Advertisement

Journal of Derivatives & Hedge Funds

, Volume 18, Issue 1, pp 42–52 | Cite as

The problem with hedge fund fees

  • Rob BrownEmail author
Original Article
  • 602 Downloads

Abstract

The structure of hedge fund fees is meaningfully flawed. These imperfections are so great that institutional investors will force substantive change over the coming years. The problem arises when the basis on which fees are calculated serves to break the required clean connection between value and price. Unfortunately, the structure of hedge fund fees when superimposed on how hedge fund portfolios are managed results in a significant breakdown between the value received and the price paid. Moreover, this failure has in large measure remained dormant as a result of the abnormally low returns generated by traditional asset categories. However, eventually, normalcy will return, and equities, credit and commodities will once again generate their traditional returns, and at that time hedge fund clients will revolt – insisting on a reconfiguration of the hedge fund fee structure. This article explores the nature and severity of this issue.

Keywords

hedge fund fee structure incentive alpha beta alignment 

INTRODUCTION

A common hedge fund fee structure consists of a fixed base management fee (for example, 1.5 per cent) and an incentive fee (for example, 20 per cent) applied to the performance net of the base management fee.1 Given the frequency with which hedge funds generate double- or even triple-digit returns, application of the incentive fee percentage can generate a sizeable payment back to the hedge fund manager. As a result, some have suggested that hedge fund fees are excessive. In response, I observe the old marketing adage ‘price is only relevant in the absence of value’, and suggest that this rule applies just as much to hedge funds as it does to any other good or service. There have existed in the past and continue to exist today hedge fund managers who deliver high levels of alpha with clearly superior consistency (Jagannathan et al 2010, Ammann et al 2010). Such managers fully justify their high performance-based fees.

However, by suggesting that relatively high fees can be justified by relatively high value-added, I do not mean to imply that the structure of hedge funds fees is without its challenges. In fact, I would suggest that hedge fund fees are meaningfully flawed and that the nature of these imperfections is of sufficient magnitude such that institutional investors will force substantive change over the coming years.

The problem arises, I argue, when the basis on which fees are calculated serves to break or otherwise confound the required tight, clean connection between value and price. Unfortunately, the structure of hedge fund fees when superimposed on how hedge fund portfolios are actually managed results in a significant breakdown between the value received and the price paid. Moreover, I argue that this failure has in large measure remained dormant as a result of the abnormally low returns generated by traditional asset categories over the past 10 or more years. However, eventually, normalcy will return, and equities, credit and commodities will once again generate returns more in line with historic norms, and I conjecture that at that time hedge fund clients will revolt – insisting on a reconfiguration of the hedge fund fee structure. This article explores the nature and severity of this issue.

STRUCTURAL NATURE OF THE PROBLEM

A hedge fund's before-fee return can be decomposed into three component parts, that is, the risk-free interest rate, that portion attributable to systematic factor exposures (betas) and that which is attributable to idiosyncratic bets (alpha) (Brown 2011a, Brown 2011b). This is no different from traditional long-only managers. The structural nature of the problem with hedge fund fees is illustrated by the simple example within Figure 1.
Figure 1

The structural problem with hedge fund fees.

In this simple example, the hedge fund generates a before-fee return of 20 per cent during the year. This return can be decomposed into three parts, that is, 1 per cent attributable to the risk-free interest rate, 13 per cent attributable to the fund's exposure to systematic factor risks such as equities or credit, and 6 per cent attributable to idiosyncratic bets (alpha). This last component, the 6 per cent, is the value-added provided by the fund manager. The annual base management fee of 1.5 per cent is subtracted from this return, leaving the client with an 18.5 per cent return before the incentive fee. In this example, the hedge fund imposes a 20 per cent haircut, equivalent to a 3.7 per cent incentive fee (20 per cent × 18.5 per cent=3.7 per cent).

As a result, the total fee paid back to the hedge fund manager is 5.2 per cent or 87 per cent of the pre-fee alpha generated by the fund. At the end of the year, the spoils are divvied up with 87 per cent going to the hedge fund manager and 13 per cent going to the client who provided the original at-risk capital – a rather poor alignment of interests.

