30.1.1 Stated and Permitted Investment Strategies

A critical aspect of a fund's risk is the potential divergence between the stated investment strategy of the fund and the actual investment strategy. The stated investment strategy of a fund is the investment strategy that a diligent investor would expect the fund to pursue, based on a reasonable analysis of information made available by the fund. The actual investment strategy of a fund at a particular point in time is the investment strategy being implemented by the fund. Finally, the permitted investment strategies of a fund delineate the range of investment strategies that the fund's managers have communicated and are mandated as allowable for the fund to implement. The range of permitted investment strategies is defined by the organizational documents and any other documents or communications provided by managers to clients or prospective clients.

At the core of each fund is the investment mandate. An investment mandate is an explicit or implicit statement of the allowable and intended strategy, goals, and/or risks of an investment program. The investment mandate and its stated investment strategy are typically disclosed in the various documents provided to investors prior to their decision to invest in the investment vehicle. Descriptions of investment mandates vary tremendously in length and specificity.

30.1.2 Deviation of Actual Strategies from Stated Strategies

Risk analysis should include examination of the potential for the actual investment strategy to differ from the stated investment strategy or the permitted investment strategies. Further, an analysis of a fund's risk management practices should include an analysis of the fund's ability to detect and correct the deviation of the actual investment strategy from the stated investment strategy or permitted investment strategies.

The actual investment strategy may differ from the stated investment strategy for three primary reasons: style drift, operational errors, and fraud.

Style drift (or strategy drift) is the change through time of a fund's investment strategy based on purposeful decisions by the fund manager in an attempt to improve risk-adjusted performance in light of changing market conditions. For example, a manager may find that a strategy is not working as anticipated and may make changes in an attempt to better control risk and/or generate better returns.

Operational errors can cause unintentional changes to an investment strategy through human error, including poor risk management systems and controls. Intentional deviations of the actual investment strategy from the stated investment strategy in disregard of investor interests, however, are examples of fraud. Fraud regarding investment strategies ranges from deception regarding the actual implementation of an investment strategy to misstating returns or even blatant theft.

30.1.3 The Investment Process

The investment process includes the methods a manager uses to formulate, execute, and monitor investment decisions, and spans the range of investment activities, from the design of the investment strategy, through the implementation of the ideas into decisions, and ultimately to the placing and execution of trading orders. The process includes sourcing of ideas, determination of transactions, setting of leverage, and execution and allocation of trades. The investment process can be divided into two stages: (1) the conception, development, and specification of the stated investment strategy, and (2) the application and implementation of the actual investment strategy.

Understanding and documenting a fund manager's investment process are not always straightforward tasks. Some fund managers rely on a largely quantitative, systematic investment process, such as a black-box process, discussed in Part 3. Other managers pursue a more discretionary process that relies on the investment decisions made by just one person or a small team of investment professionals.

Analysis of quantitative investment processes focuses on the software that captures the fund's investment strategy. Analysis of discretionary investment processes focuses directly on the judgment and skill of individuals. The investment process in discretionary cases centers on the investment management governance process, which is the explicit or implicit set of procedures through which investment decisions are made. For example, one fund may assemble talented managers who function independently in an entrepreneurial culture and compete to persuade a centralized leader or committee that their decisions have merit. In other cases, major investment decisions are tightly controlled by committee processes, with portfolio managers exercising reduced latitude.

30.2.1 Investment Process Risk

Investment process risk is economic dispersion caused by imperfect application of the stated investment strategy by the investment team. Investment process risk emanates from the difference between the proper implementation of the investment strategy as stated and the investment strategy as actually determined and implemented. Investment process risk is especially prevalent in the hedge fund industry because of the industry's skill-based nature. Investment process risk is an idiosyncratic risk caused by the fund manager's structure and operations. Although investment process risk cannot generally be eliminated, it is a risk that investors should strive to reduce.

The detection of investment process risk can include quantitative analysis. Historical and current investment returns of a fund should be analyzed against the returns of market indices and similar funds to infer the extent to which the fund's returns are consistent with the market environment. Investment process risk can also be detected qualitatively. The key to optimal control of investment process risk is a sound, well-defined, well-designed investment process. A well-defined investment process clearly specifies the functions of the investment team; a well-designed investment process maximizes the potential benefits of the stated investment strategy while minimizing the costs of investment process risk.

