Nothing like putting your foot in mouth in public and chewing.
Avinash Persaud, who is listed at VoxEU as “Chairman, Intelligence Capital Limited; Emeritus Professor, Gresham College; Senior Fellow, London Business School,” put up a “nothing to see here” post on the Libor scandal.
It’s clear from this piece that Persaud hasn’t deigned to do basic homework, like reading the FSA’s letter to Barclays, which describes its findings from its investigation and recounts the regulatory violations. For instance, Persaud describes the scandal as having “two phases”, June 2007 to June 2008, and after the collapse of Lehman. The FSA also described two phases of manipulation, but the first was 2005 to 2007, in which Barclays derivatives traders were seeking to move specific Libor indexes, most often the one month or three month Libor, apparently by small amounts (targets of single basis point moves were mentioned), to improve the value of trades they had on. The later phase was in 2007 to 2009, when banks were lowering Libor in an effort to signal that they were healthier than they were. (Within this period, Persaud makes a case that immediately post Lehman, there was no interbank lending market, so it made sense to have the banks keep posting Libor, given all the instruments priced off of it). And his air-brushing out the 2005-7 manipulation also means he does not have to deal with charges made by many and recounted by the Economist, that the gaming of Libor goes back 15 years or more before that 2005 date.
To get to the embarrassing part, we need to review how Libor is set. 16 banks submit a figure that is supposed to represent their cost of borrowing/lending at various maturities in various currencies. For each index value, say one month Euribor, of the 16 value provided, the highest four and lowest four are omitted and an average is taken of the remaining eight.
This is the part of Persaud’s piece that is the stunner:
How much collusion would be necessary to manipulate rates?
This method of estimation – where the extremes are trimmed – makes it impossible for one bank, which has a large net lender or borrower position, to manipulate the Libor rate on its own.
• Any manipulation would need collusion by a minimum of a quarter plus one of the banks being surveyed.
• These banks would have to have similar net positions – so that their interests were aligned in favour of the collusion – and the offsetting position (these banks are lending and borrowing from each other and so someone’s lending is another’s borrowing) would have to be so widely distributed so as not to trigger an offsetting collusion.
In other words, manipulation of these rates for private profit would be difficult to arise or sustain.
This is completely wrong. The ease or difficulty of manipulating Libor is a function of the dispersion of the submissions, and whether the submitter can make an accurate guess as to how high or low a bid it will take for his value to be knocked out on the high or low end. If your submission would have been among the ones in range and you move your submission to a level where you are pretty sure you’ll be excluded, you’ll effect a change in the submissions chosen to be averaged. And the Barclays traders understood that this was an effective strategy. The e-mails included in the FSA release has multiple instances of traders pleading with the submitters to put in a bid that was high or low enough to be excluded from the calculation.
To illustrate, just take a series of submissions that are dispersed (for convenience, percentage signs omitted):
1.0, 1.1, 1.2, 1.3, 1.4, 1.5….2.4, 2.5
Let’s say your submission would have been one of the ones in the middle. We’ll pick 1.8.
So the average per the Libor rules would be:
Knock out four bottom, so 1.3 and below.
Knock out 4 top, so 2.2 and above.
Average the rest (1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0. 2.1): 1.75
Now assume you have nefarious reasons for wanting to move Libor lower. Instead of submitting 1.8, you bid someplace really low but acceptable, say 1.0.
The effect in the 8 numbers to be averaged is to replace 1.8 (the number that would have been in) with 1.3 (the number formerly excluded but now included) This is a difference of 0.5 in the total to be averaged. Divide by 8, and you’ve moved Libor by 6.25 basis points (If you do it the hard way, you’ll see this is correct).
Now if there is less dispersion or less ability of the submitters to make good guesstimates of where they need to place their bid (actually, luck will do, but people who are out to manipulate markets don’t want to rely on being lucky), then it will take collusion to effect a change. But again, that does not mean it takes collusion on the scale asserted by Persaud.
Persaud presents some other peculiar ideas in his article. For instance:
…it was felt that banks are incentivised to submit realistic estimates of the underlying conditions because they are both substantial lenders (who would want higher Libor) and borrowers (who would want lower Libor) and they could not easily predict which they would be more of on a given day.
Note that while Persaud here is giving the theory of Libor, the rest of his piece does not voice disagreement with this part.
First, the banks borrow in the market where that particular bank can borrow. If counterparties want to lend to Bank of America today at 1% at 3 months, that’s where the price is. The fact that it might be expressed in terms of Libor is irrelevant since this borrowing does not reset. When Libor was suppressed, I was getting regular comments that the real rates were 30 to 40 basis points higher, and recent revelations have been consistent with that.
Second, it charmingly speaks of “bank” incentives. Employees, particularly traders, game their organizations all the time. The inmates may well have been running the asylum, yet Persaud refuses to acknowledge that possibility.
Third, Persaud like many others, thinks only of the implications for lending and borrowing, when as we’ve said repeatedly, the action in the Libor scandal was all about the derivatives. And remember, there are certain types of exposures where it is highly likely the dealers were largely positioned on one side of a trade and end customers on another. So dealer incentives may well have been largely consistent. For instance, one customer trade done is scale is borrowing floating rate debt short and swapping into fixed. It’s hard to think of customers who’d be natural swap providers (although the dealers may be accompanied by some Treasury departments and hedgies who are taking a speculative position).
What is the lesson of this flagrantly off-base article? That being an economist means you don’t have to bother with facts? That defending banks is so highly paid that someone like Persuad is willing to write credible-sounding drivel on their behalf? Oh, but I forgot. Mainstream economists gave advice that wrecked the global economy and they haven’t shown any remorse, much the less changed their ways.