What Is LIBOR? How It Was Calculated and What Replaced It

Search Fintrev

Search our guides, reviews and glossary.

LIBOR, short for the London Interbank Offered Rate, was a benchmark interest rate that estimated what large banks would pay to borrow unsecured funds from one another for short periods. It was published every London business day for several currencies and maturities. Floating-rate loans, bonds, mortgages and derivatives were priced against it, usually as “LIBOR plus a margin.” LIBOR has been discontinued. Panel publication ended in stages between December 31, 2021, and June 30, 2023, and the last temporary “synthetic” settings stopped in September 2024. In U.S. dollar markets its main replacement is the Secured Overnight Financing Rate (SOFR). Other currencies moved to their own overnight benchmarks.

What LIBOR Measured

LIBOR was meant to capture the cost of unsecured wholesale funding for large, highly rated banks active in London. Unsecured means no collateral backed the loan. A bank lending to another bank at LIBOR was taking the borrower’s credit risk, so the rate contained two things: the general level of short-term interest rates in that currency, and a premium for the chance that a bank might fail to repay.

That second component is what separated LIBOR from government or central bank rates. When banks trusted each other, the premium was small. When trust broke down, LIBOR rose even if the central bank’s policy rate stayed put.

Each LIBOR rate was identified by currency and tenor (maturity). The final seven tenors were overnight (spot next for some currencies), one week, and one, two, three, six and 12 months. A “3-month USD LIBOR” figure was the estimated annualized rate for borrowing U.S. dollars for three months. That setting was the most widely referenced and was often called simply “LIBOR.” Rates were quoted on an annual basis even for short tenors. For USD the interest calculation used an actual/360 day count, while sterling used actual/365.

How LIBOR Was Calculated

The panel and the submission

For each currency, the administrator kept a panel of contributor banks chosen for their activity in the London money markets. Panel membership was reviewed periodically. The U.S. dollar panel in later years had 16 banks, including Bank of America, Barclays, Citibank, Deutsche Bank, JPMorgan Chase and UBS.

Each morning, every panel bank answered a version of the same question: at what rate could it borrow funds in a reasonable market size, by asking for and accepting interbank offers, just before 11 a.m. London time? The question asked about the bank’s own borrowing cost, not what it would charge others. It also asked what the bank could do, not what it did. On days with few real trades, the answer was an informed estimate.

Submissions were collected, processed, and the rates were published once a day, around 11:55 a.m. London time in the ICE era.

The trimmed mean

The administrator did not take a simple average. It ranked the submissions for each currency and tenor, discarded the highest and lowest quartiles, and averaged the rest. This is a trimmed mean (also called an interquartile mean). It is designed to blunt the effect of an outlier or a single bank trying to drag the rate.

Hypothetical example: a 16-bank panel. Suppose the 16 banks on a USD panel submitted these 3-month rates, already sorted from low to high:

5.10, 5.12, 5.13, 5.14, 5.15, 5.15, 5.16, 5.17, 5.17, 5.18, 5.19, 5.20, 5.22, 5.25, 5.30, 5.60

1. Drop the lowest four (5.10 to 5.14) and the highest four (5.22 to 5.60).
2. Add the middle eight: 5.15 + 5.15 + 5.16 + 5.17 + 5.17 + 5.18 + 5.19 + 5.20 = 41.37.
3. Divide by eight: 41.37 ÷ 8 = 5.17125%.

The published figure would be 5.17125%. The 5.60% submission had no effect because it fell in the top quartile. That is the method’s strength against a lone outlier, and also its weakness. A bank whose submission stays inside the middle band moves the result by one-eighth of any change it makes. If several banks shade their submissions in the same direction, the whole middle band shifts and the trimming offers no protection. The rigging cases turned on that weakness.

