Forensic Accounting Investigation

Forensic Accounting Investigation

Supporting analysis of transactions when books and records are incomplete.

Client

Bankruptcy Trustee of Mortgage Backed Security Hedge Fund

Challenge

A hedge fund with a $600 million portfolio of securities claimed to be “interest rate neutral”, therefore interest rate changes would not impact its net asset value. The fund was leveraged with repurchase agreements (“repos”). However, when interest rates spiked, asset values plunged, and the repo lenders demanded payment. The fund was unable to make the payments, and the repo lenders took the securities and liquidated them causing the fund to collapse. Subsequently, the fund filed bankruptcy. Trading records were in disarray; the general ledger was incomplete; documentation of due diligence and/or modeling of mortgage securities did not exist.

Solution

The bankruptcy trustee engaged legal counsel, forensic accountants, and valuation experts. The forensic team, using the hedge fund’s copies of broker-dealer trade documents reconstructed the fund’s activities. Since the fund had no records of the analysis performed on potential security purchases, the team performed its own analysis on each security. The forensic team also provided the trustee and bankruptcy counsel with detailed trade activity to support depositions of the broker-dealers that sold securities to the fund.

Results

The bankruptcy trustee issued a comprehensive report of the funds’ operations, of the margin calls and liquidation of securities by the broker-dealers, and an analysis of the valuation and volatility of the funds’ investments. The reconstruction of the funds’ activities was significant to the ability of the trustee to extensively depose the broker-dealers’ representatives. The report resulted in establishing a litigation advisory board of investors, and then subsequent litigation against the broker-dealers resulting in significant recoveries by the investors. Headcount to extract the required data from the securities, and manually input into the valuation model, was reduced from approximately 20 staff, to only 5 staff.