The past thirty years have witnessed an increased separation between the ownership and the control of financial wealth. The emergence of modern portfolio theory, the increased efficiency of markets, and the growing sophistication of financial instruments have convinced many, if not most, investors to delegate the management of their portfolios to professional asset managers and their collective investment vehicles. Investment advice is now becoming a commodity.
Initially, actively managed funds took the lead and intermediated much of the consumers' investments in financial securities. However, their dismal average performance simply provided more general evidence of just how difficult it is to beat the market. It also opened the way for passive strategies and indexed funds, which were then perceived as a cost-effective way of buying equity market exposure - a strategy that made sense in an environment of rapidly rising market valuations. However, the end of the technology bubble and the subsequent bear market significantly froze the development of passive funds and provoked interest in alternative investments, such as hedge funds and private equity. Since then, the number of highly specialised, non-traditional asset management firms has been growing exponentially. Many of them are born from the ashes of the failures of mainstream fund managers.
Whatever the investment vehicle and investment strategy selected, the delegation of portfolio management activities can be seen as a particular case of the principal-agent model initially introduced in the seminal work of Jensen and Meckling (1976). Since the costs to wealth owners of monitoring those who are charged with managing their financial holdings are rather large, agency theory's most basic suggestion is that principals (investors) should compensate the agents (portfolio managers) through incentive contracts in order to align their respective interests. The nature and intensity of these incentives should depend upon a series of parameters, such as the incremental profit generated by an additional unit of effort from the manager, the precision with which investment performance and risk can be measured and monitored, the risk tolerance of the portfolio managers, and their responsiveness to incentives.
Traditional vs alternative incentives
From a theoretical perspective, incentive contracts may combine three elements, namely, a profit sharing rule (ie fee structure) to align incentives in terms of returns; a relative performance component measured against a benchmark to monitor performance, make returns comparable, and audit for common uncertainty; and checks on risk-taking, such as maximum allowable tracking error, reporting requirements and constraints on available investment choices.
How are incentive contracts implemented in practice? Surprisingly, the empirical evidence seems to suggest that traditional and alternative asset managers have taken diametrically opposed choices. Most traditional investment managers are monitored and evaluated against appropriate style benchmarks, but their compensation is not linked to their relative performance. Rather, they charge a management fee that is generally expressed as a fixed percentage of the assets of their fund. The level of this fee varies depending upon the complexity of the strategy and the asset class considered, but is typically between 1 and 3% per annum. Over recent years, asset-based fees have been subject to highly competitive pressures and declined. This is not surprising, as investors have the option of shifting their assets to another asset manager or investment vehicle as soon as they identify a better opportunity.
By contrast, alternative asset managers target an absolute performance, and charge both a management fee (typically 1% of assets under management) and an incentive fee (typically 20% of profits) based on their fund's overall performance. Anecdotal evidence suggests that for most hedge funds, the management fee is roughly equal to operating costs2 and the primary compensation is the incentive fee. In most cases, a hurdle rate of return must be exceeded by some multiple and any prior losses must be repaid before the fund manager is eligible to receive any incentive income. Over recent years, these fees have risen, particularly those of established managers who have been able to create a scarcity for their fund, which they then use to increase fees and introduce a lock-up clause3. On the contrary, with start-up funds in the course of raising capital, investors often obtain discounts on the fees in exchange for early money.
Asset-based fees vs incentive fees
One may wonder which of the two models, asset-based or incentive fees, is preferable to reduce the agency costs of portfolio management delegation. Fees uniquely based on the size of the assets under management offer a small implicit incentive to managers. As the assets in the fund grow, due to capital inflows or the appreciation of the underlying holdings, the fee collected will grow in tandem. If on the contrary, assets decrease, then the fee collected will be reduced proportionately. Several empirical academic studies have confirmed the positive relationship that exists between a fund's relative performance and subsequent inflow of new investments [Sirri and Tufano (1998)], as well as the fact that some investment funds voluntarily waive their stated fees in an attempt to boost net performance and, thereby, to attract additional assets (fee waiving). This suggests that, even though the link between performance and compensation is not direct, it nevertheless appears to be an important factor in determining fund managers' behaviour. However, we should also note that academic research has evidenced the convex nature of the relationship between fund flow and performance. That is, while superior relative performance generates an increase in the growth of assets under management and, in turn, managerial compensation, there tends to be no symmetric outflow of funds in response to poor relative performance, at least over the short term. The convex flow/performance relationship creates an incentive for fund managers to increase risk taking, especially after poor performance. Therefore, the effective incentive of an asset-based fee needs to be carefully assessed on a case-by-case basis. However, in the case of skill-based and capacity constrained strategies, asset-based fees may also create a fiduciary conflict because adding new assets can harm the interests of existing ones. Managers who have developed a strategy that works may continue selling it past the asset capacity for which it was designed, just because they are rewarded essentially on the basis of the size of their assets under management.
