British Virgin Islands (351) and Bermuda (103). The table also shows that offshore funds have more assets under management on average compared to onshore funds. This is consistent with our hypothesis that the registration exemption requirements faced by onshore funds, like the limit on the number of investor accounts and restrictions on public advertising, restrict capital flows to onshore funds.

Finally, Table 1 shows that 87 percent of onshore funds are limited partnerships (LPs) while only 4.3 percent of offshore funds endorse such a structure. In the case of offshore funds, the most frequently observed legal structure is open ended investment company (47.5 percent), but the proportion varies across locations (42.1 percent for Cayman Island and 70.9 percent for Bahamas). Corporate structures are less frequent among onshore funds because corporations are subject to double taxation in the United States.

18

4. Analysis and Results

In this section we report results from testing the four hypotheses developed in Section 2.

4.1. Share Restriction Hypothesis

In Table 2 we report results from comparing the usage of share restrictions between onshore and offshore funds. Our findings indicate that onshore funds impose tighter share restrictions than offshore funds. For example, the average lockup period of onshore funds (5.7 months) is more than twice the lockup period of offshore funds (2.4 months). On average, onshore funds require higher minimum investment amount, and have longer redemption, redemption notice, and subscription periods than those of offshore funds. All the differences are statistically significant at the 1 percent level with the t-statistics ranging from 3.64 to 19.61. Overall, our results support the share restriction hypothesis according to which, due to regulatory requirements on the number of investors and the way to offer fund shares, onshore funds are more likely to use lockups to attract long-term investors, retain investor capital and deter redemptions.

4.2. Asset Illiquidity Hypothesis

In Table 3 we report the results from comparing market liquidity risk and asset illiquidity between onshore and offshore funds. We find that offshore funds hold assets with both greater illiquidity and higher market liquidity risk. For example, the Sadka (2006) liquidity beta is 28.47 for offshore funds on average, as compared to 9.32 for onshore funds. A similar, though statistically insignificant, pattern is also found when measuring liquidity risk using the PS (2003) factor. In addition, we find that offshore funds hold assets with greater illiquidity, measured

19

either using the first-order return autocorrelation or the GLM measure. Taken together, we interpret these results as support for the asset illiquidity hypothesis-offshore funds can manage asset illiquidity better than their onshore counterparts through easy equity issuance. In contrast, onshore funds cannot as easily reverse investor outflows by raising new equity, thereby making it comparatively more costly for these funds to hold illiquid assets. This finding is consistent with Brunnermeier and Pedersen (2009) who argue that a shock to a hedge fund’s capital can lower the liquidity of assets that it trades.

The above results show that the organizational form and asset illiquidity of offshore funds differs significantly from that of onshore funds. Specifically, offshore funds have lower share restrictions and hold asset with greater illiquidity and market liquidity risk. This is somewhat surprising in light of prior research showing a positive relation between share restrictions and asset illiquidity. For example, Aragon (2007) reports a positive relation and argues that share restrictions allow funds to efficiently manage illiquid assets. By separating offshore funds from onshore funds, we extend the literature and suggest that a fund’s domicile (i.e., onshore or offshore) is an important control variable in testing the relation between share restrictions and asset illiquidity. In Table 4 we present results from a multivariate logit analysis of the fund’s decision to use a lockup provision. We estimate the model separately for the onshore and offshore fund subsamples. As explanatory variables we include asset illiquidity (ρ), illiquidity factor loading (βSadka), fund age, and a limited partnership (LP) dummy. We also include the investment style dummies to adjust for the style effect. To make parameter estimates comparable, we normalize continuous variables to have a mean of zero and a standard deviation of one across all funds.

20

Consistent with the prior literature (see Aragon (2007)), we find that both asset illiquidity and the age of a fund are related to lockups. In particular, lockups are less common among funds with greater asset liquidity as the need has been reduced. Younger funds are more likely to impose lockup as managers are eager to cumulate assets. This result holds for both onshore and offshore funds, although the magnitude of the coefficient is stronger for offshore funds. The partnership structure restricts managers from getting more investors so they need a lockup provision to prevent money withdraw; this is true especially for onshore funds.

4.3 Fund Flow Hypothesis

To examine the flow-performance relationship, we use a methodology similar to Sirri and Tufano (1998) and Fung et al. (2008). Specifically, we measure capital flows into a fund during a year by using the growth rate of net new money, which is defined as Flowi,t= (TNAi,t-TNAi,t-1 (1+ Ri,t))/ TNAi,t-1. TNAi,t is fund i’s total net assets at the end of year t, and Ri,t is the fund’s return during the year. That is, Flowi,t represents the percentage growth of a fund during the year in excess of the growth that would have occurred if no new money had flowed in. As in previous research, the top and bottom 1 percent of the flows are winsorized to mitigate the effect of outliers.

We run a piecewise linear regression of investor flows on relative performance variables Lowt, Midt, and Hight. These variables are defined using a fractional rank (FRANK) that represents a fund’s percentile performance relative to other funds in the same investment style during the same period. FRANK ranges from 0 to 1. The bottom performance quintile (Lowt) is defined as Min (0.2, FRANKt-1), the middle three performance quintiles are combined into one