30.2.2 Process Risk of Implementing a Strategy

Given an investment strategy, the process of implementing that strategy from concept to actual portfolio positions involves risk. A starting point for risk analysis is precisely identifying the nature of the fund's investment strategy. Does the fund have an active and highly hedged strategy, such as a quantitative equity market-neutral hedge fund, or does the fund have a more passive and unhedged strategy, such as a fund of funds?

Assessing the stability and reliability of a fund manager's investment process is essential to assessing risk. In the case of quantitative, systematic investment processes, much of the investment process resides in a fund's software or computer algorithms. A fund's computer algorithms are procedures within the software that determine trading decisions or other outputs. The fund manager may be unwilling to reveal the details of the computer algorithms because they represent part or all of the fund manager's competitive advantage. Prospective investors may decide not to invest in any investment process they do not understand. This is a blunt risk management policy, but an investor who cannot understand the investment process may not be able to comprehend the risks associated with the process. This lack of transparency in the investment process contributes to the investment process risk borne by the investor. Investors may mitigate this risk through diversification into other funds but must ultimately decide whether the potential returns merit the risks of a process that lacks transparency.

Given transparency, the way to manage the risk is to carefully examine and understand the details of the investment process. It is not necessary to read the underlying computer code behind every computer algorithm of a quantitative manager. Instead, the investor must understand the structure of the algorithms. The investor should determine which computer algorithms are used to evaluate different financial instruments and whether each computer algorithm includes all relevant variables. For instance, with respect to convertible bond arbitrage, appropriate economic inputs might be the underlying stock price, the historical volatility, the implied volatility, the current term structure of interest rates, the credit rating of the instrument, the duration of the bond, the convertible strike price, and any call provisions in the bond indenture. The investor should understand the purpose of the computer algorithms with respect to pursuit of alpha and control of risk.

All investment processes involve interaction of both people and computers. For example, even with a black-box investment process, there are people and a governance process involved with the design and adjustment of the mechanical process. Whether quantitatively or qualitatively focused, the investment process needs to be qualitatively analyzed for soundness through analysis of its framework of governance. This framework includes the organizational chart, along with the existence, nature, composition, and documentation of the responsibilities of all investment personnel and investment committees. The committee framework includes the processes for selecting members, selecting officers, scheduling meetings, forming agendas, and voting, as well as the process for recording, approving, distributing, and preserving the minutes. The framework also includes the detail, accuracy, and importance of procedure manuals. The analysis seeks to ascertain the susceptibility of the investment process to conflicts of interest, fraud, and incompetence.

30.2.3 Investment Process Risk and Leverage

Leverage is a key factor in risk, and therefore leverage management is a key factor in understanding and controlling investment process risk. Investment process risk stems from the leverage inherent in the stated investment strategy as well as the risk that the actual leverage will differ from the targeted leverage.

Another major concern of a leveraged fund is risk from following an investment strategy that generates a portfolio that is shared by other leveraged and active traders. For example, if numerous highly leveraged funds are pursuing very similar strategies, the funds may begin to act like a herd by having similar positions. If one fund liquidates or deleverages its positions quickly, the market impact of such trading may trigger liquidations and deleveraging by the other funds, which can spiral into rapid, large losses. Analysis of the risk from herd trading requires knowledge of the fund's positions, knowledge of the likely or actual positions of similar funds, and an understanding of the degree of liquidity in the relevant markets, so that the potential price impacts of herd trading can be evaluated.

30.2.4 Investment Process Risk from Style Drift

Investors first decide whether to accept the fundamental economic risks of the stated investment strategy and its tactical and strategic asset allocations. Investors then need to assess the potential for risks that are not fundamental to the stated strategy. Style drift shifts the actual investment strategy of the fund away from the stated or mandated strategy of the fund. Style drift may allow a fund manager to stray into markets in which the fund manager has limited expertise in search of attractive investment opportunities when the manager's previous investment strategy is no longer generating attractive opportunities. Style drift is a major component of investment process risk.