Consider the implications of this fee structure as highlighted through this simple example. One is paying an exorbitant price for simple beta. In this instance, the marginal cost for beta is 20 per cent of the annual reward to that beta. This compares unfavorably with the current market price for simple factor exposures as identified by the expense ratio for institutional index funds, institutional Exchange Traded Funds (ETFs), or the pricing on futures or swaps. Essentially, the institutional price for beta is near-zero.

Herein lays the confounding element of existing hedge fund fee structures – that serve to break the linkage between price and value. Ostensibly, one invests in hedge funds in order to experience the benefits of their alpha. The fund's embedded beta is just along for the ride and provides no inherent value, as the market price for beta is near-zero. Moreover, alpha is really quite finite and constrained. It is unlikely that even a successful hedge fund could generate double-digit alpha (after stripping out embedded betas) year after year. Far more likely is a fairly modest single-digit alpha (Fung and Hsieh 2004, Ibbotson et al 2011). However, as the current structure for hedge fund fees has the incentive percentage also applying to the return on beta, clients are at significant risk that their incentive fee could swamp their realized alpha – as in the example above. This is really no different from the problem that would be created if traditional vanilla long-only managers started charging incentive fees on their total return – it is just in their case the outcome would be somewhat worse as they tend to generate slightly lower levels of alpha and embed slightly higher levels of beta. Thus, this begs the question as to what is the magnitude of this problem.

HEDGE FUND EMBEDDED EXPOSURE TO SYSTEMATIC BETAS

One of the most popular composite hedge fund indices is provided by Hedge Fund Research Inc. (HFRI), that is, the HFRI Fund Weighted Composite Index. I took the 168 monthly returns for this index ending on 31 October 2010, grossed them back up for fees (added the fees back in), and regressed them (net of the risk free rate) on alternate systematic risk factors. The results of this multiple OLS regression are provided in Table 1.
Table 1

Hedge fund exposure to systematic betasa

Regression results for HFRI Fund Weighted Composite Index

Effective exposure

Coefficient (beta value)

Factor exposure

t-statistic

P-value

Short volatility

−0.0041

Proportionate change in volatility of daily Russell 3000 returns over rolling 6-month windows

5.3

0.0000

Long smidcap growth stocks

0.1741

MSCI North America SMID Growth

15.3

0.0000

Long emerging country stocks

0.0805

MSCI BRIC Standard (large and midcap)

8.9

0.0000

Leverage

−1.8146

Euro Cash LIBOR 1 month

3.7

0.0003

Monthly before-fee alpha (regression intercept)=62.6 bp.

R2=0.896.

Adjusted R2=0.894.

aBefore running the regressions, both the fixed base management fee and the incentive fee were added back into the return series provided by Hedge Fund Research Inc. or Hedge Fund Network. Similarly, before running the regressions, the risk-free interest rate was subtracted from all asset class returns and from the grossed-up hedge fund return series. A stepwise regression technique was used to identify the systematic factor exposures. Note that it is possible that composite hedge fund return series provided by Hedge Fund Research Inc. and by Hedge Fund Network may not be representative of a typical institutional investor's diversified portfolio of hedge funds. Such differences could result from differentiated selection and composition issues. It is possible that higher or lower levels of embedded factor exposures and consequent R2 measures could result. However, based on data gathered from several institutional investors with long-standing, well-diversified hedge fund portfolios based in the United States, we believe that the results presented here are representative and directionally correct.

Statistics based on the 168 monthly returns ending on 31 October 2010.

