LIBOR manipulation, its death, and SOFR

Barclays admits it

On 27 June 2012, Barclays became the first bank to admit that it had manipulated the benchmark rate underlying trillions of dollars of contracts. It agreed to pay $200 million to the US Commodity Futures Trading Commission (CFTC), $160 million to the US Department of Justice, and £59.5 million to the UK Financial Services Authority, a combined total commonly reported as around $450 million. The settlement confirmed what had been rumoured for years: traders at major banks had been asking their own colleagues to nudge a benchmark interest rate, the London Interbank Offered Rate (LIBOR), in whichever direction suited their derivatives positions that day.

A rate built on opinions, not trades

LIBOR was never measured from actual transactions. Each morning, a panel of banks, typically 11 to 16 for a given currency and tenor, answered a survey question: at what rate could the bank borrow, unsecured, from another bank? The published rate was a trimmed average of these submissions. Nobody required that any real loan at that rate ever took place. Between roughly 2005 and 2009, traders at multiple banks routinely asked their firm's LIBOR submitters to move the number a fraction of a basis point in a direction that helped a swap or futures book. During the 2007 to 2009 financial crisis, some banks also lowballed their submissions to avoid looking financially weak in public, a subtly different form of the same problem.

The dominoes fall

Once Barclays admitted it, the rest followed. UBS settled on 19 December 2012 for a combined total exceeding $1.5 billion across the CFTC, the FCA, Swiss regulator FINMA and the US Department of Justice, with a Japanese subsidiary pleading guilty to a criminal charge. Royal Bank of Scotland settled on 6 February 2013 for a combined $612 million. On 4 December 2013, the European Commission fined eight banks and brokers a combined €1.71 billion for cartel conduct in Euro and Yen interest rate derivatives, its largest cartel fine at the time. Deutsche Bank's turn came on 23 April 2015, with a combined settlement of more than $2.5 billion, the single largest LIBOR-related penalty. Citigroup settled with the CFTC for $425 million on 25 May 2016, covering Yen LIBOR, Euroyen TIBOR and USD ISDAFIX conduct across two separate same-day orders. Across the whole industry, regulatory fines from the US, UK and EU are commonly summarised as exceeding $9 billion, and that figure is conservative: the six settlements above already sum to roughly $7.8 billion, before Rabobank, Lloyds, JPMorgan, Société Générale, brokers such as ICAP, and private litigation are added.

The deeper fault: a market that had gone quiet

Punishing the panel banks did not fix the deeper problem. After 2008, the unsecured interbank lending market that LIBOR was supposed to measure had mostly stopped trading. Banks were no longer lending each other large sums overnight without collateral, so even an honest submission had become a guess about a market that barely existed. On 27 July 2017, the FCA's chief executive Andrew Bailey announced, in a speech in London, that the regulator would no longer compel or persuade panel banks to keep submitting LIBOR after the end of 2021. A replacement rate was needed, and one was already being built. On 3 April 2018, the Federal Reserve Bank of New York published the first print of the Secured Overnight Financing Rate, SOFR, calculated not from a survey but from actual, high-volume Treasury repo transactions. The wind-down that followed took years: most LIBOR settings stopped after the end of 2021, the remaining US dollar panel settings ceased on 30 June 2023, and the last "synthetic" US dollar LIBOR rates, a temporary bridge calculated from SOFR plus a fixed spread rather than from bank submissions, were published for the final time on 30 September 2024, ending LIBOR entirely, 38 years after it began.

The reckoning, revisited

The story has one more twist, and it came years later. On 23 July 2025, the UK Supreme Court unanimously overturned the 2015 conviction of Tom Hayes, a former UBS and Citigroup trader who had become the public face of the scandal, along with that of former Barclays trader Carlo Palombo. The Court ruled that the original trial judge had misdirected the jury; it did not exonerate either man of wrongdoing, and the Serious Fraud Office said it would not seek a retrial. More than a decade after the first settlement, the legal chapter of the case was still not fully closed.

Why the mechanism worked

The economics of the case matter as much as the scandal itself. LIBOR was the floating-rate leg of the vast majority of interest rate swaps written anywhere in the world, the reference rate for Eurodollar and LIBOR futures, and the rate used to set countless forward rate agreements and floating-rate loans. A trader who received floating LIBOR on a swap wanted the fixing set high; a trader who paid floating wanted it set low. Because rate submitters often sat close to the derivatives desks whose books depended on the fixing, a trader could ask a colleague to move the number by a fraction of a basis point on a day that mattered for a large position. On its own, that looks like rounding. Multiplied across a notional base measured in the hundreds of trillions of dollars, it becomes a serious and detectable form of fraud.

Replacing LIBOR also meant changing how a floating-rate contract works, not just relabelling the number. LIBOR was set at the start of an interest period, so a borrower or swap counterparty knew the floating payment in advance. SOFR-based rates are typically compounded from realised overnight rates and are only known at the end of the period. That is a genuine change in mechanics, and it is why the transition needed new fallback language in existing contracts, new discounting conventions, and new futures contracts: SOFR futures at the CME replaced Eurodollar futures as the standard way to hedge short-term US dollar interest rate risk.