30.2.5 Process Risk and Market Volatility

Risk caused by changes in market values is an important component of the total risk of an investment in a fund. Market risk emanates both from the stated investment strategy of a fund and from the tendency of market risk to cause and interact with other types of risk brought on by divergence of the actual investment strategy from the stated investment strategy.

Improperly implemented investment processes, such as style drift, inappropriate application of leverage, and inappropriate systematic and idiosyncratic risk exposures, are especially problematic when they occur during periods of high market volatility. Therefore, analysis of investment process risk should include analysis of the ability of the fund to avoid and rectify investment process errors during turbulent market conditions.

Losses due to market risk, including losses so extreme as to cause fund failures, may happen very quickly due to the speed at which market prices change during a time of crisis. Even if the increased market risk of a fund can be detected prior to large fund losses, the fund's investors may not have time to request and attain a return of their investments before the market has caused substantial losses.

30.4.1 Operational Errors

Operational errors are inadvertent mistakes made in the process of executing a fund's investment strategy. Operational errors range from minor errors with inconsequential losses to major errors that can cause a fund to fail. In a fund failure, all or nearly all of the assets under management are lost. Losses due to operational errors can be exacerbated by market risk.

For example, in rare circumstances, errors in executing trades may occur, such as the execution of a trade in the wrong size or direction. The size of the appropriate trade may be miscalculated or the size may be mistakenly altered between the time that the size is calculated and the time that the trade instructions are transmitted. The risks from such errors vary in severity based on market conditions.

During calm market conditions, the probability of trading errors is especially small, and the economic impact of those errors tends to be minimal, since the errors are likely to be detected quickly and can be corrected before prices have changed substantially. However, trading errors become more possible and the likely impact of those trading errors increases during chaotic markets, when errors may not be detected quickly and prices may move substantially before corrective action is taken.

The risk of operational errors can be reduced through well-designed and implemented operational systems, which can help prevent and detect errors. The purpose of detecting errors is to put corrections in place before losses increase. Improved operational systems are discussed in a later section.

30.4.2 Agency Conflicts

Operational risk includes the intentional actions of fund employees that are contrary to the interests of investors. Agency costs can be reduced when incentive-based compensation schemes are used to bring the interests of the fund manager (the agent) into closer alignment with the interests of the investors (the principals). However, optimal compensation schemes allow the interests of investors and the interests of the fund manager's employees to remain in conflict because the expense of resolving some conflicts may be too costly, and all interests can never be perfectly aligned.

An extreme example of a conflict of interest is a rogue trader. A rogue trader intentionally establishes substantial positions well outside the investment mandate. Rogue trading is most often caused by strong incentives or pressures to generate performance, combined with losses that jeopardize a trader's career if not recouped. A rogue trader can single-handedly cause the collapse of a fund or even a large bank. However, well-designed operational systems (including carefully monitored position limits) can reduce or even eliminate a rogue trader's ability to establish dangerously large positions.

In less extreme cases than rogue traders, operational risks can be caused by insufficient incentives, leading to lackadaisical management, or overly strong incentives, leading to gaming of the compensation structure. Gaming refers to strategic behavior to gain benefits from circumventing the intention of the rules of a particular system. In the context of a fund manager as an economic agent, gaming includes a sequence of efforts intended to generate gains to the agent from flaws in the compensation structure without regard for losses to investors. For example, if an investor develops a measure on which performance will be judged, such as tracking error, gaming would refer to strategies of a fund manager—such as managed returns, discussed in Chapter 10—designed to underreport tracking error and overstate perceived risk-adjusted performance while not improving actual performance. Another example of performance gaming relates to incentive fees, discussed later in this chapter.

30.4.3 Operational Fraud

Some of the most severe examples of operational fraud were described in Chapter 29 on fund collapses, wherein fund managers absconded with investor funds. However, operational fraud from the perspective of an investor is any intentional, self-serving, deceptive behavior in the operational activities of a fund that is generally harmful to the investor. Operational fraud can be reduced through separation of duties. For example, no single person in the fund should be responsible for more than one of the following areas: trading, risk management, and accounting. The fund problems discussed in Chapter 29 illustrate the importance of separating trading authority from risk management functions and separating portfolio management functions from the investment accounting and valuation processes. Further, strong controls to maintain the segregation of assets ensure that investor assets are not misused for the personal benefit of the fund manager.