21

group labeled as Midt, defined as Min (0.6, FRANKt-1 – Lowt), and the highest performance quintile (Hight) is defined as Min (0.2, FRANKt-1 – Lowt – Midt). For example, if a fund’s FRANK was 0.82 last year, its Lowt is 0.2, Midt is 0.6, and Hight is 0.02. We include the logarithm of the size in the previous period (Log (TNAt-1)) as a control variable because an equal dollar flow has a larger percentage impact on smaller funds. We also include the standard deviation of fund returns, flows to the investment style, share restrictions, fees, high-water mark (HWM) dummy, leverage dummy, and open-to-public dummy variables as control variables. We conduct the regressions annually from 1994-2005, then calculate the Fama-MacBeth (1973) coefficients as well as the t-statistics. In Table 5 we report the results from estimating the piecewise linear regression of investor flows on relative performance. We find that the sensitivity of net fund flows to past performance, especially strong performance, is greater for offshore funds. For example, a change in rank from 80th percentile to 90th percentile increases inflows by 11.5% for offshore funds, and the increase is significant at the 5 percent level. However, the same jump in rank in onshore funds does not lead to a statistically significant increase in inflows. This result supports our key identifying assumption that offshore funds find it easier to raise capital from investors who chase past performance. For example, offshore funds face fewer restrictions on advertising than onshore funds, thereby attract more capital after generating superior performance, not mentioning onshore funds can close for new money when reaching their design capacity. In addition, it is easier for offshore funds to attract new investors because these funds do not face a limit on the number of accounts. This result helps motivate our prediction that, due to decreasing returns to scale, offshore funds have lower performance than onshore funds.

22

4.4. Fund Performance Hypothesis

In this subsection we compare the performance of onshore funds with offshore funds. In Table 6 we compare the monthly returns, Sharpe Ratios, and Fung and Hsieh (2004) seven-factor alphas for the two fund subgroups. In Panel A we find that offshore funds generally have worse performance than onshore funds regardless of the performance measure. For example, the sevenfactor alpha of onshore funds is 0.77%, higher than 0.55% for offshore funds. The 0.22% spread between the two fund groups is statistically significant at the 1% level and transfers to an annual performance difference of 2.64%.

A similar pattern holds when we measure performance using raw returns or the Sharpe Ratio. In addition, the higher average performance among onshore funds is very stable across different style categories. For example, the onshore/offshore alpha spread is positive for all styles except for convertible arbitrage funds, where the difference is insignificant from zero. This suggests that the performance difference is not driven by omitted factors from our benchmark model. In Panel B of Table 6 we compare performance after further subdividing the sample funds depending on whether a fund has a lockup provision. The results again show a positive differential between onshore and offshore funds, although this difference is larger for funds without a lockup (0.18%) than those with a lockup (0.07%).18 Note that in Panel B of Table 6, the difference between the lockup funds and non-lockup funds is 0.26% for onshore funds and 0.37% for offshore funds. Therefore, offshore funds earn a larger lockup premium than onshore

                                                        18

The t-statistic for the performance difference between onshore lockup funds and offshore lockup funds is 1.19 and it is 4.38 for the performance difference between onshore non-lockup funds and offshore non-lockup funds.

23

funds, consistent with the results in Table 3 that offshore funds have greater asset illiquidity and liquidity risk. Our earlier results highlight significant differences between onshore and offshore funds in the degree of share restrictions, asset liquidity, and market liquidity risk. Prior research shows that these variables have significant explanatory power on hedge fund returns.19 Therefore, we also compare performance of onshore and offshore funds after further controlling for differences in market liquidity, asset liquidity, and share liquidity. We estimate this “liquidity-adjusted alpha” using a two-step procedure. First, we use a time-series regression with the excess return of a hedge fund as the dependent variable, and the seven factors of Fung and Hsieh (2004), plus the market liquidity factor of Sadka (2006) as explanatory variables to obtain alphas. Then, we use a cross-sectional regression of the estimated eight-factor model alphas ( ˆ i ) on asset illiquidity () and share restriction variables as in Equation (4).

ˆ i   0  1   i   1  Lockupi   2  RNP   3  MinInvi   4  RNPi 2   5  MinInvi2   6  Lockup  RNPi  style dummies  

( 4)

where RNP2 and MinInv2 are included to test whether the relation between alpha and share restriction is linear, and LockupRNP is used to test whether an extra period of redemption notice matters for investors who have agreed for a lockup period. The results are reported in Table 7. Overall our main finding here is consistent with Table 6-onshore funds perform better than offshore funds. For example, we estimate a liquidityadjusted alpha of 0.67% for onshore funds with a significance level at 1%, but for offshore funds                                                         19

A positive relation between fund returns and share restrictions, like lockups and notice periods, is reported by Liang (1999), Bali, Gokcan and Liang (2007), Liang and Park (2007), and Aragon (2007). More recently, Sadka (2010) finds that funds that significantly load on liquidity risk subsequently outperform low-loading funds by about 8% annually over the period 1994-2007.

24

this number is only 0.08% and statistically insignificant. In addition, the positive coefficients on the share restrictions variables are consistent with prior findings on share illiquidity premium in hedge fund returns. However, the higher returns attributable to fund lockups, or, “lockup premium”, has greater economic and statistical significance for offshore funds (0.37%) than onshore funds (0.15%). We further subdivide our sample funds depending on whether they are part of a masterfeeder (MF) or stand-alone (SA) structure. Strikingly, Table 7 shows that the greater performance of onshore funds is driven entirely by the positive liquidity-adjusted alphas of the SA onshore funds. In contrast, onshore funds that are part of a MF structure have similar performance to their offshore counterparts. We interpret this finding as the decision of hedge fund managers who observe a limited amount of arbitrage profits in capital markets not to raise additional capital because they want to avoid the diluting effect on the outstanding claims. For example, when hedge fund managers compete to attract capital, “hot hands” can raise enough capital without going offshore and choose to remain as SA onshore funds because size matters less under the incentive fee structure of hedge funds. In contrast, managers with lower performance are more likely to use MF structure to attract more capital seeking for a size-related compensation.20 As advertising constraint makes flow to onshore funds become less sensitive to superior performance, investors cannot chase away performance through the channel of decreasing returns to scale in the SA onshore funds. Incentive fees in hedge funds may have motivated SA onshore fund managers (who started their career with wealthy US individual investors and built a superior track record) not to accept more                                                         20

This finding is consistent with Nohel, Wang, and Zheng (2010), who examine side-by-side management of hedge funds and mutual funds. They find that hedge fund managers who accept capital from both hedge fund investors and mutual fund investors underperform their peers. 

