The transition from LIBOR to risk-free rates was the most significant structural change in interest rate markets in decades. At its peak, LIBOR underpinned an estimated $300 trillion of financial contracts globally — from floating rate bonds and syndicated loans to interest rate swaps and trade finance facilities. When regulators concluded that LIBOR was no longer fit for purpose, the task of replacing it required the simultaneous re-engineering of market conventions, legal documentation, systems, and risk models across the entire financial system.

Why LIBOR Was Scrapped

LIBOR — the London Interbank Offered Rate — was a daily benchmark representing the rate at which panel banks said they could borrow from each other in the wholesale unsecured money market. It was published for five currencies (GBP, USD, EUR, CHF, JPY) and multiple tenors (overnight through twelve months). The critical flaw was structural: LIBOR was based on submissions from a panel of banks, not on actual transactions. In a world of abundant central bank liquidity and reduced interbank lending, the transactions underlying the submissions were too thin to support a reliable benchmark — banks were making educated guesses, not reporting actual trades.

This structural weakness created the opportunity for manipulation. The LIBOR scandal, which emerged in 2012, revealed that traders at multiple banks had systematically manipulated their LIBOR submissions to benefit derivatives positions or to project an appearance of creditworthiness during the financial crisis. Barclays, UBS, RBS, and others paid enormous fines; individuals were prosecuted. The FCA concluded that LIBOR's model — survey-based, panel-dependent, not anchored in observable transactions — was fundamentally unsound and could not be reformed to a standard the market could trust.

In 2017, Andrew Bailey (then FCA CEO) announced that the FCA would no longer compel panel banks to contribute to LIBOR after the end of 2021. This set the clock running on the transition.

SONIA: The Sterling Risk-Free Rate

The working group on sterling risk-free reference rates, convened by the Bank of England, selected SONIA — the Sterling Overnight Index Average — as the recommended alternative to GBP LIBOR. SONIA had been published since 1997; the Bank of England reformed it in 2018 to broaden its transaction base.

SONIA is calculated daily by the Bank of England as the trimmed mean of actual overnight sterling unsecured transactions reported to it by banks. Because it is anchored in real transactions — not estimates — SONIA is more robust than LIBOR. It is published each business day for the previous day's transactions, typically mid-morning.

The fundamental difference between SONIA and LIBOR is that SONIA is an overnight rate. LIBOR was a term rate — it gave a direct read on the expected cost of borrowing for specific tenors (three months, six months, one year). SONIA reflects only overnight borrowing costs. To use SONIA in a floating rate instrument, it must be compounded over the relevant interest period.

Compounding Conventions

The standard convention for SONIA-linked instruments is daily compounding in arrears. At the end of each interest period, the daily SONIA fixings for each business day in the period are compounded together to produce the period rate. For a three-month GBP floating rate bond paying SONIA + 50 basis points, the coupon for each period is calculated as:

Coupon = Notional × (Compounded SONIA rate + 0.50%) × Day Count Fraction

Where the Compounded SONIA rate is the product of daily (1 + SONIA/365) factors for each business day in the period, minus one.

Compounding in arrears creates an operational challenge: the final coupon amount is not known until near the end of the interest period, because the last SONIA fixings needed for the calculation are published only two to five business days before the payment date. This is a material change from LIBOR, where the coupon was known at the start of the period (LIBOR was fixed at the start). Borrowers and lenders who need advance notice of cash flows — for liquidity management, payment systems, or operational reasons — were disadvantaged by late knowledge of the coupon.

Observation Period Shifts and Lookback

To address this, market participants developed conventions that shift the observation period backwards by five business days, so that the SONIA fixings used in the calculation end five days before the coupon payment date. The Bank of England provides a version of SONIA with a five-day lag (SONIA compounded index) to facilitate this. The lookback convention — using fixings from a period ending before the payment date — has been widely adopted for floating rate notes and loans.

Term SONIA

Daily compounded SONIA addresses most of the transition use cases, but some market participants — particularly in the loan market — argued strongly for a term version of SONIA that, like LIBOR, would be fixed at the start of the interest period and known in advance. The ICE Benchmark Administration publishes Term SONIA Reference Rates (TSRR) for tenors of one, three, six, and twelve months, derived from SONIA OIS (overnight index swap) prices rather than from actual transactions.

Term SONIA is permitted for certain use cases — specifically floating rate loans and trade finance — but is not recommended for the derivatives market, where daily compounded SONIA is the standard. The FCA's guidance is that term rates should not be used where overnight compounded SONIA is a viable alternative, to preserve the liquidity and robustness of the overnight market.

ISDA 2020 Fallback Protocol

One of the largest legal challenges of the transition was the treatment of contracts that referenced LIBOR but had no adequate fallback language specifying what happened if LIBOR ceased to exist. Millions of legacy derivatives contracts fell into this category. ISDA's 2020 IBOR Fallbacks Protocol and Supplement addressed this: by adhering to the Protocol, counterparties agreed that, upon LIBOR cessation, their contracts would automatically convert to the relevant RFR (SONIA for GBP contracts) plus a credit spread adjustment — the fixed spread added to compensate for the difference between LIBOR (which included a credit component reflecting bank credit risk) and the near risk-free RFR.

The credit spread adjustments were calculated as the five-year median historical difference between LIBOR and the compounded overnight RFR over a lookback period ending on the date LIBOR cessation was announced. For three-month GBP LIBOR, this adjustment was approximately 11.93 basis points.

Tough Legacy Contracts

Despite widespread Protocol adherence and active transition programmes, a residual population of legacy contracts proved impossible to transition actively — either because the counterparty could not be located, or because the contract contained change-of-rate provisions that prevented unilateral amendment, or because the product was structured in a way that made amendment impractical. These were the "tough legacy" contracts.

The UK's Financial Services Act 2021 gave the FCA the power to require the LIBOR administrator to publish a synthetic version of LIBOR for a limited period after cessation, using the SONIA term rate plus the ISDA credit spread adjustment. Synthetic GBP LIBOR was published for one, three, and six-month tenors, providing a temporary fallback for genuine tough legacy contracts. The FCA signalled from the outset that synthetic LIBOR was a temporary measure, not a permanent replacement.

Operational Challenges

The LIBOR transition required banks to update every system that referenced LIBOR: trading systems, risk systems, collateral management systems, loan origination and servicing platforms, treasury systems, and regulatory reporting systems. For large banks with legacy technology stacks, this was a multi-year programme involving thousands of system changes. Risk model calibration also had to be redone — VaR models, CVA models, and pricing models that had been calibrated to LIBOR-based yield curves needed to be recalibrated to SONIA-based curves.