The waterfall methodology

After the scandal, the administrator tried to anchor submissions to real trades. In April 2018 the ICE Benchmark Administration (IBA) proposed a standardized waterfall methodology, which it then rolled out across the remaining currencies. Each bank worked down three levels:

• Level 1, transaction-based. A volume-weighted average of the bank’s eligible unsecured funding transactions, with more weight on trades booked closer to 11 a.m. London time.
• Level 2, transaction-derived. If the bank lacked enough eligible trades for Level 1, it submitted a rate derived from transaction data using set methods, such as interpolation between tenors or adjustments to recent trades.
• Level 3, expert judgment. If neither of the first two levels was possible, the bank submitted the rate at which it judged it could fund itself at 11 a.m. in the unsecured wholesale market, using its own documented methodology.

The trimmed mean was still applied to the final submissions. The waterfall improved the process, but it could not create transactions that did not exist. Interbank term lending had shrunk after 2008, and many submissions continued to rely on the lower levels.

Interpreting LIBOR

A LIBOR reading combined expected policy rates with bank credit and liquidity conditions, so analysts used it in two ways.

As a rate level, it showed the cost of short-term money. A falling LIBOR made floating-rate borrowing cheaper and pointed to easier credit conditions. A rising LIBOR made borrowing more expensive and could slow lending and business activity.

As a spread, it showed stress in the banking system. Two comparisons were widely followed:

• The TED spread: 3-month USD LIBOR minus the 3-month U.S. Treasury bill yield. Treasuries carry essentially no default risk, so a wider gap meant lenders were demanding more to lend to banks than to the government.
• The LIBOR-OIS spread: 3-month LIBOR minus the overnight indexed swap rate for the same term, which reflects expected policy rates. Because the expected path of policy rates is removed, what remains is mostly bank credit and liquidity risk.

Both spreads widened sharply during the 2007 to 2008 financial crisis, when banks hoarded cash and doubted each other’s solvency. That sensitivity was useful for reading the market. It became a problem when the same rate also set millions of loan payments.

How LIBOR Was Used

LIBOR sat underneath a wide range of products:

• Interbank and derivatives products: interest rate swaps, forward rate agreements (FRAs), interest rate futures such as Eurodollar futures, caps, floors, options and swaptions.
• Commercial lending and securities: floating-rate notes, floating-rate certificates of deposit, corporate loans and syndicated loans (loans funded by a group of lenders).
• Structured and hybrid products: collateralized debt obligations (CDOs), collateralized mortgage obligations (CMOs), and various accrual, callable and perpetual notes.
• Consumer credit: adjustable-rate mortgages (in the U.S., often tied to 1-year LIBOR), private student loans and some other variable-rate loans.

Beyond direct pricing, LIBOR served as a discount curve for valuing future cash flows, a reference for clearing and settlement, and a gauge of market expectations for central bank rates. U.S. dollar LIBOR alone was referenced by contracts estimated in the hundreds of trillions of dollars of notional value, which made its replacement slow and complex.

Examples of LIBOR-Based Products

A floating-rate note

Hypothetical example. A company issues a $1,000,000 note paying 3-month USD LIBOR + 0.50%, reset quarterly, with interest calculated on an actual/360 basis.

• At the first reset, 3-month LIBOR is 2.00%. The coupon rate is 2.50%. For a 91-day period: $1,000,000 × 2.50% × 91/360 = $6,319.44.
• At the next reset, LIBOR has fallen to 1.50%, so the coupon rate is 2.00%. For a 92-day period: $1,000,000 × 2.00% × 92/360 = $5,111.11.

The margin (0.50%) stayed fixed for the life of the note and reflected the issuer’s own credit. Only the LIBOR component moved. The same structure applied to a corporate loan or an adjustable-rate mortgage, where the margin was set in the contract and the index reset on a schedule.

An interest rate swap

Hypothetical example. Two parties agree on a swap with a $10,000,000 notional amount. Party A pays a fixed 2.25% a year and receives 3-month LIBOR. Party B takes the other side. Payments are quarterly and netted, and for simplicity each quarter is treated as 90/360 (one-quarter of a year).

