Form PF and Hedge Funds: Risk-measurement Precision for Option Portfolios by OFR
Mark D. Flood
Office of Financial Research
Office of Financial Research
Mangrove Partners Narrowly Avoids “Extinction-Level Event”
The Securities and Exchange Commission’s Form PF is the implementation of Congress’s post-crisis mandate for risk reporting by hedge funds to help protect investors and monitor systemic risk. We extend the methodology of Flood, Monin, and Bandyopadhyay  to assess the risk measurement tolerances of Form PF for portfolios including options exposures. We generate a range of simulated portfolios of equities and equity options, where the weights are calibrated so that portfolios appear identical on Form PF. We assess the measurement tolerances of Form PF by examining the minimum-maximum range of actual risk exposures as measured directly from portfolio details. We find that the possible range of variation is significant. For portfolios that include options but do not report value at risk on Form PF, the range is especially large.
Form PF And Hedge Funds: Risk-Measurement Precision For Option Portfolios – Introduction
Form PF implements a Congressional mandate, enacted in the wake of the recent financial crisis, for reporting of hedge fund risk exposures. Given that risk measurement is the goal, it is important to understand the precision of Form PF in capturing portfolio risks. We assess the risk measurement tolerances Form PF has for hedge fund portfolios that include options exposures. We generate a range of simulated portfolios of equities and equity options where the portfolios have observable risk characteristics, but where the weights are calibrated so that portfolios appear identical on Form PF. We then assess the measurement tolerances of Form PF by examining the range of actual risk exposures as measured directly from portfolio details. The paper reaffirms the feasibility of the constrained risk-maximization methodology of Flood, Monin, and Bandyopadhyay , who considered market-neutral portfolios of exchange-traded equities, without options. We find that the inclusion of options has a significant impact on actual portfolio risks. For portfolios that include options but do not report value at risk (VaR) on Form PF, the range of permitted actual risks is especially large.
The new provisions for enhanced regulatory reporting on private funds, including hedge funds, appear in Section 404 of the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act; see U.S. Congress ) and have the twin goals of investor protection and systemic risk assessment. The provisions are only a small part of the much larger package of interconnected reforms in the Act. Although the Dodd-Frank Act is an overall response to the financial collapse of 2007-09, Congress’s concerns about the systemic risks posed by hedge funds clearly began earlier. The crisis of Long-Term Capital Management (LTCM) in 1998 generated substantial interest in studying the potential systemic risks posed by hedge funds (see President’s Working Group on Financial Markets  and Bernanke ). The financial crisis of 2007-09, which included a disruption to quantitative funds as a significant foreshock in August 2007 (see Khandani and Lo ) reinvigorated these concerns. We document a series of House and Senate hearings, occurring before and during the crisis, on the systemic threat posed by hedge funds. A particular concern in the pre-crisis discussions was the increase in overall leverage in the system generated by the reliance of large broker-dealers on hedge funds for risk transfer services through the derivative and securitization markets.
The precision of Form PF in capturing risk exposures is an important question, because Form PF is the primary supervisory tool for measuring these risks systematically across the sector. Moreover, because Form PF is relatively new, and because the data collected are confidential, it is difficult to assess the form’s precision directly. However, although the data regulators collect on Form PF are confidential, the form itself is public information. Our methodology relies only on the form and its instructions, along with standard market data sources for constructing and analyzing simulated portfolios. In the science of measurement, or metrology, a “tolerance” is the limit on acceptable deviations between the underlying true value of the measurand and the actual measured result. Any such deviations we detect in our analysis are implicitly acceptable (or tolerated) by Form PF. Note that our assessment focuses on the risk-measurement precision of Form PF, rather than its accuracy (or statistical unbiasedness); we do not provide a basis here for assessing the accuracy of hedge funds’ actual Form PF reports, which have been used effectively in other contexts (e.g., OFR, 2015).
To examine the risk-measurement tolerances of Form PF, we generate a large collection of simulated hedge funds and report their risk exposures according to the instructions of Form PF. Each fund’s portfolio consists of an equities sub-portfolio and an equity options sub-portfolio. We construct the portfolios by applying purely quantitative textbook investment strategies: (1) a dollar-neutral stock screen based on alphas from a factor model for the equities sub-portfolio, and (2) a portfolio of short straddles for the equity options sub-portfolio. Each fund is constrained to have an identical presentation on Form PF, where the constraint is satisfied by careful calibration of the portfolio weights for each fund. That is, any differences in actual risks across funds would not be observable on the official report.
Examination of the cross-sectional distributions of risk measures of the simulated portfolios reveals significant dispersion in the actual portfolio risks of funds with identical presentation on Form PF. For instance, among several variants of VaR and expected shortfall (ES) that we calculate, we find that the maximum portfolio risk is on average 42 percent higher than the median risk, despite all funds reporting the same VaR on Form PF Question 40 out to one-hundredth of a percentage point of net asset value. Furthermore, for funds that do not report their VaR on Form PF Question 40, the maximum risk conveyed by these risk measures is more than ten times higher, averaging 535 percent higher than the median risk.
We also directly examine the impact of options on the measurement tolerances of Form PF by comparing to tolerances for portfolios without options. For funds that do not report their VaR on Form PF Question 40, we find that the portfolios including options have maximal risk about 6.35 times the median risk, almost three times the value of 2.39 for this statistic in the equities-only case. (The median portfolios themselves, with and without options, will be similar in our simulations, because all are constrained to appear identical on Form PF.) This suggests that options used in a speculative manner can greatly increase the measurement tolerances of Form PF.
The remainder of the paper proceeds in four sections. Section 1 discusses the legislative and scholarly background for hedge fund risk reporting. Section 2 outlines our simulation methodology and Section 3 presents the results. Section 4 concludes.
See full PDF below.