This analysis suggests that four systematic factors explain almost 90 per cent of the variation in monthly returns for this composite hedge fund performance measure. Ranked in order of their t-statistics, the most important factor, carrying a 17.41 per cent weighting, is smidcap growth stocks as measured by the MSCI North America SMID Growth Index. The second most important factor, carrying an 8.05 per cent weighting, is large and midcap emerging country stocks as measured by the MSCI BRIC Standard (large and midcap) Index. Filling out the third and fourth positions are short volatility with a t-statistic of 5.3 and leverage (borrowing) with a t-statistic of 3.7, respectively.2 The P-values in the far right-hand column serve to further emphasize the statistical significance of these four systematic betas. We can conclude from these results that hedge funds have, on average, delivered a 25.46 per cent exposure to equities over the 14 years examined. Finally, the monthly before-fee alpha (regression intercept) comes in at 62.6 bp per month and accounts for only 10 per cent of the variation in monthly returns – a modest proportion.

These data allow us to estimate the impact of hedge fund fee structures on large, well-diversified pools of individual hedge funds – what one is likely to find at a pension or endowment. In other words, we can estimate the fee sensitivity to beta reward. Table 2 provides a summary of these results.
Table 2

Portion of alpha paid back to the hedge fund manager in feesa

Results for HFRI Fund Weighted Composite Index

Annual return

Monthly returns

Give up

Equities (%)

Alpha (before-fee) (bp)

Base management fee (bp)

Incentive fee (bp)

Total fee paid to hedge fund manager (bp)

Portion of alpha paid to hedge fund manager (%)

−10

62.6

12.4

0

12.4

19.8

  0

62.6

12.4

0

12.4

19.8

 10

62.6

12.4

3.7

16.1

25.7

 20

62.6

12.4

7.4

19.8

31.7

 30

62.6

12.4

10.9

23.3

37.2

 35

62.6

12.4

12.5

24.9

39.8

aStatistics appearing in Table 2 assume that monthly alpha remains constant from month to month. Similarly, volatility and the 1-month Euro Cash LIBOR rate remain constant over time – only the return to equities varies. The percentages in the far right-hand column show the proportion of each month's alpha that must be returned to the hedge fund manager as fees.

Statistics based on the 168 monthly returns ending on 31 October 2010.

Assumes constant levels of monthly alpha, Russell 3000 volatility and cash return – equal to the average levels as determined by the earlier OLS regression.

To create Table 2, all components were held constant with the exception of the return to equities (smidcap growth and emerging markets), which were allowed to vary. Alpha (before-fee) was assumed to remain constant each month at its historic average of 62.6 bp. Volatility and leverage (Euro cash) were held constant at their average levels, experienced over the 14 years of data on which the Table 1 regression was based. The first column identifies the assumed annual equity return, ranging from a low of −10 per cent to a high of +35 per cent. The next four columns show the resulting monthly alphas and fees generated. Finally, the far right-hand column identifies the portion of the alpha that clients will need to return to the hedge fund manager as fees (both base management and incentive). For equity market returns at or below zero, clients pay 19.8 per cent of the resulting alpha back to the hedge fund manager in fees. But as equity market returns increase, so does the portion of generated alpha, which must be returned to the manager. By the time equity market returns reach 35 per cent, almost 40 per cent of generated alpha must be returned as fees. Unfortunately, examining this structural problem through the lens of such broad-based averages as represented by the HFRI Fund Weighted Composite Index tends to hide the true nature of the problem.

PROBLEM IS FAR WORSE FOR CERTAIN STYLES

Given the structural nature of the problem, one can anticipate that the breakdown between value received and price paid would be significantly more troublesome for certain hedge fund styles. Those styles with the greatest embedded beta and the least alpha would be the most at risk. The notion that hedge funds are absolute return vehicles delivering pure alpha and no beta has no basis in reality. Their embedded factor exposures may be a bit more complex, are frequently multidimensional, but are no less pertinent or potent than for traditional long-only managers. A classic example is provided by one of the most popular and easily understood hedge fund sectors, that of the long/short equity funds. Table 3 shows the regression results for the Hedge Fund Network (HFN) Long/Short Equity Index.
Table 3

Long/short hedge fund exposure to systematic betasa

Regression results for HFN Long/Short Equity Index

Effective exposure

Coefficient (beta value)

Factor exposure

t-statistic

P–value

Short volatility

−0.0035

Proportionate change in volatility of daily Russell 3000 returns over rolling 5-month windows

3.0

0.0030

Long smidcap growth stocks

0.2273

MSCI North America SMID Growth

12.2

0.0000

Long US stocks

0.1541

AMEX Composite Index

5.3

0.0000

Leverage

−3.4152

Japan Cash LIBOR 1 month

6.6

0.0000

Monthly before-fee alpha (regression intercept)=21.8 bp.