By contrast, performance fees seem to do a better job at aligning the interests of managers (desire for high fees) and investors (desire for high excess returns). When subject to a performance fee, a manager will sell his strategy only up to the asset capacity for which it was designed. Then, he will close his fund to additional investment, as he has stronger incentives for performance than for asset growth. Adding too much assets means being forced to put some money into second-best ideas, and these ideas do not often deliver the kind of returns desired, so asset growth is de-facto limited. At some point, managers may even have to implement net share repurchases. In this context, an increase in revenues should essentially come from improving the excess returns delivered to investors rather than by increasing the assets under management. This partially explain the relatively small size of hedge funds - about 80% of the hedge funds reporting to commercial databases manage less than US$100 million of equity capital.
However, performance fees also have their drawbacks. The most important ones are linked to their asymmetric nature, the manager participates in the upside, but not in the downside. This corresponds to a potentially perpetual call option with a path-dependent payoff - the payoff at any time depends on the high-water mark, which is related to the maximum asset value achieved. This option-like payoff structure may lead to possible adverse incentive effects, because the manager simultaneously owns the option and controls its underlying asset (the portfolio), as well as its volatility. Therefore, near the end of an evaluation period, some managers may decide to increase portfolio risk in order to increase the value of their option4. On the contrary, outperforming managers may attempt to lock-in their positive performance and dampen portfolio volatility. Alternatively, some fund managers may also try to improve the return of their portfolios by window dressing them, for example by using stale prices rather than real market values (or vice-versa) for illiquid stocks or non-traded assets around the end of an evaluation period. Between the lack of agreed-upon standards, different views about illiquid marks, and moral hazard, valuation can be akin to numerical quicksand.
It is interesting to note that although mutual funds and hedge funds seem to disagree on what is the best choice between asset-based and performance-based fees for their external investors, they both agree on their own internal compensation structures, which involve asset management firms and individual fund managers. The compensation of portfolio managers tends to be performance-based, with a fixed base salary topped by bonuses based, partially or entirely, on relative performance. This should be kept in mind, as a complete discussion on the incentives facing mutual funds must consider two layers of agency problems: the agency relationship between the fund company and the fund investors and the agency relationship between the fund company and fund management [Chevalier and Ellison (1999)].
The regulatory view
An interesting viewpoint on the question of asset management fees is that of regulators, which varies from one country to another. In the US, for example, mutual funds are registered investment companies and they are highly regulated by the SEC. The latter allows performance incentive fees and enables a fund to charge higher fees when it beats a benchmark, so long as it is willing to charge less when it fails to beat it. As one could expect, many fund managers are perfectly happy to sell their funds to the public on the grounds that it can beat the market, but despite the offer, very few of them are willing to put their own money where their mouths are and take the other side of the bet. According to the Lipper database, less than 2% of the US equity mutual funds apply a performance fee.
In Europe, a European Council Directive sets the general legal framework within which undertakings for collective investment in transferable securities (UCITS) may carry on their business. It establishes that 'the law or the fund rules must prescribe the remuneration and the expenditure which a management company is empowered to charge to a unit trust and the method of calculation of such remuneration.' Therefore, legal restrictions to the way companies managing mutual funds can be compensated for their services, if any, are to be found only at the national level. Several countries, such as Spain, France, or the UK, have left a large degree of latitude when it comes to portfolio managers deciding on the mechanism and the value of their compensation. Strikingly, in practice, even though it is legally permissible, most mutual fund companies are almost never compensated through incentive contracts. Instead, they are paid a fixed percentage of assets under management, and the incentive intensity is set to zero. At the other extreme, hedge funds and other lightly regulated private investments companies are primarily charging incentive fees.