25

capital from offshore investors in order to prevent the decreasing returns to scale problem of their arbitrage strategies. Meanwhile, the performance of MF onshore funds is lower because the fund manager (who is seeking higher compensation through growth) circumvents the regulation on the number of investors by accepting additional capital from offshore investors through a MF structure.

5. Robustness

We find that the liquidity-adjusted alpha of SA onshore funds is positive and significantly different from zero. To show that this alpha is not generated by pure luck, we apply the robust bootstrap methodology developed by Kosowski et al. (2006) to our analysis as follows.21

First, we conduct the time-series regression of the excess return on the eight risk factors, and save the parameter estimates and the time-series of residuals for each fund. Then, we do the cross-sectional regression of the estimated eight-factor model intercept on asset illiquidity and share illiquidity as in Equation (4), and save the t-statistic of the intercept ( tˆ ). Second, we draw 143 months with replacement from January 1994 to November 2005. Then, for each fund, we create the bootstrapped excess return (BER) observations using the parameter estimates and residuals of the time-series regression without including the estimated intercept. That is, the bootstrapped returns have a true alpha of zero. Call the new excess return data generated by this procedure as bootstrap sample BER 1. Note that this procedure is designed to keep the higher order correlation between the regressors and residuals as well as the crosssectional correlation of the residuals across funds in the original return data.

                                                        21

See Kosowski et al. (2007) and Fung et al. (2008) for details on bootstrap analysis of hedge fund performance.

26

Third, we delete funds that have less than twenty-four return observations from BER 1. Then, using BER1, we repeat the time-series and the cross-sectional regressions in the first step and find the t-statistic of the intercept from the cross-sectional regression ( tˆ1 ). Finally, we repeat the second and third steps 1,000 times to generate BER 1,…, BER 1,000, and { tˆ1 ,…, tˆ1, 000 }. We compare tˆ from the original return data with the 99th percentile of the bootstrapped empirical distribution of the t-statistics of alpha to examine whether the probability of obtaining a t-statistic higher than or equal to tˆ is lower than 1 percent. We find that the t-statistics of alphas ( tˆ ) of all funds, onshore funds, and the SA onshore funds are all higher than the 99th percentiles of the bootstrapped empirical distributions. This confirms that the liquidity-adjusted alpha of hedge funds reported in this paper cannot be attributed to pure luck.

6. Conclusion

This paper analyzes the impact of regulation on share restrictions, capital flow, and performance of hedge funds. By separating i) offshore funds from onshore funds, ii) stand-alone funds from master-feeder funds, and iii) illiquidity of hedge fund assets from illiquidity of hedge fund shares, we make three contributions to the hedge fund literature.

First, we find that the positive relation between share restrictions and asset illiquidity reported in previous research should be reexamined. By distinguishing onshore funds from offshore funds, we show that onshore funds on average impose stronger share restrictions than offshore funds but invest in more liquid assets in order to better manage equity funding risk caused by restrictions on advertising and the number of investors.

27

Second, we find that onshore and offshore hedge funds have a different flowperformance relationship and this difference affects risk-adjusted performance. We show that capital flow to onshore funds is less sensitive to past performance due to prohibition on advertising, and high capital flow to offshore funds chase away alpha due to decreasing returns to scale. Finally, we analyze the risk-adjusted performance of hedge funds after controlling all three types of liquidity risk (market liquidity, asset liquidity, and share liquidity), and find that liquidity-adjusted alpha is positive and significant only in stand-alone onshore funds. Masterfeeder funds that accept capital from both onshore and offshore investors underperform standalone funds. This finding is consistent with previous research that shows increased capital flows to actively managed funds lead to decreasing returns to scale as documented by Berk and Green (2004) as well as Pastor and Stambaugh (2010).

28

References Ackermann, C., McEnally, R., and Ravenscraft, D., 1999, The Performance of Hedge funds: Risk, Return, and Incentives, Journal of Finance 54, 833-874. Agarwal, V. and Naik, N., 2005, Hedge Funds, Foundations and Trends in Finance 1, 103-170. Agarwal, V. and Naik, N., 2004, Risks and Portfolio Decisions Involving Hedge Funds, Review of Financial Studies 17, 63-98. Aggarwal, R., and Jorion, P., 2010, The Performance of Emerging Hedge Funds and Managers, Journal of Financial Economics 96, 238-256. Amihud, Y., 2002, Illiquidity and Stock Returns: Cross-Section and Time-Series Effects, Journal of Financial Markets 5, 31-56. Amihud, Y., and Mendelson, H., 1986, Asset Pricing and the Bid-Ask Spread, Journal of Financial Economics 17, 223-249. Amihud, Y., and Mendelson, H., 1991, Liquidity, Maturity and the Yields on US Treasury Securities, Journal of Finance 46, 1411-1425. Aragon, George, 2007, Share Restriction and Asset Pricing: Evidence from the Hedge Fund Industry, Journal of Financial Economics 83, 33-58. Asness, G., Krail, R., and Liew, J., 2001, Do Hedge Funds Hedge?, Journal of Portfolio Management 28, 6-19. Bali, T. G., Gokcan, S., and Liang, B., 2007, Value at Risk and the Cross-Section of Hedge Fund Returns, Journal of Banking and Finance 31, 1135 – 1166. Berk, J., and Green, R., 2004, Mutual Fund Flows and Performance in Rational Markets, Journal of Political Economy 112, 1269- 1295. Brown, S., Goetzmann, W., and Ibbotson, R., 1999, Offshore Hedge Funds: Survival and Performance, Journal of Business 72, 91-117. Brown, S., Goetzmann, W., and Liang, B., 2004, Fees on Fees in Funds of Funds, Journal of Investment Management 2, 39-56. Brown, S., Goetzmann, W., Ibbotson, R., and Ross, S., 1992, Survivorship Bias in Performance Studies, The Review of Financial Studies 5, 553-580. Brown, S., Goetzmann, W., and Park, J., 2001, Careers and survival: Competition and risk in the hedge fund and CTA industry, Journal of Finance 56, 1869-1886.