R2=0.855.

Adjusted R2=0.852.

aThe measure of volatility was based on daily total returns for the Russell 3000 Index and compared the percentage change in the standard deviation of these daily returns during the most recent month versus the prior 4 months. The measure of leverage was represented by borrowing at the 1-month Japanese Cash LIBOR rate.

Statistics based on the 168 monthly returns ending on 31 October 2010.

Over the 14 years ending in 2010, four systematic factors explained over 85 per cent of the variation in monthly returns. Alpha explains less than 15 per cent. The single most important factor is that of smidcap growth stocks (MSCI North America SMID Growth Index), with a weighting of 22.73 per cent. This analysis also shows a weighting of 15.41 per cent to the AMEX Composite Index, resulting in a total combined equity exposure of 38.14 per cent. The factor exposures are rounded out with a short exposure to volatility and a long exposure to leverage (by effectively borrowing at the Japanese 1-month LIBOR rate). The regression's intercept term identifies before-fee alpha, which came in at 21.8 bp per month. Relative to the HFRI Fund Weighted Composite Index (intended to represent all hedge fund styles), the long/short equity sector apparently delivers higher embedded equity exposure (more beta) and lower alpha. Table 4 identifies the fee sensitivity of long/short to traditional rewarded betas.
Table 4

Portion of alpha paid back to the Long/Short Hedge Fund manager in feesa

Results for HFN Long/Short Equity Index

Annual return

Monthly returns

Give up

Equities (%)

Alpha (before-fee) (bp)

Base management fee (bp)

Incentive fee (bp)

Total fee paid to hedge fund manager (bp)

Portion of alpha paid to hedge fund manager (%)

−10

21.8

12.4

0

12.4

57.0

  0

21.8

12.4

0

12.4

57.0

 10

21.8

12.4

5.5

17.9

82.1

 20

21.8

12.4

11.1

23.5

107.8

 30

21.8

12.4

16.3

28.7

131.6

 35

21.8

12.4

18.7

31.1

142.9

aStatistics appearing in Table 4 assume that monthly alpha remains constant from month to month. Similarly, volatility and the 1-month Japanese Cash LIBOR rate remain constant over time – only the return to equities varies. The percentages in the far right-hand column show the proportion of each month's alpha that must be returned to the hedge fund manager as fees.

Statistics based on the 168 monthly returns ending on 31 October 2010.

Assumes constant levels of monthly alpha, Russell 3000 volatility and cash return – equal to the average levels as determined by the earlier OLS regression.

Observe how for annual equity market returns at or below zero, approximately 57 per cent of pre-fee alpha is returned to the hedge fund manager in order to pay the management fee. Unfortunately, as equity market returns increase, fees consume larger and larger portions of the generated alpha, eventually reaching 143 per cent of alpha at 35 per cent equity market returns.

This unsatisfactory relationship results from the following three factors. First, the modest size of long/short's alpha. Second, long/short's relatively greater exposure to equity markets. Third, the propensity for equity markets to deliver occasional high rates of return. For this hedge fund sector, the likelihood for fees to consume all or more than a manager's entire alpha is high enough so as to make this sector inherently unattractive. Given the limited availability of alpha within this space, clients are frequently overpaying for their long/short equity managers, thinking they are getting alpha net of fees, when in reality they are receiving nothing more than ridiculously over-priced beta. But it is wrong to paint all hedge fund styles with the same damning characterization. Table 5 examines some of the most problematic sectors and some that are relatively immune to this fee structure deficiency.
Table 5

Hedge fund styles differ greatly in their exposure to this fee problema

Hedge fund index

Fee sensitivity to traditional betas

 