The soft dollar arrangements
Our discussion of fees would not be complete if we did not mention soft dollar brokerage, or simply soft dollars. Soft dollar brokerage is a popular arrangement between a fund and its broker. Basically, the fund manager agrees to place a designated dollar value of trading commission business with a broker over a given period of time. In exchange for this promise, the broker provides the manager with research credits equal to some part, say 50%, of the promised commissions. Rather than rebating these credits back to investors, the manager keeps them and uses them to buy services and any of the large number of broker-approved research products (hardware, software, subscriptions, databases, etc.) supplied by third-party research vendors. The broker then pays the manager's research bill and simultaneously cancels the appropriate number of credits from the manager's soft dollar account. From a functional perspective, soft dollars are simply one form of bundling research and execution together into a single commission payment. They are unique in allowing research and execution to be provided by entirely separate firms, thereby promoting vertical disintegration of the research and execution functions.
Do soft dollars reduce or increase agency costs of delegated portfolio management? Both views are defendable. On the one hand, one may argue that soft dollars allow managers to misappropriate investor's wealth by churning their portfolios to subsidise research for which they should pay directly. This, in turn, generates various inefficiencies, such as the choice of a broker for his willingness to provide research credits rather than on expected execution quality. At the end of the day, because brokerage commissions are included in the price basis of the underlying security, investors implicitly pay the underlying research costs. Soft dollars, therefore, subsidise the manager's use of research inputs, and in some cases the existence or amount of the subsidy is unknown to investors. Thus, portfolio managers shift expenses that are normally shouldered by them onto fund shareholders. But on the other hand, one may also argue that soft dollars are aligning the interests of asset managers with those of their investors. Fund managers typically own a very small percentage of their portfolio, directly as co-investors or via an annual management fee. If managers were required to pay for all research and execution out of their own pockets, they would bear a disproportionate share of the costs of generating portfolio returns in relation to the private benefits based on their portfolio share. Seen in this light, the agency problem faced by portfolio investors is that in the absence of agreement, managers will do too little research, identify too few profitable trading opportunities, and execute too few portfolio trades. Thus, soft dollar arrangements allow investors to subsidise investment research and thereby encourage managers to do more of it, which ultimately benefits the portfolio performance.
Last but not least, soft dollars may also be unique in aligning the incentives of brokers and managers. When a broker provides soft dollar research credits to a manager, it typically does so in advance of the commission payments it expects from the manager. But the manager has no legal obligation to trade and may in particular terminate the executing broker relationship with the balance of the soft dollar account unpaid. The broker will then lose a stream of commissions that would have included a premium above the cost of providing low-quality brokerage. The threat of termination dramatically increases the expected losses to brokers who provide low-quality services, and may therefore perform an effective quality assuring function.
What makes a good performance fee?
Coming back to the main topic of our discussion, at this stage, we may wonder what the necessary characteristics of a good performance fee should be. Ideally, a performance fee should be structured to achieve five main objectives. It should reward a proficient manager for excess return earned over the measurement period, it should control portfolio risk, it should contain fair but significant consequences for manager underperformance, the performance fee agreement should be explicit in its description of the fee structure to eliminate client misunderstandings and properly frame client expectations, and it should be designed so that there is little economic incentive for the manager to grow the assets under management beyond the level at which the performance fees max out. The performance fee structure encourages investment firms to run their strategies at optimal asset levels that permit the maximisation of dollars of excess return.
Are hedge fund fees exaggerated?
Throughout the bull market of the 1990s most investors overlooked the fees charged by mutual fund companies because returns were so impressive. But times have changed. We are now in an era of difficult markets and the level of fees has come under close scrutiny. Many traditional investors who are just beginning to venture into alternative investments find their levels of fees overwhelming. If the industry standard seems to be 1% for the management fees and 20% for the performance fee, several funds among the largest and top-performing ones are far above that. For instance, Caxton Corporation which oversees more than US$10 billon charges 3% and 30%, while Renaissance's US$6.7 billion Medallion fund charges a 44% incentive fee, more than twice the industry average. Interestingly, both funds are closed to new investors and have returned money to their existing investors in 2003 in order to be able to maintain positive returns. Of course, only the best performing funds are able to dictate conditions like this. Nevertheless, the list of the top ten earners in the hedge fund industry is impressive. According to Institutional Investor, the top 10 managers earned the following sums in 2003 from a combination of their share of the fees generated by the funds they managed and the gains on their own capital in the funds: George Soros of Soros Fund Management, US$750 million; David Tepper of Appaloosa Management, US$510 million; James Simons of Renaissance Technologies, US$500 million; Edward Lampert of ESL Investments, US$420 million; Steven Cohen of SAC Capital Advisors, US$350 million; Bruce Kovner of Caxton Associates, US$350 million; Paul Tudor Jones of Tudor Investment, US$300 million; Kenneth Griffin of Citadel Investment, US$230 million; Daniel Och of OCH-Ziff Capital Management, US$150 million; and Leon Cooperman of Omega Advisors, US$145 million.