30.5.1 Incentives Can Increase Operational Risk

The desire of portfolio managers, traders, and, indeed, all fund employees is to retain existing assets, attract new assets, and increase revenues. Therefore, it is in their best interest to report high and consistent performance. However, the incentives that motivate fund managers to report high and consistent performance can cause unintended consequences.

For example, as discussed in Part 3, there is an incentive for a fund manager to reduce investment risks when substantial profits have accrued to lock in an acceptable level of performance and the continued opportunity to manage assets. Conversely, there is an incentive for a fund manager to increase risks when substantial losses have accrued to increase the chance of recouping the losses and reaching an acceptable level of performance to sustain retention of assets. In particular, a rogue trader with large losses or a trader who has suffered large losses through errors has an especially strong incentive to recoup losses before they are fully revealed. Greatly increased risk taking may provide such a trader with a higher chance of recovering the losses and thus salvaging a career.

The call-option-like nature of performance-based fees also provides fund managers with the incentive to take high risks. The risk-taking incentive varies, based on the moneyness of the option. The incentive to take high risks is greatest when the fund is near or below its high-water mark, as discussed in Part 3. The net result is that portfolio managers and traders have incentives to take higher risks (and, in some cases, lower risks) than are appropriate, given the investment mandate and market conditions.

30.5.2 Internal Control Procedures

Detection of operational risk relies on strong risk management and risk monitoring systems. Using many of the quantitative tools described in other chapters, the fund management team should constantly monitor (1) the current market values of the positions, (2) the recent financial performance of the positions, and (3) the market risks of the current positions.

The portfolio manager and the risk manager need to agree on the measurement of risk, the risk limits, and the methods of ensuring that positions do not cause risk limits to be exceeded. Risk limits are the maximum levels of measured risk that are allowed in a portfolio, in terms of both individual risks and aggregated risks. Risk may also be controlled through position limits. A position limit is a specific restriction on the size of the holdings of a particular security or combination of securities.

All trades need to be entered and managed in an electronically monitored system that enforces position limits and assists in the control of risk limits. Position limits are relatively easily enforced through a system that prevents placing orders that would cause position limits to be exceeded. Risk limits are less easily enforced because, for example, risks can be altered by changes in market conditions. For example, a change in market volatility can cause a risk measure such as value at risk to change, even when no additional transactions have been entered. Further, while prevention and correction of violations of position limits are objective and clear processes, the correction of risk limits is more complex and can be accomplished in many ways.

If traders are allowed to conceal the size of their positions and/or trading losses from others or are allowed to consistently exceed risk limits, it will be only a matter of time before the fund experiences catastrophic losses. The risk manager needs to have sufficient skills, systems, and authority to execute the responsibility of limiting risk consistent with the investment mandate.

Traders and portfolio managers may also become aware of errors in the systems of measurement and management of risk. Thus, through time, these traders and portfolio managers may become equipped with methodologies to obscure true risk from the risk manager. The risk manager must have the skills, dedication, authority, and support to design, implement, and monitor risk management systems that provide reasonable protections from investment process risks.

The ongoing monitoring of the financial performance, risks, and current market values of portfolios and positions allows early detection of investment process problems and other operational problems. Proper valuation is the core of the monitoring process, since valuation is central to performance measurement and risk measurement.

30.5.3 Valuation Procedures

It is imperative that traders and portfolio managers do not value (i.e., price) their own positions for risk management and reporting purposes. Valuation of positions for reporting and risk control purposes should be determined through an independent process. Valuation can be the object of powerful conflicts of interest. Investors select funds on the basis of risk-adjusted returns, and incentive fees are paid on reported returns. Accordingly, some managers may manipulate reported returns to enhance their income or maximize the probability of retaining or growing investor assets. These conflicts of interest may provide incentives for the fund manager to (1) obscure losses, (2) smooth returns by shifting performance between reporting periods, and (3) vary risks to recoup losses or lock in profits.