29

Brunnermeier, M. and Pedersen, L., 2009, Market Liquidity and Funding Liquidity, Review of Financial Studies 22, 2201–2238. Buscema, J, 1996, Master-Feeder Funds: Structuring for Maximum Efficiency, Hedge Fund News, February 1996 Issue, International Fund Administration (IFA). Chordia, T., 1996, The Structure of Mutual Fund Charges, Journal of Financial Economics 41, 3-39. Chow, G., 1960, Test of Equality between Sets of Coefficients in Two Linear Regressions Econometrica 28, 591-605. Christoffersen, P., and Goncalves, S., 2005, Estimation Risk in Financial Risk Management, Journal of Risk 7, 1-28. Cumming, D., 2007, A Law and Finance Analysis of Hedge Funds, Working Paper, Rensselaer Polytechnic Institute (RPI). Ding, B., Getmansky, M., Liang, B., and Wermers, R., 2009, Investor Flows and Share Restrictions in the Hedge Fund Industry, Working Paper, University of Massachusetts, Amherst. Edelen, R., 1999, Investor Flows and the Assessed Performance of Open-end Mutual Funds, Journal of Financial Economics 53, 439-466. Fama, E. and French, K., 1992, The Cross-Section of Expected Stock Returns, Journal of Finance 47, 427-465. Fama, E. and French, K., 1993, Common Risk Factors in the Returns on Stocks and Bonds, Journal of Financial Economics 33, 3-56. Fama, E. and MacBeth, J., 1973, Risk, Return and Equilibrium: Empirical Tests’, Journal of Political Economy, 38, 607-636. Fung, W., and Hsieh, D., 1997, Empirical Characteristics of Dynamic Trading Strategies: The Case of Hedge Funds, Review of Financial Studies 10, 275-302. Fung, W., and Hsieh, D., 2000a, Performance Characteristics of Hedge Funds and CTA Funds: Natural versus Spurious Biases, Journal of Financial and Quantitative Analysis 35, 291-307. Fung, W., and Hsieh, D., 2000b, Measuring the Market Impact of Hedge Funds, Journal of Empirical Finance 7, 1-36. Fung, W., and Hsieh, D., 2001, The Risk in Hedge Fund Strategies: Theory and Evidence from Trend Followers,” Review of Financial Studies 14, 313-341.

30

Fung, W., and Hsieh, D., 2002a, Benchmarks of Hedge Fund Performance: Information Content and Measurement Biases, Financial Analyst Journal 58, 22-34. Fung, W., and Hsieh, D., 2002b, Asset-based Style Factors for Hedge Funds, Financial Analyst Journal 58, 16-27. Fung, W., and Hsieh, D., 2002c, The Risk in Fixed-Income Hedge Fund Styles, Journal of Fixed Income 12, 6-27. Fung, W., and Hsieh, D., 2004, Hedge Fund Benchmarks: A Risk Based Approach, Financial Analyst Journal 60, 65-80. Fung, W., Hsieh, D., Naik N., and Ramadorai T., 2008, Hedge Funds: Performance, Risk and Capital Formation, Journal of Finance 63, 1777-1803. Gentilini, A. and Tantiyakul, P, 2009, Lipper TASS Asset Flows Report: Hedge Funds - Second Quarter 2009, Thomson Reuters. Getmansky, M., Lo A., and Makarov I., 2004, An Econometric Model of Serial Correlation and Illiquidity in Hedge Funds Returns, Journal of Financial Economics 74, 529-610. Goetzmann, W., Ingersoll, J., and Ross, S., 2003, High-water Marks and Hedge Fund Management Contracts’, Journal of Finance 58, 1685-1717. Goetzmann, W., Ingersoll, J., Spiegel, M., and Welch, I., 2007, Portfolio Performance Manipulation and Manipulation-Proof Performance Measures, Review of Financial Studies 20, 1505-1546. Huang, J., Wei, K., and Yan, H., 2007, Participation Costs and the Sensitivity of Fund Flows to Past Performance, Journal of Finance 62, 1273-1311. Ibbotson, R., and Chen, P., 2006, The A, B, Cs of Hedge Funds: Alphas, Betas, and Costs, Working Paper, Yale School of Management. Jagannathan, R., Malakhov, A., and Novikov, D., 2010, Do Hot Hands Exist among Hedge Fund Managers? An Empirical Evaluation, Journal of Finance 65, 217-255. Khandani, A., and Lo, A., 2007, What Happened to the Quants in August 2007? Journal of Investment Management 5, 5-54. Kosowski, R., Naik, N., and Teo, M., 2007, Do Hedge Funds Deliver Alpha? Journal of Financial Economics 84, 229-264.