Not surprisingly, traditional investors' first reaction may be to dismiss the hedge fund industry due to excessive layers of fees. Performance fee structures with 25 and 35% carry can work out to be tremendous fees, and immediately prompt the question: 'Does the return justify the fee?' The answer is twofold. Firstly, outsiders invest in a hedge fund because they believe the manager has an expertise that they can not replicate for themselves, or that replication is too costly. This is a fact to remember when looking at hedge fund fees - you get what you pay for. Secondly, if investors achieve their objectives after expenses, the fees are justified, even if their level is an especially hard pill to swallow5. But if a fund delivers poor performance, it is not worth a low fee; in fact, it is worth no fee at all. Thus, fees should be directly related to providing what the investor wants. Consequently, when evaluating or selecting an investment fund, the fee charged should not be the unique determinant. The investment philosophy and quality and tenure of management are also important considerations, amongst others.
Why so much resistance?
So, in conclusion, are performance-based fees a desirable feature for asset management? One argument often encountered is that poorly performing managers will be paid less and, therefore, benefit the plan sponsor. On the other hand, managers who perform well will also be paid more. But since the fund earns more, this extra fee will really not cost anything at all. Perhaps, proponents contend, the carrot of higher fees and the stick of lower ones will make the managers work harder.
The objectives of performance fees are to reduce them for flat and negative performance and to reward managers for positive absolute performance. Structured properly, this makes a lot of sense for the investor and the manager if added value is properly identified. Then the client and manager are simply entering a profit-sharing plan, and profit sharing is effective in aligning incentives. The problem with performance fees starts when they are not structured properly, that is, if the client is giving a manager a fee based on something other than added value (the true alpha). This is not sustainable in the long-run. Nevertheless, many traditional managers are still reluctant to use performance fees. If the entire industry shifted to performance fees, one of the things that might happen is a reduction in fees in general. For instance, if two-thirds of the managers underperformed, they would draw one-third of their normal fees, while the one-third that outperformed would draw four-thirds of their normal fees. The industry-wide fees would then be cut by a third. Not surprisingly, at least two thirds of the asset management industry will keep fighting such a trend.
Carpenter J. N., 2000, "Does option compensation increase managerial risk appetite?" Journal of Finance, 50, 2311-2331
Chevalier J. and G. Ellison, 1997, "Risk taking by mutual funds as a response to incentives," Journal of Political Economy, 105, 1167-1200
Goetzmann, M., J. Ingersoll, and S. Ross, 1998, "High water marks," NBER Working Paper 6413, National Bureau of Economic Research, Cambridge, MA
Jensen, M. C., and W. H. Meckling, 1976, "Theory of the firm: Managerial behavior, agency costs, and ownership structure," Journal of Financial Economics, 3, 305-360
Liang B., 1999, "On the performance of hedge funds," Financial Analysts Journal, 55, 72-85
Sirri E. R. and P. Tufano, 1998, "Costly search and mutual fund flows," Journal of Finance, 53, 1589-1622
1 This article was recently published in the Capco Journal of Financial Transformation, vol 13.
2 Liang (1999) calculated the average annual management fee for hedge funds to be 1.36%, with a median of 1%. This base fee proved to be much smaller than total management fees surveyed from retail mutual funds.
3 As an illustration, Steve Mandel at Lone Pine Capital can charge half of the performance fee (i.e. 10%) of any gain the fund makes from its low. This 10% performance fee continues until the fund has made up 150% of the drawdown from the previous high, then the standard 20% fee kicks in again.
4 Carpenter (2000) studies the optimal portfolio strategy of a manager compensated with a convex option-like payoff and proves this is optimal behaviour.
5 As an illustration, Goetzmann et al. (2001) use an option approach to calculate the present value of the fees charged by a hedge fund manager and show that the present value of the incentive fees can be quite high (i.e. for a volatility of 15%, the fee can be as high as 13% of the assets under management).