On September 6, 2022, the Commodity Futures Trading Commission (“CFTC”) issued an order accepting an offer of settlement, in which a trader neither admitted nor denied the CFTC’s findings that he mismarked the forward price curves related to U.S. dollar interest rate derivatives (“Speaking Order”). The CFTC imposed a three-year trading ban and $250,000 civil monetary penalty against the trader, who was a managing director at a global bank. In this alert, we summarize the CFTC’s findings and conclusions, and offer insights about controls that may help your organization protect against regulatory exposure resulting from similar conduct.
The trader was head of an interest rate derivatives trading desk (the “Desk”) in the New York office of a global bank. The Desk acted as a market maker and built and distributed its own pricing for OTC derivatives. The trader managed all activities of the Desk, and his responsibilities included marking the USD LIBOR forward curve, which typically was derived from observable market data such as prices of bonds, futures, and treasuries.
According to the CFTC, the Desk calculated and maintained two separate USD LIBOR curves: first, a “Live Curve” that updated automatically during the day based upon live observable price data, and second, a “Closing Curve” that the Desk marked based upon end-of-day settlement prices. The bank’s policies required both curves to be marked based upon observable mid-market prices. The bank used the Closing Curve to calculate the profit and loss (“PnL”) of the trader’s trading book.
The CFTC’s Misconduct Findings
The CFTC found that the trader “mismarked and caused to be mismarked the end-of-day and end-of-month Closing Curves,” and that his curves “diverted from observable mid-market prices.” According to the Speaking Order, the trader marked his Closing Curves in a manner that benefitted the positions in his trading book. For example, the trader would mark the Closing Curve higher if a higher mark would increase the PnL of a long position in his book. The trader allegedly engaged in this behavior to hide his substantial losses.
The Bank’s Internal Controls Team
The CFTC found that the trader took steps to conceal his scheme from others at the bank. One way he concealed his conduct was by gaming the bank’s internal control policies. At the end of each month, the trader sent the Closing Curve that he calculated to another division at the bank that was responsible for internal controls relating to valuation and accuracy of the bank’s PnL. That division calculated its own version of the Closing Curve (the “Control Curve”) based upon observable mid-market price data, and compared each point on the Control Curve to the trader’s Closing Curve.
The bank had a policy that required it to adjust the Desk’s PnL if any point on the Closing Curve diverged from the Control Curve by more than one basis point. The CFTC noted that, given the general market liquidity of the products traded by the Desk, numerous individuals at the bank had observed that a one-basis-point tolerance on either side of the bid/ask was too wide.
The CFTC found that the trader was aware of the one-basis-point tolerance, and that he generally marked his Closing Curves to keep them within one basis point of directly observable mid-market rates. By doing so, he was able to avoid a PnL adjustment by the internal control division. The alleged mismarking overstated the PnL of the Desk by approximately $25 million.
According to the Speaking Order, the trader generated his Closing Curves using a series of complex spreadsheets that he designed. The trader typically calculated the Closing Curve without input or assistance from others on the Desk. When he was out of the office, the junior traders on the Desk used the spreadsheets to generate the curves, but they did so by following specific instructions that obscured how the calculations resulted in the Closing Curve marks. As a result, the Closing Curves generated by the junior traders allegedly incorporated existing mismarks without the junior traders’ knowledge.
Discovery of the Conduct
In March 2018, the bank noticed a significant divergence between the Live Curve—which updated automatically based upon observable data—and the Closing Curve—which the trader had marked. While being questioned by his supervisors, the trader represented that these anomalies occurred within the past quarter, despite the trader having allegedly mismarking the curves since at least 2015. The trader did not inform his supervisors that he had been considering the potential impact on the Desk’s PnL when he calculated the Closing Curves.
Marking forward price curves is not an exact science. The quantity and quality of observable price data are directly related to market liquidity and publicly available price information. Less liquid markets require traders to use judgment about expected future market fundamentals, particularly further out the forward curve.
The CFTC’s Speaking Order reflects the difficulty in developing sophisticated controls to identify intentional mismarking of forward price curves. The bank’s internal controls were designed to discourage mismarking, because the bank’s policy was to adjust the Desk’s PnL in situations where any given point on the Control Curve differed from the Closing Curve by more than one basis point. In this circumstance, which involved a highly liquid market, a one-basis-point tolerance proved to be too wide to detect intentional mismarking, which enabled the trader to game the policy.
When considering internal controls for PnL adjustments, it is reasonable for a company to permit a tolerance between a trader’s marks and the available market price data. The tolerance should be calibrated to the specific market, with a focus on liquidity, the availability of observable price data, and reliability of prices provided by consensus pricing services (i.e., services that provide market data based upon aggregated price information). For markets with abundant observable price data, companies should consider setting a narrower tolerance band. In contrast, in illiquid markets with limited observable price data and a wider disparity of prices, a wider tolerance band may be appropriate.
Finally, the Speaking Order highlights the potential collateral consequences of intentional mismarking. Specifically, the CFTC noted that the trader’s mismarking caused his employer to provide inaccurate valuation data to a swap data repository, and separately, to provide swap counterparties with inaccurate daily marks and margin calculations. These type of inaccuracies expose a company to further regulatory action. Accordingly, a company that implements robust controls can capture benefits that extend beyond the direct focus of those controls.