Managers may have an incentive to obscure losses to avoid reporting inferior performance and provide time to recoup losses. Managers may also have an incentive to smooth returns. If performance is smoothed over a long period of time, the fund appears to have lower volatility and higher risk-adjusted returns, which can assist in attracting and retaining assets. Managers smooth returns by moving profits from periods of high performance to periods of low performance and by moving losses from periods of low performance to periods of high performance. For example, at the end of a very successful period, a manager can smooth returns by placing a reduced value on the end-of-period positions. The reduced valuations lower the prior period's financial performance and provide an immediate gain for the next period. At the end of an unsuccessful period, a manager can smooth returns by placing an increased value on the end-of-period positions.

Finally, fund managers have an incentive to vary risks to game performance evaluation. A fund manager with a benchmark has an incentive to lock in profits by lowering risk when profits have already exceeded the benchmark. In so doing, the manager is increasing the probability of outperforming the benchmark over the reporting period. Conversely, a manager underperforming a benchmark has an incentive to take high risks in an effort to reach the benchmark by the end of the reporting period.

For exchange-traded securities, the fund's own back office team can easily value the fund using pricing feeds from market data systems, combined with position reports from the custodian, prime broker, and administrator. Less liquid positions traded in the over-the-counter (OTC) market, such as mortgage-backed or convertible securities, should be valued through the use of external dealer quotes rather than by the models or prices supplied by a trading desk. Objectivity and independence in the valuation process are essential.

Systems of valuation, performance reporting, and risk measurement that permit ambiguity or subjectivity are fertile ground for operational risk. Relatively minor incidents of bending rules to one's advantage, if unchecked, can evolve into more reckless behavior and create incentives for additional rule breaking to cover up past indiscretions. Chapter 29 described fund collapses that began with minor losses that were obscured and became a gateway to increasingly desperate risk taking in an attempt to restore profitability.

30.5.4 Custody of Assets

Investors must also be concerned about the custody of assets. Custody refers to the safekeeping of the cash and securities of a fund. Are assets segregated for the fund investors as a group or for individual investors in the fund? Are controls in place to prevent the movement of funds into the fund manager's personal accounts? Are legal agreements with prime brokers effective in preventing commingling of assets with the assets of the brokers, or will fund manager assets become commingled with brokerage assets, making it difficult to disentangle assets, such as during the bankruptcy of the European prime brokerage operations of Lehman Brothers?

Managed accounts, or separate accounts, can provide the safest and most transparent arrangement for investors. In a managed account, the assets managed by the fund manager remain in the investor's account at a brokerage firm. The investor retains custody of the assets, so it is virtually impossible for the fund manager to withdraw the funds. Given that the account belongs to the investor, the investor can access account positions, balances, and performance in real time. This allows the investor to see returns on a much more frequent and independent basis, if desired, and to evaluate them. There is likely to be no question about the valuation of the securities in the account, as the prices of the positions should be determined by the custody bank or broker, not by the fund manager.

30.5.5 The Culture of the Fund

The previous sections discussed potential conflicts of interest and incentives that could lead a manager to manipulate reported performance and/or alter risks in an attempt to game investors. Of course, most fund managers are professionals who serve the interests of their clients and perform their responsibilities with integrity. How can a prospective investor gauge the integrity of a fund manager? Each fund manager may be viewed as possessing a fund culture.

A fund culture is a generally shared set of priorities and values within the fund's organization. At the most positive end are investment teams with a fund culture of rigid adherence to the highest professional standards of conduct. However, some investment teams may develop fund cultures that sacrifice such standards in pursuit of pragmatism or even disregard for investor interests. One of the strongest protections against operational risk is a fund culture that fosters competence, honesty, and diligence. Evidence of the true culture of a fund can be found by examining its systems of risk management and methods of performance valuation, performance reporting, and employee compensation.

30.6.1 Option Sensitivities and Put-Call Parity

Chapter 6 discussed the concept of put-call parity, which can be expressed as follows:

(30.1)

It is assumed that the put and call are both European options with identical maturities and strike prices. Intuitively, the long position in a put (+Put) removes the downside risk exposure from being long the underlying asset (+Stock), whereas the short position in the call (–Call) gives up the profit potential from price increases in the underlying asset. The combination removes all risk, creating the equivalent of a long position in a riskless bond (+Bond). Therefore, the hedged position in Equation 30.1 is equal to owning a riskless bond. Specifically, the bond has a maturity date equal to the expiration of the options, a face value equal to the strike price, no coupon, and a price equal to the strike price discounted at the riskless rate.

