«The Principal–Agent Problem in Finance (a summary) Sunit N. Shah Published 2014 by the CFA Institute Research Foundation Summary prepared by Sunit ...»
The Principal–Agent Problem in Finance
Sunit N. Shah
Published 2014 by the CFA Institute Research Foundation
Summary prepared by Sunit N. Shah
Despite its prominence in economics today, the study of incentives is relatively
new. Included in this field is the study of the principal–agent relationship.
The principal is one who, within predefined terms, assigns a task to an agent,
who performs the task on the principal’s behalf. If the agent’s incentives are
not aligned with those of the principal and the principal cannot monitor the agent’s actions, the agent has both the motivation and the ability to act undetected against the principal’s interests. This scenario is referred to as the “principal–agent problem.” Early in the history of economics, researchers focused primarily on the behavior of market participants on an aggregate level. Ronald Coase is widely credited with taking the analysis one level deeper in the 1930s with his examination of the firm. In the 1950s and 1960s, economists began to examine how differences in incentives among members of a given team cause the various members to act differently from each other, sparking the subdiscipline of economics known as “optimal contract theory.” At the same time, another set of economists, including Robert Wilson and Kenneth Arrow, were exploring the nature of risk sharing among groups of people with different preferences for risk. Such studies brought to light a set of difficulties that can occur when parties to a contract involving risk transfer alter their actions after the contract is agreed upon. This area of work came to be known as “agency theory.” These two lines of study share a common thread: Each party to a contract must recognize that the other parties may change their behavior after the contract has been struck. The key question is how to structure the contract to minimize potential problems. These two fields eventually merged, and from this union, the principal–agent model and the study of the principal–agent problem were born.
For the principal–agent relationship to be problematic, two ingredients are needed: conflicting incentives and private information. Without the former, the principal may simply leave the agent to his or her own devices; without the latter, the principal need only structure the contract to cover each realization of private information ex post.
©2014 The CFA Institute Research Foundation The Principal–Agent Problem in Finance (a summary) It is not surprising that the financial services industry finds itself rife with potential principal–agent problems. The interconnectedness of the industry has created a myriad of agency relationships in which monitoring is difficult, and many of these relationships involve risk transfer or risk sharing within groups. Consequently, ethical standards within the field must be high, lest the power of participants’ own incentives drive them to act counter to their fiduciary duty to their clients.
Unfortunately, strong ethical standards have not been upheld. As several studies have shown, many participants in the finance industry who have witnessed wrongdoing do not report it and, worse yet, many would engage in illegal activity to get ahead if they were sure they would not be caught. These patterns exist even though other studies have shown that trust is the attribute that is most important to investors when hiring investment managers—even more important than an ability to achieve high returns.
Examining the literature that discusses such problems in two important areas of finance, asset management and the banking industry, is an important step in promoting the introspection needed to reevaluate the industry’s practices.
Compensation Structures in Asset Management Asset managers often receive two streams of revenue for their services. One, called a “management” fee, is a fixed percentage of the assets under management and is often justified by overhead expenses, such as payroll, rent, and infrastructure. The second revenue stream is referred to as the “performance” or “incentive” fee and is a predetermined percentage of the return the manager generates on the investor’s assets.
On the one hand, because the manager does not pay the investor back for losses generated, the incentive fee is an option-like component of the management contract. The management fee, on the other hand, is an equitylike position in the fund because it grows commensurately with the assets.
Consequently, the choice between the incentive fee and the management fee boils down to a choice between option compensation and equity compensation.
Much research has been done on these fee structures. Some of the research examines this topic in a static framework; that is, some studies use a one-period model without considering changes to the fund over time. Most of this subset of the research concludes that incentive fees do not have a high correlation with actual manager incentives and that the management fee component of the structure does a better job than the incentive fee of aligning a manager’s motives with those of the investor.
Research examining this topic in a dynamic framework has incorporated the question of how capital flows react to fund performance over time. Such research finds that, although outperformance in a given period does benefit the 2 ©2014 The CFA Institute Research Foundation The Principal–Agent Problem in Finance (a summary) manager through higher fees in that period, the larger benefit to the manager comes from the additional fees he or she will receive in future periods on the capital inflows that result from this outperformance. Furthermore, the potential for redemptions in periods of underperformance can cause managers to take on less risk than is optimal, with the result that higher incentive fees are often required to counter this effect by motivating managers to take on more risk.
To collapse risk and return into one metric, some investors evaluate managers based on the Sharpe ratio a manager has achieved historically, but research has cast doubt on the effectiveness of such a metric in promoting prudent manager behavior. If the manager has an incentive to maximize his or her Sharpe ratio, that manager may execute strategies that generate a small profit in almost all scenarios but lose an enormous sum in the remaining few. An example is pairing an out-of-the-money short call position with an out-of-the-money short put position. Although such an approach may maximize a manager’s Sharpe ratio in some circumstances, it is unlikely to be in the investor’s best interest.