• Quarter 1, LIBOR = 2.75%. Floating leg: $10,000,000 × 2.75% × 0.25 = $68,750. Fixed leg: $10,000,000 × 2.25% × 0.25 = $56,250. Party B pays Party A the net $12,500.
• Quarter 2, LIBOR = 1.75%. Floating leg: $43,750. Fixed leg: $56,250. Party A pays Party B the net $12,500.

The notional never changes hands. A company with a LIBOR-based loan could use the receive-floating side to offset its loan payments, which turned its floating cost into an effectively fixed one. An investor holding fixed-rate assets could do the reverse to gain floating exposure.

A Brief History of LIBOR

In the early 1980s, a fast-growing market in syndicated loans, swaps and other interest rate products needed a common reference rate. The British Bankers’ Association (BBA) introduced BBA interest-settlement rates in 1984, and these were formalized as BBA LIBOR in 1986. It quickly became the default benchmark for interbank and cross-border lending.

At its widest, LIBOR was published for 10 currencies and 15 maturities, about 150 rates a day. The arrival of the euro in 1999 removed several European currency settings. After the scandal, the 2012 Wheatley Review recommended cutting rates with too little underlying trading, and the list was reduced to the five-currency, seven-tenor set of 35 rates.

The UK’s Financial Conduct Authority began regulating LIBOR in 2013. In February 2014 administration passed from the BBA to ICE Benchmark Administration, a subsidiary of Intercontinental Exchange, and the rate became ICE LIBOR.

The LIBOR Rate-Rigging Scandal

Because submissions were estimates, a bank could shade them without any trade contradicting it. Investigations found two main patterns.

The first was trader-driven manipulation. Derivatives traders asked their banks’ submitters to nudge rates up or down on specific days to benefit positions tied to the fixing. Some coordinated with traders at other banks and with interdealer brokers.

The second was lowballing during the financial crisis. Some banks submitted rates lower than their true borrowing costs so they would not look weak. A high submission could signal that a bank was struggling to raise funds.

Press reporting in 2008, including in The Wall Street Journal, questioned whether LIBOR reflected real borrowing costs, and regulators in the U.S., UK and Europe opened investigations. In June 2012 Barclays settled with U.S. and UK authorities, the first major resolution, and the scandal became public at scale. Barclays, UBS, Royal Bank of Scotland, Rabobank, Deutsche Bank and broker ICAP were among those fined, and individual traders and brokers faced prosecution.

Why LIBOR Was Discontinued

Manipulation was the trigger, but the deeper problem was structural. After 2008, banks relied far less on unsecured term borrowing from each other. A benchmark meant to measure an active market was increasingly measuring one with few trades, so it depended on judgment however the submission process was policed.

In July 2017 the FCA said it would no longer compel panel banks to submit after the end of 2021. On March 5, 2021, it confirmed the end dates:

• After December 31, 2021: all sterling, euro, Swiss franc and Japanese yen settings, plus the one-week and two-month USD settings, ceased or lost representativeness.
• After June 30, 2023: the remaining USD settings (overnight, one, three, six and 12 months) ended as panel-based rates.

To protect contracts that could not be amended in time, the FCA allowed temporary synthetic LIBOR for some settings, calculated from the replacement rate plus a fixed spread rather than from bank submissions. Synthetic sterling and yen settings were wound down first. The last synthetic USD settings (one, three and six months) ended in September 2024. No LIBOR rate is published today.

What Replaced LIBOR

Each currency moved to a nearly risk-free overnight rate based on actual transactions:

• U.S. dollar: SOFR (Secured Overnight Financing Rate), published by the Federal Reserve Bank of New York, measures the cost of borrowing cash overnight against U.S. Treasury collateral in the repo market.
• British pound: SONIA (Sterling Overnight Index Average), administered by the Bank of England.
• Euro: €STR (euro short-term rate), published by the European Central Bank.
• Japanese yen: TONA (Tokyo Overnight Average Rate).
• Swiss franc: SARON (Swiss Average Rate Overnight).