31

Kosowski, R., Timmerman, A., White, H., and Wermers, R., 2006, Can Mutual Fund Stars Really Pick Stocks? New Evidence from a Bootstrap Experiment, Journal of Finance 61, 22512595. Lerner, J., Schoar, A., 2004, The Illiquidity Puzzle: Theory and Evidence from Private Equity, Journal of Financial Economics 72, 3–40. LePree, S., 2008, Taxation of US Tax-exempt Entities’ Offshore Hedge Fund Investments – Application of the Section 514 Debt-financed Rules to Leveraged Hedge Funds and Derivatives and the Case for Equalization, Holland & Knight, LLP. Liang, B., 1999, On the Performance of Hedge Funds, Financial Analysts Journal 55, 72-85. Liang, B., 2000, Hedge fund: the Living and the Dead, Journal of Financial and Quantitative Analysis 35, 309-326. Liang, B., 2004, Alternative Investments: CTAs, Hedge Funds and Funds-of-Funds, Journal of Investment Management 2, 76-93. Liang, B., and Park, H., 2007, Risk Measures for Hedge Funds: A Cross-Sectional Approach, European Financial Management 13, 317-354. Lo, A., 2001, Risk Management for Hedge Funds: Introduction and Overview, Financial Analysts Journal 57, 16-33. McCrary, S. A., 2002, How to Create and Manage a Hedge Fund: A Professional’s Guide, John Wiley & Sons, Inc. Nanda, V., Narayanan, M., and Warther, V., 2000, Liquidity, Investment Ability, and Mutual Fund Structure, Journal of Financial Economics 57, 417-443. Nohel, T., Wang, Z. J., and Zheng, L., 2010, Side-by-Side Management of Hedge Funds and Mutual Funds, Review of Financial Studies 23, 2342-2373. Pastor, L., and Stambaugh, R., 2003, Liquidity Risk and Expected Stock Returns, Journal of Political Economy 111, 642-685. Pastor, L., and Stambaugh, R., 2010, On the Size of the Active Management Industry, University of Chicago working paper. Sadka, R., 2006, The Role of Liquidity Risk, Journal of Financial Economics 80, 309-349. Sadka, R., 2010, Liquidity Risk and the Cross-Section of Hedge-Fund Returns, Journal of Financial Economics 98, 54-71.

32

Scholes, M. and Williams, J., 1977, Estimating Betas from Nonsynchronous Data, Journal of Financial Economics 5, 309-327. Shleifer, A. and Vishny, R., 1997, The Limits of Arbitrage, Journal of Finance 52, 35-55. Sirri, E. and Tufano, P., 1998, Costly Search and Mutual Fund Flows, Journal of Finance 53, 1589-1622. Strasburg, J. and Gongloff, M., 2010, Soros’s Ex-Manager to Close Shop, The Wall Street Journal, August 19, C1 and C10.

33

Table 1 Legal Structure of Hedge Funds by Domicile Country This table compares offshore hedge funds with onshore funds in terms of fund size and legal structure. The data is from the TASS database, and the sample period is from January 1994 to November 2005. AUM_Total is the total assets under management (AUM) as of November 2005. Legal Structure (%) No. of Funds

AUM _Average ($mm)

AUM _Total ($ billion)

Onshore Funds

1,524

83.6

Offshore Funds

2,049

Limited Partnership

Limited Liability Company

Open Ended Investment Company

Exempted Company

Open Ended Mutual Fund

Others

71.1

87.0

9.8

1.4

0.0

0.0

1.8

170.7

193.7

4.3

6.6

47.5

12.7

5.2

23.6

1,193

157.5

101.7

4.8

7.9

42.1

21.4

2.9

20.9

British Virgin Islands

351

201.5

40.9

3.4

4.3

67.6

0.0

2.8

21.9

Bermuda

216

169.8

22.7

3.7

9.7

46.8

1.9

27.2

10.7

Bahamas

103

137.4

3.8

3.9

1.0

70.9

0.0

1.0

23.2

Others

186

216.3

24.5

3.8

2.7

31.7

1.1

1.1

59.6

All Funds

3,573

132.5

264.8

39.6

8.0

27.8

7.3

3.1

14.2

Domicile Country

Cayman Islands

 

 

34

Table 2 Share Restrictions and Other Characteristics of Hedge Funds by Investment Style and by Domicile Country This table compares offshore funds and onshore funds in terms of share restrictions, size, age, and fees. Reported numbers are sample averages across all funds within the same investment style. Panel A shows share restriction variables; Panel B displays fund size and age; Panel C presents fees and other fund characteristics. ***, **, and * denote that the difference in the characteristics of offshore funds and onshore funds is significantly different from zero at the 1%, 5% and 10% level, respectively.

Panel A: Share Restrictions Investment Style

Convertible Arbitrage Dedicated Short Seller Emerging Markets Equity Market Neutral Event Driven Fixed Income Arbitrage Global Macro Long/Short Equity Hedge Multi-Strategy All Funds

Lock-Up Period (months)

Minimum Investment ($mm)

Redemption Notice Period (Days)

Redemption Frequency (Days)

Subscription Frequency (Days)