(Slope between ‘Portion of alpha paid back to manager as fees’ and ‘rate of return earned by simple equity, credit, or commodity exposures’)

Problematic hedge fund styles

 HFN Value Index

40.9

 HFN Long/Short Equity Index

35

 HFRI Emerging Markets: Latin America Index

32.9

 HFN Long-Only Index

26.4

 HFRI Emerging Markets: Global Index

25.5

 HFRI EH: Quantitative Directional

20.1

 HFN Hedge Fund Aggregate Index

19

 HFN Emerging Markets Index

15

 HFRI EH: Sector – Energy/Basic Materials Index

14

 HFN North America Index

12.3

 HFN Fixed Income Arbitrage Index

12

Hedge fund styles that carry NO simple traditional embedded betas

 HFN Macro Index

2.3

 HFRI FOF: Market Defensive Index

1.9

 HFN Regulation D Index

1.8

 HFRI EH: Equity Market Neutral Index

1.7

 HFN Asset-Based Lending Index

0

 HFN Market Neutral Equity Index

0

 HFN Options Strategies Index

0

 HFN CTA/Managed Futures Index

−2.5

aThe fee sensitivity statistics appearing in the right-hand column are based on monthly data and show the slope between the proportion of monthly alpha that must be returned to the manager as fees and the monthly return on simple, traditional equity, credit and commodity exposures.

Statistics based on the 168 monthly returns ending on 31 October 2010.

Only the returns earned by simple exposure to traditional equity, credit or commodities are allowed to vary. Other factor exposures such as volatility or interest rates were held constant.

The top-half of the table identifies 11 hedge fund styles, drawn from the HFRI and the HFN databases, which are the most exposed to the problem. The right-hand column shows their fee sensitivity to traditional betas. These data indicate how much the proportion of alpha returned to the manager as fees increases with the monthly return earned by traditional, simple equity, credit or commodity exposures. For example, the HFN's Value Index of hedge funds is unusually exposed to equity market returns (that is, 55.9 per cent weighting to equities) and has earned a modest before-fee alpha (that is, 27.3 bp per month, pre-fee) over the past 14 years. For each 1 per cent increase in simple equity market monthly returns (for example, from 1 per cent to 2 per cent per month), the percentage of before-fee alpha returned to the manager as fees goes up by 40.9 per cent (in absolute terms).

The bottom-half of this table shows eight sectors that are relatively immune to the problem. These particular styles have near-zero embedded exposure to traditional betas. The HFN's Market Neutral Equity Index carries zero exposure to traditional systematic market risks and therefore avoids the structural fee problem of many other hedge fund styles (although it is quite exposed to short volatility and a value/growth tilt).

PROPOSED SOLUTION

To date, institutional investors have not yet demanded a solution to this structural failure. In large measure, their patience has resulted from abnormally low or near-zero returns for the traditional equity, credit and commodity markets over the last 14 years.

Inevitably, the returns on these common factor risks will return to normal. Under such an environment, Table 6's mixed portfolio might be expected to return close to 10 per cent instead of 3.3 per cent per annum. Those hedge fund styles carrying the highest exposure to such betas would come under the greatest client pressure to repair their deficient fee structures. In practice, only four potential solutions exist.
  • First, institutional investors could abandon those hedge fund sectors (for example, long/short equity, emerging markets, or quantitative directional) that have the greatest fee sensitivity to traditional betas. Although this solves the problem, it also abandons a number of potentially quite attractive alpha opportunities. Nevertheless, it is far more sensible to avoid such opportunities if their use requires one to assume the risk when they fail and to pay back to the manager between 50 per cent and 150 per cent of the opportunity when they succeed.

  • Second, hire an independent third party to construct an ex post performance benchmark consisting of the hedge fund manager's estimated ex post embedded traditional factor exposures. Incentive fees would then be paid only on performance net of this third-party benchmark. Although this approach is conceptually attractive, the operational and legal issues probably leave it impractical, at best.

  • Third, set up a separate account with the specific hedge fund manager and contractually specify that the account is to be managed neutral to traditional systematic factor risks. This third approach solves the structural fee problem, but imposes the inefficiencies of a separate account. Moreover, certain market sectors or strategies cannot, practically, be run in a market neutral manner – or even close.