In the U.S., some lenders also adopted credit-sensitive alternatives such as Ameribor, which reflects unsecured borrowing costs across a broad group of American banks and financial institutions. SOFR, however, became the market standard and the legal default.

Legacy contracts and the LIBOR Act

Contracts written before the transition often had fallback language that broke down without LIBOR. Many simply referred to “the last published rate,” which would have frozen a floating loan into a fixed one. Market participants used three routes:

1. Contractual fallbacks. Derivatives documentation from ISDA added fallbacks to compounded overnight rates plus a fixed spread adjustment. That adjustment was set as the five-year historical median difference between LIBOR and the replacement rate for each tenor.
2. Renegotiation, where parties could agree on new terms.
3. Legislation. The U.S. Adjustable Interest Rate (LIBOR) Act, signed in March 2022, replaced USD LIBOR in contracts without workable fallbacks. The Federal Reserve then selected SOFR-based benchmarks, with tenor spread adjustments, as the replacements under the law.

The spread adjustment matters for borrowers. SOFR runs below LIBOR because it carries no bank credit premium, and longer LIBOR tenors also included a term premium. Without an adjustment, switching a loan from LIBOR + 2% to SOFR + 2% would have quietly cut the lender’s return. The adjustment closes that historical gap so the switch does not transfer value between the parties on day one.

LIBOR vs. SOFR and Other Rates

LIBOR vs. SOFR. The two rates differ in how they are built and what they measure:

• Source: LIBOR came from bank submissions, often estimates. SOFR is calculated from a large volume of actual overnight repo trades.
• Security: LIBOR measured unsecured borrowing. SOFR measures borrowing secured by Treasuries, so it contains almost no credit risk.
• Tenor: LIBOR was published for terms up to 12 months and was known at the start of each interest period. SOFR is an overnight rate. Loans use either an average or a compounded SOFR, often calculated over the period and known only at its end, or a forward-looking Term SOFR derived from SOFR futures markets.
• Stress behavior: LIBOR tended to rise when banks were under stress. SOFR can fall in a crisis, because demand for Treasury-backed safety rises.

LIBOR vs. the fed funds rate and prime rate. The federal funds rate is the overnight rate for unsecured lending of reserves between U.S. banks, and the Fed targets it as a policy rate. The prime rate is a rate banks set for lending to customers, usually moving with the fed funds target. LIBOR was neither a policy target nor a lender’s posted rate. It was an average of submitted funding costs in London, set across several currencies.

LIBOR vs. EURIBOR, TIBOR, HIBOR and MIBOR. The Euro, Tokyo, Hong Kong, Shanghai and Mumbai interbank offered rates are separate IBORs with their own administrators, panels and rules, and they did not end simply because LIBOR did. EURIBOR, for example, was reformed with a hybrid methodology and continues alongside €STR, and HIBOR continues in Hong Kong. Each benchmark’s status depends on its own regulator.

What LIBOR’s End Means for Borrowers

If you hold an adjustable-rate mortgage, student loan or business loan written before the transition, it most likely moved to a SOFR-based index through its fallback clause or the LIBOR Act. The contract margin generally stayed the same, with the spread adjustment added to the new index, and the rate still resets on the schedule in your loan documents.

Two practical differences can follow. With an average or compounded SOFR, the rate for a period may reflect conditions during that period rather than a figure fixed at its start. And a period of banking stress no longer pushes up your index the way it could under LIBOR. Your loan notices and servicer disclosures name the exact replacement index and adjustment that apply.

The Bottom Line

Anyone holding an adjustable-rate mortgage, student loan or business loan written before the transition should know it most likely moved to a SOFR-based index. The contract margin generally stayed the same, with a spread adjustment added so the switch did not shift value on day one. Loan notices and servicer disclosures name the exact replacement index and adjustment that apply.