Onshore

Offshore

t-stat

Onshore

Offshore

t-stat

Onshore

Offshore

t-stat

0.63

40.9

42.2

-0.30

82.2

66.5

1.49

37.8

31.3

2.35**

0.63

-0.58

27.2

21.8

0.84

132.7

44.1

3.03***

60.1

32.1

3.21***

0.56

0.38

2.43**

33.9

27.0

1.61

92.8

50.9

3.83***

40.5

31.0

2.27**

5.35***

0.71

0.68

0.30

32.7

25.2

2.88***

78.5

38.8

6.95***

47.5

30.7

5.25***

4.54

2.21**

1.31

1.11

1.27

51.8

48.9

0.97

160.2

90.1

6.55***

51.8

35.5

4.66***

4.58

1.98

3.56***

1.18

1.07

0.53

44.2

31.4

3.31***

92.3

65.3

2.41**

40.7

32.4

2.69**

3.91

1.19

3.16***

0.85

0.76

0.47

27.5

18.1

3.25***

63.1

38.0

4.52***

37.9

30.8

2.24**

6.40

2.36

13.55***

0.78

0.59

4.08***

36.8

29.4

6.36***

117.0

51.6

17.24***

50.3

34.0

10.47***

4.59

2.43

2.44**

1.16

1.08

0.20

39.6

40.4

-0.16

80.6

61.9

39.4

31.4

2.74***

5.69

2.39

15.60***

0.90

0.73

3.64***

38.7

31.4

8.12***

112.4

56.4

47.9

32.9

13.95***

Onshore

Offshore

t-stat

Onshore

Offshore

3.43

3.38

0.06

1.16

0.97

4.58

4.59

0.00

0.57

4.00

1.68

2.21**

4.51

1.15

6.21

t-stat

 

35

1.90 19.61***

Panel B: Size and Age Number of Funds

AUM _Total ($ billion)

AUM _Average ($mm)

Age (months)

Investment Style Onshore

Offshore

Onshore

Offshore

Convertible Arbitrage Dedicated Short Seller Emerging Markets Equity Market Neutral

70 19 45 124

112 17 255 164

1.4 0.8 1.5 5.5

Event Driven

219

256

78

149

Global Macro Long/Short Equity Hedge Multi-Strategy

69 819 81

All Funds

1,524

Fixed Income Arbitrage

Onshore

Offshore

t-statistic

Onshore

Offshore

6.1 0.6 32.2 8.7

52.4 49.6 47.7 56.0

216.7 56.4 160.2 124.5

-2.80*** -0.16 -4.45*** -1.71*

74.3 103.2 61.8 52.0

62.8 75.0 65.6 46.0

1.69* 1.40 -0.50 1.37

16.0

34.0

191.6

257.1

-1.43

72.6

56.8

3.18***

4.8

18.4

100.0

200.3

-3.26***

52.8

50.2

0.46

198 793 105

1.9 30.9 8.4

15.2 62.1 16.3

57.1 61.2 123.3

148.7 132.2 315.7

-2.37** -4.97*** -2.50**

61.4 64.3 55.3

52.6 52.9 52.5

1.32 5.06*** 0.44

2,049

71.1

193.7

83.6

170.7

-7.51***

64.2

54.9

5.93***

t-statistic

Panel C: Fees and Other Characteristics Investment Style

Management Fee (%) Onshore

Offshore

Incentive Fee (%) t-stat

High-Water Mark (%)

Onshore

Offshore

t-stat

Onshore

Offshore

Leveraged (%) Onshore

Offshore

Convertible Arbitrage Dedicated Short Seller Emerging Markets Equity Market Neutral Event Driven Fixed Income Arbitrage Global Macro Long/Short Equity Multi-Strategy

1.22

1.34

-1.64*

17.9

19.0

-1.38

65.7

65.2

70.0

81.2

1.17

1.35

-1.11

19.4

18.8

0.34

52.6

52.9

36.8

35.3

All Funds

1.45

1.53

-1.04

17.9

17.5

0.42

44.4

43.9

57.8

57.6

1.21

1.43

-4.28***

18.9

19.9

-1.57

66.1

72.0

52.4

57.3

1.26

1.42

-3.81***

18.7

19.1

-0.82

67.6

68.4

51.6

59.0

1.24

1.36

-1.57

19.1

19.8

-0.92

79.5

61.1

78.2

77.9 80.3

1.47

1.57

-0.83

18.5

17.8

0.76

68.1

48.0

72.5

1.15

1.32

-8.16***

19.2

19.1

0.49

71.7

62.9

61.9

64.1

1.44

1.47

-0.35

19.8

18.1

2.27**

76.5

67.6

61.7

59.0

1.22

1.40

-10.70***

19.0

18.8

1.01

69.8

60.7

60.9

65.1

 

36

Table 3 Illiquidity of Assets in Hedge Funds by Investment Style and by Domicile Country This table compares offshore funds and onshore funds in terms of market illiquidity beta and asset illiquidity measures. βPS represents for the Pastor-Stambaugh (2003) liquidity beta, βSadka is the Sadka (2006) liquidity beta, Autocorrelation () is the first-order autocorrelation in fund returns, and GLM Measure () is the liquidity measure similar to Getmansky, Lo, and Makarov (2004). ***, **, and * denote that the difference in the characteristics of offshore funds and onshore funds is significantly different from zero at the 1%, 5% and 10% level, respectively. Market Illiquidity Investment Style

Convertible Arbitrage Dedicated Short Seller Emerging Markets Equity Market Neutral

Asset Illiquidity

PS (2003)

SADKA (2006)

Factor Loading (βPS)

Factor Loading (βSadka)

Autocorrelation ()

