Goldman Sachs: Trades Designed to Get the Benchmark Print

The Goldman Sachs ISDAFIX CFTC settlement required a $120 million civil penalty over allegations that traders tried to move an interest-rate benchmark and submitted false rates to benefit the bank’s positions. The Commodity Futures Trading Commission alleged that traders targeted the 11 a.m. fixing through swap spreads, U.S. Treasuries and Eurodollar futures. For someone who knows the desk, the revealing connection is between the instruction to move a rate and the separate position that would profit when it moved.

The commission announced the settlement on December 21, 2016. Its action covered The Goldman Sachs Group, Inc. and Goldman, Sachs & Co., the group’s banking, securities and investment management subsidiary. The alleged conduct ran from January 2007 through March 2012 and involved New York traders across the bank’s interest-rate business, including the head of its U.S. Interest Rate Products Trading Group, according to the commission.

This CFTC enforcement action shows how a published reference rate allegedly became the target of the desks whose contracts depended on it. The agency’s account connects the broker’s screen, the bank’s submissions and the settlements they were allegedly meant to improve.

What the ISDAFIX Print Meant

The ISDAFIX print was the snapshot used to establish a benchmark for interest-rate swaps. The dollar benchmark was supposed to represent the prevailing market rate for the fixed leg of a standard fixed-for-floating swap at 11 a.m. Eastern Time. That rate helped determine swaption cash settlements, swap-futures settlements and the value of other interest-rate products.

The commission’s order describes the machinery. At the swaps broker, dealers’ brokers called out bids and offers across a shared desk while connected to their clients by direct phone lines and speaker boxes. An employee manually changed the swap-spread levels on the live 19901 screen to reflect that trading. The screen also displayed Treasury yields and prices from the broker’s electronic platform.

For most maturities, the broker combined a swap spread with the corresponding Treasury yield to generate a reference rate. The one-year rate worked differently: it drew on Eurodollar futures yields and the broker’s assessment of prevailing short-term swap rates. Moving a Treasury price, a swap spread or a futures input could therefore affect the rate used elsewhere.

After taking the snapshot, the broker sent reference rates to a panel of banks. Each bank’s submission was supposed to reflect the midpoint of where it would bid and offer a swap to a dealer of good credit. Most banks normally accepted the broker’s figures. According to the order, those reference rates usually became the published benchmark. That made the screen at the fixing time particularly valuable to anyone trying to influence the result.

The consequences reached beyond dealer trading books. The commission identified pension funds and local and state governments among the users of ISDAFIX-based instruments to hedge changes in interest rates. A reference rate meant to help manage risk was also a number against which money changed hands.

Why Goldman Wanted the Rate to Move

The profitable direction depended on the position. Goldman’s Swap Desk held swaps and swap futures, while its nearby Volatility Desk held swaptions and more complex interest-rate derivatives. The order alleged that traders tried to influence ISDAFIX to improve payments or reduce losses on those positions.

Internal trades connected the desks. The Volatility Desk often transferred its exposure from cash-settling swaptions to the Swap Desk through an internal swap. If the Swap Desk would receive the ISDAFIX rate, a higher fixing helped it. If it would pay the rate, a lower fixing helped. The commission alleged that these internal transfers gave the Swap Desk an incentive to move a benchmark that also determined what Goldman’s outside counterparties received or paid.

Swap futures supplied another incentive. Their value moved inversely to the relevant ISDAFIX rate at settlement. A net long position benefited from a lower rate, while a net short position benefited from a higher one. A desk could therefore want different maturities to move in opposite directions on the same day.

The order alleged that traders were sometimes willing to buy higher or sell lower than the market required because they expected the benefit on cash settlements to cover or exceed the trading loss. We see the problem: a losing trade could still be a winning strategy for the desk. The payoff sat in another contract.

That connection is central to commodity market manipulation across linked products. The 2024 working paper An Analysis of Market Manipulation Definitions Around the World, developed as part of a World Federation of Exchanges research project, describes strategies spanning instruments, counterparties and desks, with profits taken elsewhere. Its review distinguishes trading that changes a price from false submissions that feed a benchmark. Goldman’s alleged conduct involved both.

How Goldman Sachs Traders Tried to Get the ISDAFIX Print

For Goldman Sachs’s ISDAFIX traders, to get the print meant securing a fixing favorable to the positions settling against it, according to the CFTC’s account. The order alleged that traders timed transactions to be the last event moving the screen before the snapshot. Some instructions specified the desired rate at the fixing rather than an amount the trader wanted to buy or sell.