Investor Short-Termism The direct incentives in the field of asset management are generally governed by the manager’s compensation structure, but investors themselves often indirectly provide managers with incentives to act counter to the investors’ interests through a focus on short-term results. That investors tend to focus on the very recent past, a phenomenon known as “short-termism,” is well documented. In fact, most of the literature considers this behavior as simply one aspect of human nature in investing. The behavior is grounded in several well-established psychological effects, such as recency and saliency, and studies have shown that such behavior not only exists but also is increasing in magnitude over time. This tendency on the part of investors may force asset managers to manage for the short term, thereby causing them to neglect the best interests of those very investors.
The bulk of the research on investor short-termism has examined this behavior in a corporate management context. This research finds that when investors focus on short-term earnings results, corporate managers react to this myopic view by managing for the short term. In the resulting equilibrium, the managers artificially boost short-term earnings and the market rationally anticipates this behavior.
Subsequent research has focused specifically on asset management, beginning with the mutual fund industry. This research finds similar outcomes. Investors chase performance. They pour funds into recent sensational performers, yet often fail to exit underperformers at the same pace. Managers respond to this behavior by decreasing risk in scenarios in which additional outperformance would not help the fund as much as underperformance would ©2014 The CFA Institute Research Foundation The Principal–Agent Problem in Finance (a summary) hurt it and by increasing risk in opposite scenarios. That finding is worth repeating: The empirical evidence shows that when the incentives resulting from investor short-termism contradict what is actually best for the investors, managers tend to act in their own self-interest rather than abide by their fiduciary duty to those investors.
Analysis of investor short-termism in the hedge fund industry is sparse, in large part because of the opacity of the industry and the consequent lack of data for analysis. The research that has been performed demonstrates that investors in this industry also react strongly to short-term results. Outflows tend to respond to underperformance more quickly than inflows respond to outperformance, however, which is a striking difference between investor behavior in the hedge fund industry and in the mutual fund industry.
Systemic investor behavior can have sweeping effects on markets. It can make markets less efficient. The absolutist version of the efficient market hypothesis maintains that the existence of one rational trader in the market is sufficient to nullify the effect of many irrational ones; the one rational trader will simply counter any irrational trades from others and return the market to efficiency. Research has shown, however, that to the extent that the rational traders have limited capital or limited time horizons, such behavior would not be rational because the market could, in words attributed to John Maynard Keynes, remain irrational longer than the rational investor remained solvent.
Many asset managers, using Warren Buffett as a model, have taken steps to escape the effects of investor short-termism by setting up permanent capital vehicles resembling Buffett’s Berkshire Hathaway. Berkshire Hathaway has provided its investors with a return that far outpaces the risk-free rate and has also decisively outperformed the S&P 500 Index. It has, however, suffered its own ups and downs over time. For example, as the tech bubble expanded in the late 1990s, Berkshire Hathaway lost almost half its capital while the stock market gained 32%. But in such cases, Buffett’s firm was able to hold its positions until the market corrected because, unlike a mutual fund, Berkshire Hathaway is a firm with a fixed number of shares. As a result, anyone wanting to sell his or her shares has to sell them to another investor rather than back to the issuer.
Management firms that have followed suit include Pine River Capital Management, which set up two publicly traded REITs, and many hedge funds, such as Moore Capital, Greenlight, Paulson, SAC, and Third Point Management. All of them have created reinsurance companies in which the hedge funds manage the assets.
banking industry creates an environment that is ripe for potential incentive conflicts. In the run-up to the financial crisis, bank employees often faced conflicts between the incentives created by their compensation contracts and their duty to the shareholders of their companies. The government certainly faced tension between its duty to represent taxpayer interests and its stance of not interfering with private industry. And the bailouts themselves forged a new principal– agent relationship between bank managers and the taxpayers whose money was used to stabilize their institutions. The complex relationships among the various parties, combined with the density of relevant information in each transaction, provided a fertile breeding ground for principal–agent issues. Furthermore, the amount of money involved amplified each individual’s incentives significantly, making those incentives more difficult to ignore.
Many bank employees faced conflicts between their own incentives and their responsibility to their banks’ shareholders. Traders, who take on risk on behalf of bank shareholders, often have unlimited upside to their bonuses, whereas the shareholders participate equally in upside and downside outcomes. This situation motivates traders to take on much more risk than is optimal for the shareholders. During the financial crisis, this misalignment became apparent, with many traders receiving enormous bonuses, even though their trades brought down the very banks for which they worked.
The research is more forgiving in its evaluation of bank managers than traders. Like traders, some bank executives received large bonuses over time frames that included periods of loss that wiped out a decade’s worth of prior earnings. However, the research that examines the relationship between the banks’ losses and the structural alignment of their managers’ incentives is mixed. Some of the evidence suggests that the banks that did better during the crisis had management compensation structures that were better aligned with the interests of shareholders. Other research demonstrates that banks whose managers had incentives closely aligned with shareholder interests actually performed worse. A third stream of research has shown that bank CEOs did not move to liquidate their personal holdings before the crisis, which demonstrates that they did not anticipate the losses.
Overall, the literature in this area seems to conclude that, for better or worse, bank managers were as unaware as the rest of the market that the crisis was going to materialize. It also shows, however, that bank managers’ personal incentives probably played a role in the banks’ behavior under the Troubled Asset Relief Program (TARP). Banks whose CEO compensation was above the TARP limit were more likely to reject TARP, and this likelihood increased significantly as manager compensation rose.