GLM Measure () Onshore

Offshore

t-stat

0.16

0.31

0.33

-0.69

0.06

-0.16

-0.04

0.11

-2.49**

0.19

0.17

0.62

0.12

0.15

-0.75

0.55

0.09

0.06

0.75

0.13

0.08

0.81

Onshore

Offshore

t-stat

Onshore

Offshore

t-stat

Onshore

Offshore

-2.47

-3.30

0.75

-25.19

-31.74

0.38

0.39

0.39

-2.40

-5.60

0.75

-60.84

-4.46

-0.70

0.05

8.03

8.50

-0.14

21.43

110.63

-1.20

0.58

0.70

-0.12

36.86

25.12

t-stat

Event Driven

1.11

0.51

0.56

53.96

41.62

0.75

0.22

0.22

0.01

0.21

0.22

-0.20

Fixed Income Arbitrage

0.06

0.45

-0.26

-25.48

17.50

-1.91*

0.28

0.25

0.76

0.27

0.22

1.11

Global Macro

1.84

2.08

-0.11

18.79

46.63

-0.93

0.01

0.08

-2.58***

-0.06

-0.01

-0.85

Long/Short Equity Hedge

2.62

2.79

-0.21

3.31

9.32

-0.43

0.10

0.12

-1.42

0.08

0.13

-3.02***

Multi-Strategy

3.04

-0.97

1.45

-22.82

-24.25

0.03

0.16

0.18

-0.57

0.12

0.19

-1.47

All Funds

1.92

2.30

-0.71

9.32

28.47

-2.17**

0.14

0.16

-2.07**

0.11

0.15

-2.88***

37

Table 4 Logit Analysis of the Lockup Provision This table reports the parameter estimates and Pseudo-R2 from the logistic regression of lockup provision. Independent variables are asset illiquidity (ρ), market illiquidity (βSadka), fund age, and a limited partnership dummy variable (LP). Investment style dummy variables are also included as control variables. To make estimates comparable, variables are normalized to have a mean of zero and a standard deviation of one across all funds. ***, **, and * denote statistical significance at the 1%, 5% and 10% level, respectively. All Funds

Onshore

Offshore

Panel A. Univariate Analysis with Asset Illiquidity Asset Illiquidity (ρ) 2 Pseudo-R (%)

0.21*** 3.18

0.13* 2.68

0.42*** 5.54

Panel B. Univariate Analysis with Market Illiquidity Market Illiquidity (βSadka) 2 Pseudo-R (%)

0.00 2.46

0.10 2.59

-0.08 3.55

0.18* 0.10 -0.63*** 0.45* 6.18

0.46*** -0.04 -0.40** 0.17 5.84

Panel C. Multivariate Analysis Asset Illiquidity (ρ) Market Illiquidity (βSadka) Age LP 2 Pseudo-R (%)

0.30*** 0.01 -0.53*** 1.10*** 10.13

   

38

Table 5 The Effect of Performance on Capital Flows: Onshore vs. Offshore Hedge Funds This table presents the effect of relative performance on capital flows to onshore and offshore funds. Similar to Sirri and Tufano (1998) and Fung et al. (2008), capital flows are measured by using the growth rate of net new money, which is defined as Flowi,t = (TNAi,t - TNAi,t-1*(1+ Ri,t))/ TNAi,t-1. TNAi,t is fund i’s total net assets at time t, and Ri,t is the fund’s return over the prior period. The coefficient estimates are presented from the piece-wise linear regression of investor flows on relative performance variables, Low, Mid, and High, which are defined using a fractional rank (FRANK) that represents a fund’s percentile performance relative to other funds in the same investment style in the same period. FRANK ranges from 0 to 1. The bottom performance quintile (Lowt) is defined as Min (0.2, FRANKt-1), the middle three performance quintiles are combined into one grouping labeled as Midt, defined as Min (0.6, FRANKt-1 – Lowt), and the highest performance quintile (Hight) is defined as Min (0.2, FRANKt-1 – Lowt – Midt). For example, if a fund’s FRANK was 0.98 last year, its Lowt is 0.2, its Midt is 0.6, and its Hight is 0.18. As control variables, risk, size, flows to the investment style, share restrictions, fees, high-water mark (HWM) dummy, leverage dummy, and open-to-public dummy variables are included. The regressions are run annually, and standard errors and t-statistics are calculated from the annual results as in Fama and MacBeth (1973). t-statistics are given in parenthesis below the coefficient estimates. ***, **, and * denote statistical significance at the 1%, 5% and 10% level, respectively. ALL

 

Onshore

2.36 (6.50)

***

2.81 (5.26)

Relative Performance Bottom Performance Quintile (Low) nd th 2 -4 Performance Quintiles (Mid)

0.82 (2.93) 0.67 (5.80)

**

0.29 (0.42) 0.72 (7.64)

Top Performance Quintile (High)

0.86 (1.91)

Intercept

Std. dev. Monthly Returns Log (TNAt-1) Flow to the Investment Style High Water Mark (HWM) Lockup Period Redemption Frequency Subscription Frequency Management Fee Incentive Fee Open to Public Leveraged 2

Adj-R (%)

-0.04 (-4.66) -0.14 (-8.56) 0.18 (1.44) 0.20 (6.34) 0.00 (0.72) 0.00 (0.59) -0.06 (-3.65) 0.04 (1.60) 0.01 (1.83) 0.12 (1.67) 0.05 (1.03)

***

Offshore ***

***

0.74 (1.14)

* ***

-0.03 (-2.39) -0.17 (-7.52) -0.01 (-0.96) 0.19 (6.49) 0.01 (1.36) 0.01 (2.74) -0.03 (-1.95) 0.02 (0.58) 0.01 (1.60) 0.01 (0.20) -0.01 (-0.15)

***

***

***

*

10.89

14.51

39

2.66 (8.58) 0.88 (1.32) 0.73 (3.79) 1.15 (2.74)

** ***

***

** **

-0.05 (-7.37) -0.16 (-9.07) 0.01 (2.54) 0.21 (5.25) 0.00 (0.87) 0.06 (2.09) -0.08 (-1.87) -0.03 (-0.89) 0.00 (0.39) 0.09 (1.18) 0.05 (1.11) 11.17

***

***

** *** *** ** ***

* *

  Table 6 Performance and Risk for Onshore and Offshore Hedge Funds Panel A compares offshore hedge funds with onshore hedge funds in terms of performance and risk. Risk-adjusted performance is measured by the Sharpe ratio and the seven-factor model alpha as in Fung and Hsieh (2004). Reported numbers are sample averages. Panel B lists a two-way sorting result: i) onshore vs. offshore, ii) lock-up vs. non-lockup, and compares the risk-adjusted performance. ***, **, and * denote that the difference in the characteristics of offshore funds and onshore funds is significantly different from zero at the 1%, 5% and 10% level, respectively.