  • Fourth, abandon incentive fees altogether, and only invest with hedge fund managers who restrict their compensation to a fixed base management fee. This last approach may be the most practical. However, some would argue that it sacrifices the highly desirable alignment of interests that result from an incentive fee. Such an argument only holds, as has been demonstrated by this article, when traditional embedded betas have been eliminated.

Table 6

Annual returns to traditional betas over the past 14 years

Traditional systematic factor exposure

Passive index

Return (annualized geometric mean return) (%)

Equities

S&P 500 Index

5.6

Credit Spreads

Barclays Capital US Credit Index (with US Treasury term structure stripped out)

−0.1

 

Barclays Capital US Corporate High-Yield Bond Index (with US Treasury term structure stripped out)

0.8

Commodities

Reuters/Jefferies – CRB Total Return Index

2.8

Mixed portfolio

33.3% Equities, 16.7% Investment grade credit spreads, 16.7% High-yield credit spreads, 33.3% Commodities

3.3

Statistics based on the 168 monthly returns ending on 31 October 2010.

Finally, appreciate that the first three solutions increase accountability and transparency. For this reason, some hedge funds may not support such solutions. It is a lot easier to make a living by charging an incentive fee based on beta returns as opposed to generating sufficient real alpha and then restricting one's fee to this single source of remuneration. I recall well the conversation I had with one genuinely talented high-yield hedge fund manager who was starting up a credit fund back in February 2009. I asked this manager how much of a bounce back in credit spreads he expected. He replied +80 per cent, that is, he was expecting the return of a passive index exposure to credit to be +80 per cent going forward. I asked the manager whether he was going to subtract this index from his performance before applying the 20 per cent incentive fee. He replied, ‘heck no, we want to get paid 20 per cent just for providing simple exposure to the markets’.

Footnotes

  1. 1.

    Hedge fund fees vary widely. Fixed base management fees could be as low as 0.5 per cent or as high as 5 per cent. However, a 1.5 per cent base management fee is not uncommon. Similarly, incentive fees vary widely. They could be as low as 10 per cent or as high as 40 per cent. Once again, an incentive fee of 20 per cent is not uncommon. The exact mechanics of how they are applied or computed varies from fund to fund.

  2. 2.

    The measure of volatility was based on daily total returns for the Russell 3000 Index and compared the percentage change in the standard deviation of these daily returns during the most recent month versus the prior 5 months. The measure of leverage was represented by borrowing at the 1-month Euro Cash LIBOR rate.

References

  1. Ammann, M., Huber, O. and Schmid, M. (2010) Hedge Fund Characteristics and Performance Persistence. Working paper, University of St. Gallen, July 2010.Google Scholar
  2. Brown, R. (2011a) Volatility and asset class exposure drive hedge fund returns. Journal of Investment Consulting 12 (1): 15–22.Google Scholar
  3. Brown, R. (2011b) Volatility – The normally dormant factor exposure for hedge funds. Journal of Derivatives and Hedge Funds 15 (February): 231–239.CrossRefGoogle Scholar
  4. Fung, W. and Hsieh, D.A. (2004) Hedge fund benchmarks: A risk-based approach. Financial Analysts Journal 60 (5): 65–80.CrossRefGoogle Scholar
  5. Ibbotson, R.G., Chen, P. and Zhu, K.X. (2011) The ABCs of hedge funds: Alphas, betas, and costs. Financial Analysts Journal 67 (1): 15–25.CrossRefGoogle Scholar
  6. Jagannathan, R., Malakhov, A. and Novikov, D. (2010) Do hot hands exist among hedge fund managers? An empirical evaluation. The Journal of Finance 65 (1): 217–255.CrossRefGoogle Scholar

Copyright information

© Palgrave Macmillan, a division of Macmillan Publishers Ltd 2011

Authors and Affiliations

  1. 1.CFA InstituteCharlottesvilleUSA

Personalised recommendations