The number that froze the feeder-fund industry in December 2008 was reported as approximately $7.5 billion — the net asset value of Fairfield Sentry Limited, a British Virgin Islands fund and the largest of the so-called feeder funds that channeled investor capital into Bernard L. Madoff Investment Securities LLC ("BLMIS"). The figure itself is soft; depending on the reporting date and the methodology used by Fairfield Greenwich Group ("FGG"), Sentry's stated NAV in 2008 was quoted variously between roughly $6.9 billion and $7.5 billion, and the Massachusetts Securities Division's administrative complaint In re Fairfield Greenwich Advisors LLC, Docket No. 2009-0028 (Apr. 1, 2009), works from a slightly different aggregate. What is not soft is the structural fact behind the number: for roughly 15 years Sentry had collected 1 percent in management fees and 20 percent in performance fees on a strategy it did not run, custody it did not hold, and trades it did not execute. Pierre Clauss, Thierry Roncalli, and Guillaume Weisang, in Risk Management Lessons from Madoff Fraud, 10 Int'l Fin. Rev. 505 (2009), open from that arrangement and ask an engineering question: with the quantitative tools available in 2008, could a competent risk officer at a feeder have detected that something was wrong before the FBI did? Their answer is yes — and the proof is in their replication.
What separates the paper from the more polemical red-flags literature surveyed in the previous chapter (see [gregoriou-lhabitant-riot-red-flags]) is its methodological discipline. Clauss and his coauthors are quantitative finance academics, not whistleblowers. They treat Madoff's reported track record as a time series and ask whether any plausible calibration of the advertised split-strike conversion — Madoff's claimed equity-and-listed-options strategy, dissected in detail two chapters earlier (see [bernard-boyle-amazing-returns]) — can generate the observed return distribution. They then layer that statistical finding onto a governance critique of the feeder-fund structure itself. The combination — econometric forensics joined to a theory of intermediary failure — is what makes the paper a centerpiece for any student trying to understand how Madoff's billions moved through "diligence."
What Returns-Based Style Analysis Actually Does
A short, equation-free explanation is owed here, because returns-based style analysis is the analytic engine of the entire critique. The technique was introduced by William F. Sharpe in Asset Allocation: Management Style and Performance Measurement, 18 J. Portfolio Mgmt. 7 (1992). The intuition is simple. A hedge fund will not tell you what it owns; it will tell you, monthly, what it earned. Sharpe's insight was that you can treat the fund's monthly return stream as the dependent variable in a constrained regression against a menu of observable, investable benchmarks — say, the S&P 100 total return, the return on a rolling short-call overlay, and the return on a rolling long-put overlay. The regression coefficients, constrained to be nonnegative and to sum to one, describe the portfolio that would have produced returns most like the fund's reported series. If the fund is doing roughly what it claims, the residuals are small and the implied portfolio looks like the stated strategy. If the residuals are large, or if making them small requires implausible assumptions, the fund is doing something other than what it says.
Applied to Sentry, the technique fails to fit. Clauss, Roncalli, and Weisang show that no constrained combination of S&P 100 long exposure, short-call exposure, and long-put exposure reproduces Sentry's reported Sharpe ratio of roughly 2.5 alongside the observed serial correlation in its monthly returns. To make the residuals disappear at all, the model must assume a "return-smoothing" overlay — a moving average that mechanically suppresses month-to-month volatility. That smoothness is the statistical signature of either illiquid assets being marked to model (the Long-Term Capital pattern) or, more bluntly, of returns being manufactured. No genuine book of S&P 100 stocks and listed options held at a U.S. broker-dealer can produce it: options on the OEX trade in penny increments and mark to market every afternoon. Diagnostic tests for this kind of smoothing were not exotic. Mila Getmansky, Andrew W. Lo, and Igor Makarov, An Econometric Model of Serial Correlation and Illiquidity in Hedge Fund Returns, 74 J. Fin. Econ. 529 (2004), had made them standard practice four years before Madoff's arrest. Any feeder-fund risk committee in 2007 with a Bloomberg terminal and a competent graduate student could have run the analysis.
The Feeder's Conflict: One-and-Twenty on Someone Else's Strategy
The paper's governance argument lands with equal weight, and it is here that the law student should linger. FGG, Sentry's sponsor and investment manager, did not merely market access to Madoff; it represented itself as a sophisticated due-diligence intermediary. Its offering memoranda described independent verification of trades, monitoring of counterparties, and ongoing risk surveillance. As the Massachusetts complaint would later document, Sentry's "due diligence" consisted largely of asking Madoff questions and accepting his answers.
Clauss, Roncalli, and Weisang frame this as a structural pathology rather than a personal failing. The feeder's fee stream — 1 percent management, 20 percent performance — is a function of assets under management; assets under management depend on continued allocation to BLMIS; serious diligence raises the probability that the allocation must be withdrawn. The 1-and-20 was, in substance, a rent extracted in exchange for not asking the questions FGG's clients had hired it to ask. How the partners persuaded themselves that this was nevertheless an honest day's work is a separate story, sociological rather than econometric, and is taken up in the next chapter (see [stolowy-trustworthy-investment-opportunity]).
For the lawyer, several doctrines come into focus. The Investment Advisers Act of 1940, 15 U.S.C. § 80b-6, imposes a fiduciary duty on registered investment advisers and prohibits any "transaction, practice, or course of business which operates as a fraud or deceit upon any client." A feeder fund that misrepresents the scope of its own diligence — independently of whether it knew about the Ponzi scheme — likely runs afoul of § 206(2)'s negligence-based standard, as articulated in SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 191-95 (1963). Rule 10b-5 reaches further but requires scienter under Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193 (1976); a Sharpe-style regression of the kind Clauss and coauthors run is precisely the evidence a plaintiff would marshal to plead recklessness.
NAV, Custody, Counterparty
The authors distill their prescriptive lesson into three independent checks any feeder ought to perform:
- NAV verification — an administrator independent of the manager prices the portfolio against external market data. Sentry's NAV was effectively self-reported by BLMIS.
- Custody verification — assets are held by a custodian unaffiliated with the adviser. BLMIS was its own custodian, the single most glaring operational flag catalogued in the previous chapter.
- Counterparty verification — option trades of the size Madoff claimed should leave footprints in OCC clearing records and dealer confirmations. None existed.
Each verification was, in 2008, routine institutional practice for any prime-brokered hedge fund. The feeders simply did not perform them, and they billed for the omission.
Doctrinal Takeaway
For the law-student reader, the Clauss, Roncalli, and Weisang paper marks a doctrinal hinge. The rule it most directly indicts is Investment Advisers Act § 206(2), 15 U.S.C. § 80b-6(2): a feeder-fund adviser that markets independent diligence and delivers self-reported NAVs has, at minimum, negligently defrauded its clients within the meaning of Capital Gains Research Bureau. The policy lever the paper most directly motivates is the Commission's 2009 amendments to the custody rule, 17 C.F.R. § 275.206(4)-2, which tightened surprise-examination and qualified-custodian requirements precisely to foreclose the BLMIS-as-its-own-custodian configuration. The unresolved tension the paper leaves for the trustee's litigation — whether Sentry was victim, gatekeeper, or both, and how that characterization should shape clawback exposure under SIPA's net-equity regime — is the subject of a later chapter on the Second Circuit's resolution of the question (see [second-circuit-net-equity-madoff]). Quantitative finance had the diagnostic in 1992; the regulatory architecture took until 2010 to catch up; the doctrinal question of what the feeders owed their investors is, even now, only partly resolved.
The feeder funds collected fees for a service — independent risk oversight — that, on the evidence of their own track records, they did not provide.