To the extent that markets are well functioning, the total value of the hedged position on the left side of Equation 30.1 evolves through time the same as that of the right-hand side: It grows at the riskless rate from being equal to the present value of the strike price to being equal to the strike price at maturity or expiration. The sensitivity of this hedged portfolio to the price and volatility of the underlying financial asset must be zero. This insight drives the relationships discussed in the next section.

30.6.2 Using Option Sensitivities to Hedge Risk

Various sensitivities of option prices to underlying option parameters, including delta, gamma, and vega, were discussed in earlier chapters. These sensitivities are often referred to as the Greeks, since most of them are denoted with letters of the Greek alphabet. Exhibit 30.1 summarizes the most important Greeks for the underlying stock, a put, a call, and a fully hedged position assuming identical strike prices and expiration dates.

As discussed in Chapter 6, sensitivities such as those shown in Exhibit 30.1 are partial derivatives that examine the effect of each underlying risk source in isolation. Portfolio managers with positions involving explicit or implicit options often use these option sensitivities in tandem to understand and analyze the total risks of a portfolio.

Note that a call and a put with identical strike prices and identical expiration dates have identical gammas and vegas. The reason that the gamma and vega values are the same is that a portfolio that is long the put and short the call must be gamma neutral and vega neutral. The logic is as follows: If the underlying asset (stock) is gamma and vega neutral, and if +Stock + Put – Call is gamma and vega neutral, then the difference (Put – Call) must also be gamma and vega neutral. Note, however, that the delta risk of calls and puts are of opposite signs and always differ by 1.

Thus, from inspection of Exhibit 30.1, being long a call and short a put (+Call – Put) or being long a put and short a call (+Put – Call) is hedged with respect to gamma risk and vega risk but leaves delta risk at +1 or –1, respectively. The delta risk is laid off by either a long position in the underlying asset (which adds delta = +1) or a short position in the underlying asset (which adds delta = –1). This analysis ignores counterparty risk and assumes that the options have the same strike prices, expiration dates, and underlying assets. Hedging involving options that differ by moneyness or by expiration date requires additional positions and hedge ratios based on the differences in delta, gamma, and vega.

30.6.3 Viewing Options as Volatility Bets

Option traders sometimes say that calls and puts are the same. This claim is in stark contrast to the view of many market participants less familiar with option hedging. Most market participants do not hedge risk with options and view calls and puts as complete opposites because long calls are bullish positions and long puts are bearish positions. But the claim that calls and puts are the same makes sense from the perspective of a trader who maintains delta neutrality. From the trader's perspective, the difference in delta between a call and a put is trivialized by the trader's ease in managing delta by simply expanding or contracting a hedging position in the underlying asset (to maintain a target delta such as delta neutrality). An option trader may view management of vega risk and gamma risk as more problematic due to trading costs such as bid-ask spreads. The trader can buy vega and gamma with either a long position in a call or a long position in a put. Similarly, the trader can short vega and gamma with short positions in either calls or puts.

In other words, the only difference between the call and put in Exhibit 30.1 is in the delta, and the delta of a single position to a trader who maintains delta neutrality is easy to hedge using the underlying asset. Option traders often take bets with respect to volatility and view positions in the underlying asset as the slack variable that is used to control delta. A slack variable is the variable in an optimization problem that takes on whatever value is necessary to allow an optimum to be feasible but, while doing so, does not directly alter the value of the objective function. In this case, the position in the underlying asset keeps the portfolio from being affected by directional moves in the market (i.e., controls the delta) but is not a direct source of alpha and does not affect gamma or vega.

For those market participants who use options to place directional bets on the underlying assets, long positions in options provide leverage, positive skews, and limited downside risk, all of which can be especially attractive to aggressive investors with a view that an underlying asset is mispriced. Some sophisticated traders take the other side of these option bets when they perceive, for example, that volatility is being overpriced. These traders maintain delta neutrality and therefore view the positions in mispriced options as pure bets on volatility that offer alpha.