BY TERRY BELTON
GLOBAL HEAD OF FIXED INCOME AND FX RESEARCH, JPMORGAN
AND
ALEX ROEVER
HEAD OF SHORT DURATION STRATEGY, JPMORGAN

In 1984, the British Banker's Association (BBA) began producing a survey of interbank lending rates within the London market. The resulting London Interbank Offer Rate (LIBOR) has since grown from an arcane gauge of bank lending into one of the most widely used global interest rate benchmarks. It is the central metric for the world's money, credit and interest rate derivatives markets.

However, certain aspects of LIBOR are widely misunderstood and have been illuminated by the intensity of the global liquidity crisis. Some believe LIBOR is too low relative to actual bank borrowing rates due to systemic bias on the part of contributors to under-report their actual borrowing costs. As evidence, these critics point to the lower level of daily LIBOR postings relative to other objective measures of bank borrowing costs. Critics have even suggested that LIBOR be overhauled or reformulated to more accurately refl ect the state of the markets.

UNDERSTANDING LIBOR

LIBOR is like a Rorschach test. Different people can look at the same inkblot on a card and see wildly different things. Similarly, people can see very different things based on LIBOR.

Unlike many other interest rate benchmarks, LIBOR is not an observable traded commodity. Although banks regularly lend excess reserves and other funds to one another, interbank trades are private, bilateral transactions. Eurodollar deposits - dollar-denominated deposits held at a foreign bank - generally are considered the most similar observable commodity. Interbank loan rates refl ect banks' willingness and ability to lend to one another, while Eurodollar and other time deposits refl ect the credit market's willingness to lend to banks. The BBA attempts to replicate the yields at which banks are willing to lend to each other by surveying banks on each London business day. It conducts surveys across 10 different currencies, asking a specifi ed panel of banks active in each market to estimate borrowing rates for a series of maturities ranging from overnight to one year. The banks are under no obligation to prove that they can actually trade at those yields. The BBA compiles the contributed data, and then calculates the average level for each maturity using a trimmed-mean process intended to screen out abnormally high or low rates.

In the case of U.S. dollar (USD) LIBOR, the contributor panel consists of 16 banks from seven nations, including many of the world's largest banks (see table). For each maturity, BBA will exclude the highest and lowest quartile of rates, and calculate LIBOR using a simple average of the remaining contributions. The BBA publishes the results shortly after 11:30 a.m. London time.

For several years preceding the current liquidity crisis, the spread between the highest and lowest contributed data was relatively tight, as the credit environment was benign. For example, using "untrimmed" data from the USD LIBOR panel, the average distance between the highest and lowest contributed rates for the three-month LIBOR during the fi rst four months of 2007 registered about 1.5 basis points, with the largest difference reaching 2 basis points. In contrast, the average spread between the high and low quotes during the fi rst four months of 2008 was 5.9 basis points, with the widest one-day difference hitting 17 basis points.

BUILDING A BETTER LIBOR

Could BBA LIBOR somehow be changed so that it becomes a better, or at least more robust, measure of bank borrowing levels? Any potential changes should be contemplated with two goals in mind: fi rst, to avoid a discrete change in long-term LIBOR expectations that could disrupt the interest rate derivatives markets, and second, to maintain the perception that LIBOR represents the unsecured borrowing rate of a prime quality bank (generally a low double-A or better rating profi le). Including non-representative banks could detract from the index's stability.

The following represent a few of the most common, and potentially most practical, proposals:

Increase the size of the panel. Increasing the number of USD panel participants might have the joint effect of making the panel less U.K.-focused and provide a better global cross section.

The BBA would need to exercise care in expanding the group, as potential U.S.-based additions like Wachovia or Wells Fargo have signifi cant retail funding bases, and as a result could be less aggressive in pursuing interbank deposits. If not offset with other wholesale-funded institutions, the revised postings could gravitate lower, exacerbating the perception that USD LIBOR is too low.

Change the time of the quote. There is often a drop in rates when the U.S. markets open, refl ecting a surge in activity resulting from U.S. money markets where the majority of trading takes place very early in the morning. Delaying the LIBOR announcement until 10:00 a.m. New York time could allow key economic data to price into the Eurodollar deposit market.

On the negative side, aligning the setting time with the U.S. markets would demand late-day processing and settling of trades for the 13 of the 16 current panel members that are domiciled outside the U.S. In addition, smaller, mostly European banks that borrow heavily from the U.S. may be left with unrolled funding and few other funding alternatives late in the day. In a similar vein, the BBA recently declined to add a second quote when U.S. markets open, noting that it would add to market confusion.

Change the calculation to use a median rather than a trimmed mean. The idea here is to use a more representative number in the rate determination process, rather than an average. However, the actual statistical difference between median and average may be fairly minimal, even in periods like 2008 where the dispersion between high and low postings has been signifi cant.

Change the survey question. Critics believe the current definition ("the rate at which the contributor could borrow funds, were it to do so by asking for and then accepting interbank offers in reasonable market size just prior to 11:00 a.m. London time") results in too low a LIBOR. To provide a fairer picture of the level where banks are willing to transact with one another, the critics would prefer something like "the rate at which interbank term deposits within the euro area are offered by one prime bank to another prime bank." Whether or not survey wording would generate fairer results is questionable. Moreover, any defi nitional change large enough to move the market could actually change long-term LIBOR expectations and, therefore, violate our fi rst goal outlined above.

The market is smart and understands what LIBOR is: a benchmark used to price credit. The market then makes its own adjustments to the benchmark - adding or subtracting basis points - to reflect its own value judgments. The BBA should be free to restructure LIBOR if it is absolutely necessary, but some of the proposed changes amount to no more than tinkering. The markets would be best served if the BBA continues to let the market make its own adjustments.

A BENCHMARK FREED

Most of the criticisms recently leveled at BBA LIBOR are a reflection of extreme market conditions rather than fatal flaws in the benchmarks construction. It is debatable if one could design a less fl awed benchmark, and it could not effectively replace LIBOR anyway. From the banks, to the money markets, to the interest rate swaps market, to the Eurodollar futures pits, LIBOR is baked into the global fi nancial system. It is clear that the Federal Reserve and other central banks comprehend the special role LIBOR plays in the market, and it is positive that they are taking steps to address current market concerns. Conditions in the LIBOR market should gradually improve in the coming months. As they do, most of the current criticisms of BBA LIBOR will likely be forgotten.

Back to Summer 2008 Issue Home