In one recorded conversation described in the order, a trader asked the broker to arrange for the screen operator to delay reporting changes in ten-year prices that went against Goldman’s interest until after the fixing. The trader then followed up by email:

The instruction makes the alleged objective unusually clear. The trader wanted control over the number other contracts would use, and was prepared to spend money in the input market to get it. The broker’s role mattered because the trading and the displayed level had to meet at the same moment.

Eurodollar Futures Fed the One-Year Rate

The order’s September 2009 example shows traders working backward from a desired settlement. Goldman would pay less on a cash-settling straddle swaption if the one-year ISDAFIX rate came out lower. A swaps trader asked the broker how that rate was set.

According to the commission, the broker explained that trading Eurodollar futures could influence the fixing more easily than trading one-year swaps. The trader then sold one-year swaps shortly before the fixing to affect the broker’s assessment of the market and traded Eurodollar futures in a direction intended to put downward pressure on the rate. The alleged strategy reached across both inputs because the eventual payment depended on their combined result.

A Submission Chosen for the Position

The false-reporting allegation concerned a different decision: what number Goldman submitted as its view of the market. The commission alleged that traders instead selected rates to favor the bank’s derivatives positions.

In the order’s February 2009 example, Goldman stood to receive more from pension funds and other clients on cash settlements of five-year straddle swaptions if the benchmark fixed higher. The CFTC alleged that the Volatility Desk asked the submitter for a higher rate, and the submitter complied.

The order also described requests made before the fixing time. The agency’s point was concrete: a request driven by a settlement position, made before the market snapshot, could not represent an honest reading of that later snapshot. Goldman allegedly used trading and submissions separately and together to push the benchmark in the direction its positions needed.

What Evidence Is Required to Prove Fraud?

Emails, audio recordings, trades and benchmark submissions supplied the evidence in this action. The commission connected them to the positions the traders allegedly wanted to benefit. Those connections explain what the trading was meant to accomplish. The instruction gives the desired direction, the position supplies the financial reason, and the timing connects the conduct to the fixing.

The records tell different parts of that story. A broker conversation can explain why a particular market was chosen. An internal swap can show which desk took on the benchmark exposure. The submission and the desk’s message can show whether the submitted rate followed a market assessment or a request for a better settlement. We examine those connections with the person who understands what the records meant inside the business. The Commodity Futures Trading Commission’s Division of Enforcement investigated the Goldman conduct.

Knowledge can also sit outside the trading desk. The CFTC whistleblower program accepts information from customers and market observers as well as employees. Someone who understands the pricing process can explain why a trade, a message and a settlement belong to the same story.

What the CFTC Order Required

The settlement paired the civil penalty with an order to stop the violations and improve benchmark controls. The commission alleged basic gaps: Goldman had no specific procedures for determining or monitoring ISDAFIX submissions, gave submitters no formal training and did not require submissions to be documented.

In the order, Goldman represented that it had introduced senior approval of benchmark contributions, records of contributions and their methodology, and reviews of potential conflicts. Those controls addressed the same connection the allegations exposed: the person helping supply the benchmark could have a trading interest in its direction.

The order required improvements in monitoring, audits and training where needed. The obligations reached the people overseeing the desks, as well as the people submitting the rates.

The Same Benchmark, More Than One Bank

Goldman’s settlement was the CFTC’s third ISDAFIX enforcement action. Earlier actions against Barclays and Citibank also concerned alleged efforts to influence the fixing through trading and submissions for the benefit of derivatives positions. The recurring conflict was the same: banks contributing to a benchmark also held contracts whose payments depended on it. The pattern continued in the Royal Bank of Scotland’s subsequent ISDAFIX order.

Whistleblower Rewards Depend on the Program

The Commodity Exchange Act provides CFTC whistleblower awards totaling 10% to 30% of sanctions collected in qualifying actions. Goldman’s benchmark action arose under that law.

A whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers. That is a separate program. We work with you to identify the program and the case your information can support.

You Know Why the Desk Wanted That Rate

If you know which position a fixing instruction was meant to benefit, speak to one of our whistleblower experts about the fixing you saw. We connect your understanding of the desk with the trades, submissions and settlements, and bring in CFTC whistleblower lawyers who understand benchmark manipulation. We stay with you from the first call, building your strongest CFTC whistleblower case and standing beside you through it.