A scramble to replace LIBOR is under way

SITTING in his office in the Wrigley Building overlooking the Chicago river in 2012, Richard Sandor, who has spent his career inventing financial products, was reading about the scandals surrounding the London Interbank Offered Rate (LIBOR), an array of interest rates set daily by a club of banks in Britain and used to price trillions of dollars’ worth of loans, derivatives and more. “This is stupid,” Mr Sandor recalls saying to a colleague. “Let’s make a bet; LIBOR will lose its pre-eminence.”

Two years later the Federal Reserve reached the same conclusion. It formed a group, the Alternative Rate Reference Committee, which has created a new benchmark dollar interest rate, the Secured Overnight Financing Rate (SOFR). Since April, SOFR has been used for a handful of bond offerings by large institutions including the World Bank, MetLife and Fannie Mae. Central banks in Britain, the euro zone, Japan and Switzerland are also constructing new benchmark rates.

LIBOR is heading for extinction. Its fate was sealed in July 2017 when Andrew Bailey, head of Britain’s Financial Conduct Authority, a regulator, said it would be phased out in 2021. It had been undermined by twin scandals. In the first, a product of the crisis, the rate-setting banks tweaked their quotes, possibly with supervisors’ implicit support, to limit the chances of market panic. In the second traders manipulated the rates subtly, to gild their profits.

The ruckus cost Bob Diamond, the chief executive of Barclays, his job and Tom Hayes, a trader at UBS and Citigroup who was jailed for 11 years, his liberty. Oversight of LIBOR was transferred from the British Bankers Association, a trade body, to British regulators and then to Intercontinental Exchange, an American stock- and derivatives-exchange group. In June Société Générale agreed to pay American authorities $750m to settle a charge of manipulation, adding to a list of seven other big banks. (The French bank also agreed to pay a large sum to settle charges related to a bribery scheme in Libya.)

From fiction to friction

LIBOR also rests on shaky economics. Its roots go back to an informal coalition of London banks in the 1960s. This was formalised into a panel of 20, which submitted daily estimates of their borrowing costs for up to five currencies and seven maturities of up to a year. Yet some quotes are little better than guesses. In July Randal Quarles, the vice-chairman of the Fed in charge of bank supervision, noted that just six or seven transactions a day were used to set one- and three-month dollar LIBOR and an average of one for the 12-month rate, for which on “many days there are no transactions at all”. A few banks have dropped out of the panel; some are staying until 2021 only at the FCA’s request.

On this flimsy foundation a staggering $260trn-worth of financial products, from interest-rate swaps to retail mortgages, are priced, estimates Oliver Wyman, a consulting firm. Dollar LIBOR accounts for by far the biggest chunk, not far short of $200trn; sterling and yen weigh in at $30trn apiece and Swiss francs at $5trn. The chief benchmarks for euros, EURIBOR and EONIA, face an even tighter timetable for reform and replacement than LIBOR. (EONIA does not comply with a recent European Union directive and must go by the end of 2019.)

Creating and then switching to truly market-based alternatives is an almighty task. The Fed’s approach was to tap into the “repo” (repurchase) market. Banks seeking short-term cash sell securities with little credit risk, such as Treasuries, to other banks with a promise to buy them back the next day at a slightly higher price. The difference is in reality the interest rate on an overnight loan. To ensure repayment, they provide collateral. There are $700bn-worth of these transactions daily, which are reported to the Fed through the Depository Trust & Clearing Corporation and the Bank of New York Mellon. After ingesting and processing vast quantities of data to produce a weighted average, the Fed publishes the result, SOFR, at 8am.

Take-up has been slow. So far only seven or eight bonds have been sold using SOFR as a reference price. Doubtless this is partly because of investors’ unfamiliarity with a new product, compounded by the Fed’s inability to explain itself to those who do not understand its jargon. But it may also reflect difficulties with using overnight, near-risk-free rates.

For a start, an overnight rate is exactly that: a term structure has to be constructed for longer maturities, for instance from expected or actual overnight rates. SOFR also reflects the rate on extraordinarily high-quality, essentially risk-free credits, which would default only if America’s government failed. Using it as a benchmark may therefore risk creating a mismatch for the average bank. In a crunch, SOFR may fall as investors run for safe assets, pushing down banks’ revenues from SOFR-linked loans. Yet banks’ own borrowing costs on wholesale markets will increase.

In addition, legal problems loom, and time is short. Contracts continue to be written on LIBOR, of which plenty extend beyond 2021. The Bank of England noted in June that the number of such LIBOR-linked sterling derivatives had risen since the previous year. Many contracts, the bank went on, lack “fallback” clauses setting out which rate applies once LIBOR goes. British regulators wrote to banks on September 19th instructing them to provide by December a summary of their plans for mitigating LIBOR-related risks.

Meanwhile Mr Sandor has developed his own benchmark, which is steadily attracting customers. By 2015 he had convinced a handful of small banks to join his new American Financial Exchange, which now has 99 members and where $1bn-worth of loans are traded daily. From those transactions, a benchmark overnight interest rate for unsecured loans, Ameribor, has been derived.

This month Ameribor was used for the first time in pricing a loan, by ServisFirst Bank of Birmingham, Alabama, to a car dealer in Tennessee. The bank’s chief executive, Tom Broughton, says that it considered SOFR, but because it does not use Treasury repos and its liabilities are not secured, it needs a rate that can accommodate credit risk. Mr Sandor hopes that in two to three years Ameribor will become a benchmark for many of America’s 5,000 regional and community banks and their customers. Whether or not that happens, the era of LIBOR is ending.