Panel A: Performance of Onshore vs. Offshore Hedge Funds by Investment Styles Average Return (%)

Standard Deviation (%)

Sharpe Ratio

Seven-Factor Model Alpha

Investment Style Onshore

Offshore

t-statistic

Onshore

Offshore

t-statistic

Onshore

Offshore

t-statistic

Onshore

Offshore

t-statistic

Convertible Arbitrage Dedicated Short Seller Emerging Markets Equity Market Neutral

0.50 0.48 0.84 0.51

0.26 -0.42 0.57 0.50

1.92* 2.29** 0.61 0.05

1.96 6.51 6.26 2.12

1.86 6.26 6.26 1.83

0.40 0.20 -0.01 1.53

0.21 0.07 0.07 0.21

0.04 -0.14 0.18 0.12

0.95 2.30** -1.22 0.88

0.52 0.52 0.62 0.48

0.54 0.44 0.60 0.33

-0.19 0.33 0.07 1.74*

Event Driven

0.80

0.57

2.71***

2.76

1.91

3.95***

0.27

0.27

-0.01

0.83

0.57

3.09***

Fixed Income Arbitrage

0.56

0.36

2.38**

1.97

2.00

-0.11

0.55

0.34

1.32

0.69

0.40

2.99***

Global Macro Long/Short Equity Multi-Strategy

0.08 0.69 1.02

0.26 0.54 0.68

-0.59 1.67* 1.74*

4.73 4.80 2.91

3.68 4.35 2.78

1.98** 2.51** 0.28

-0.01 0.16 0.34

-0.02 0.10 0.28

0.10 1.97** 0.68

0.61 0.84 0.99

0.31 0.65 0.47

2.06** 3.48*** 2.80***

0.67

0.49

2.73***

3.97

3.69

2.42**

0.20

0.14

2.41**

0.77

0.55

6.16***

All Funds

 

40

Panel B: Performance of Onshore vs. Offshore Hedge Funds by Lockup Periods Lockup

Non-lockup Difference (αlockup- αnon-lockup)

Number (percent)

Seven-Factor Model Alpha

Number (percent)

Seven-Factor Model Alpha

All Funds

687 (30.8%)

0.90

1,546 (69.2%)

0.54

0.36***

Onshore Funds

443 (44.2%)

0.92

560 (55.8%)

0.66

0.26***

Offshore Funds

244 (19.8%)

0.85

986 (80.2%)

0.48

0.37***

Difference (αonshore - αoffshore)

0.07

0.18***

 

41

Table 7 Liquidity-adjusted Alpha This table presents the parameter estimates and adjusted-R2s from cross-sectional regressions of ^

 i   0   1   i   1  Lockupi   2  RNP   3  MinInvi   4  RNPi 2   5  MinInvi2   6  Lockup  RNPi  style dummies   i

The independent variable is a fund’s alpha from the time-series regression of the fund’s excess return on Fung and ^ Hsieh (2004) factors and Sadka (2006) liquidity factor (  i ), and explanatory variables are asset illiquidity (), share restrictions like a lockup dummy variable (Lockup), redemption notice period (RNP) in months, and minimum investment amount (MinInv) in millions of dollars., and style dummy variables. t-statistics are reported in parentheses. ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively. Intercept All Funds Onshore SA MF Offshore SA MF

0.35 (4.09)*** 0.67 (5.48)*** 0.94 *** (5.65) 0.11 (0.63) 0.08 (0.66) 0.10 (0.63) 0.07 (0.40)

Asset Illiquidity 0.26 (2.58)*** 0.20 (1.33) 0.08 (0.35) 0.62 (3.42)*** 0.36 (2.58)*** 0.28 (1.59) 0.61 (2.94)***

Lockup

RNP

MinInv

RNP2

MinInv2

0.28 (3.44)*** 0.15 (1.49) 0.17 (1.16) 0.05 (0.34) 0.37 (2.80)*** 0.51 (2.77)*** 0.13 (0.82)

0.26 (5.10)*** 0.21 (2.95)*** 0.18 (1.55) 0.25 (3.14)*** 0.21 (2.88)*** 0.06 (0.72) 0.13 (1.29)

0.00 (0.00) 0.01 (0.10) 0.04 (0.73) -0.03 (-1.54) -0.02 (-0.45) 0.06 (0.96) -0.02 (-0.43)

-0.03 (-1.86)* -0.04 (-2.13)** -0.01 (-0.25) -0.04 (-2.35)*** -0.01 (-0.32) 0.03 (0.86) -0.04 (-1.27)

0.00 (0.11) -0.01 (-0.07) -0.01 (-0.89) 0.01 (1.17) 0.01 (0.52) -0.01 (-1.02) 0.01 (1.01)

 

42

LockupRNP -0.07 (-1.34) -0.04 (-0.58) 0.01 (0.10) -0.01 (-0.19) -0.11 (-1.31) -0.16 (-1.43) -0.01 (-0.09)

Adj-R2 4.83 3.31 4.59 6.85 4.13 3.87 6.05