Photo: Peter Thal Larsen
How is LIBOR calculated?The BBA provides pretty detailed analysis of the process. The key here is what the rate is meant to be. The contributors, are supposed to submit a rate for each currency they contribute for overnight, one week, two week, and monthly out to a year. The rate is meant to be:
“At what rate could you borrow funds, were you to do so by asking for and then accepting inter-bank offers in a reasonable market size just prior to 11 am?”
This is a bit like self-reporting your weight. The bank is supposed to submit a rate where they think they could borrow, not where they actually borrowed or where they would lend to other contributors. Right from the start the question raises questions that have been discussed for years.
How can a bank “know” where some other bank will lend them money? Can’t they use transactions? Can’t they get firm “offers” from other banks? Why not have the banks submit levels where they would lend to other banks?
The very nature of the question used to solicit rates tells you all you need to know. LIBOR has always had an element of “gamesmanship” if not outright lying.
In general, banks will tend to submit lower rates and attempt to artificially lower LIBOR. There are two reasons for this:
- Signaling Effect – banks don’t want to say it would cost them more money to borrow than their peers because that would be admitting to weakness and may cause their lenders to pull back and create a financing freeze. Each contributor’s LIBOR indications are published so if a bank shows up with a particularly high estimate of what it would cost to borrow, it would attract unwanted attention. It may be the truth and should be evident in the CDS and bond markets, but for some reason banks remain concerned about the signaling impact and tended to skew LIBOR submissions lower than they should have been
- Risk and P&L Impact – The big banks always have a lot of risk associated with LIBOR. It will affect their borrowing costs, it will affect what they receive on floating rate loans and it will affect the value of their interest rate derivative books. It might have other secondary impacts, but those 3 areas are big. It is probably safe to say that banks in general benefit from lower LIBOR, but that won’t be true for all banks and won’t be true for all days. There are occasions where banks may benefit from a higher LIBOR. If they have a disproportionately large amount of floating rate loans resetting on that day, they may benefit by having LIBOR higher that day, thus locking in slightly higher income for that period. Someone at the bank will know their exposure to the LIBOR setting on any day, and it would be hard to believe that on days when the exposure is large either direction, the group that submits it would be unaware of the potential P&L impact.
That is a dangerous concoction. A question that leaves a lot of wiggle room, and banks that may have strong incentives to use that “wiggle” room.
Controls and Actual Calculations
For me, the single most important rate is the 3 month USD LIBOR rate. It certainly impacts Americans more than any other rate calculated by the BBA. Here is yesterday’s submissions, and the calculation.
Editor’s Note: These numbers correspond with July 2.
3 Month USD LIBOR Fixing
There are 18 banks that submit U.S. LIBOR. The 4 lowest rates and 4 highest rates are thrown out for purposes of the setting. Then LIBOR is set as the average of the remaining 10 rates.
You can see the wide discrepancy in rates. HSBC and Barclay’s clearly think they have easy access to money. SocGen and BNP seem to think it would cost them a lot of money relative to the others. There is no indication how much borrowing and lending is occurring in the interbank market, so there is no easy way for an outsider to tell if this reflects reality or not. JP seems conservative given that their 5 year CDS trades at 125 which is similar to HSBC’s 120 level. Barclay’s 5 year trading at 205 would indicate a possibly optimistic view of where they could get short term funding, and Citi and BAC barely behind JPM again seems a bit optimistic given their CDS trade at 235 and 250 respectively.
So by throwing out the outliers, and using a relatively large pool to calculate the average, the BBA attempts to mitigate the risk of any one bank skewing the setting. The problem is that it doesn’t do much if multiple banks collude to manipulate the setting.
If multiple banks have the same incentive to skew their own submission, and worse yet, communicate that to “friendly” banks, then the BBA methodology breaks down further.
The problems with the BBA methodology is there is no confirmation that the rate submitted is reasonable, and nothing is done to protect against group rather than individual bias in their submissions.
I don’t think this will turn into lawsuit mania. In spite of the huge notional amount of contracts and loans outstanding based on LIBOR, it may be difficult to pursue a case. We will go through the cases that might make the most sense in a moment, but here are the main reasons that I don’t think this will snowball into a massive amount of litigation
- Individuals and Many Corporations benefitted from lower LIBOR settings. To the extent the bias was to artificially lower LIBOR, the direct impact would be to reduce amounts owed on floating rate borrowings. Anything where individuals as a class were hurt would expose the banks to big problems, as they would be getting sued by a group that would have a lot of jury sympathy. But in this case, individuals that had any LIBOR exposure, typically directly benefitted from any bias to make the rates low, as they borrowed in LIBOR and little of their investment income was based on LIBOR.
- The duration of LIBOR is short. Even if you have a 10 year swap where you paid fixed and received floating, you would likely have to demonstrate that on each reset date, the bank colluded to move LIBOR setting against you. Maybe you can argue that the overall trend was enough to impact your mark to market, but that may be a stretch. Having to show that at each quarterly reset, on the day your contract resets, there was collusion, could be very difficult. The duration also comes into play on the damages side. Let’s assume you had a $10 billion swap (a reasonably big trade). If the banks all colluded against you on a particular setting and managed to move LIBOR by 10 bps (a big differential) your “loss” would be $2.5 million or 0.025% of notional. Not small, but another example of how the short term nature of LIBOR makes it hard to build a big claim.
- The complexity of the process helps the banks as a group. How would you prove you were hurt by a particular bank? If a bank submitted a “bad” price, but was already in the outlier group, it would be hard to make a claim since it didn’t affect the calculation. If a banks “bad” price moved it from the calculation group to the outlier, it is only the difference between what would have been fair for them, and what the bank that is now being used submitted. It is really hard to show how much 1 bank affected the outcome. Larger groups caught in the act would be required, but this will be more difficult to find consistently, and remember to prove real loss, you would need that group collusion on each reset date. Then those banks would split the cost. That is ignoring the difficulty of proving what a “bad” price is. The question certainly allows for the defence tactic of “well, that’s what I thought it would be”, especially during periods where interbank activity went to zero and the banks were relying heavily on central bank funding.
It is easy to salivate over the potential lawsuits and the losses the banks might have, but a dearth of sympathetic victims, the rolling nature and short duration of LIBOR based products, the lack of a test to determine if a submission was “bad” and the complexity of the setting process make it far harder to bring successful lawsuits than the headlines might suggest.
I would expect more firings at banks. Clearly Barclay’s was not acting alone. In this environment, no bank is going to support an employee who was involved in this. The banks will defend themselves in court against lawsuits based on the letter of the law, but at this stage, none are going to fight to save staff that participated in schemes to move LIBOR (or at least those dumb enough to use e-mail and other recorded forms of communication).
Will other big heads roll? That to me is less clear, but is a possibility. I’m assuming like in most other things at big banks, if the people in charge are well-respected with no big internal rival, they survive, but if the person has been on the edge and has a group happy to force a regime change, we could see one.
Ultimately the LIBOR setting process will have to change. The current process is too vague. There have been some calls for an alternative to LIBOR, but with so many contracts outstanding, I think that is unlikely. It will be far easier to just amend the methodology and try to improve the existing LIBOR process rather than starting an entirely new rate series.
There will likely be some cases brought that get settled and cost the banks some money. If anything, I suspect municipalities might form the best class action. I think many did enter into pay fixed, receive floating swaps, so as a group they might have the size and sympathy to pursue something, though I bet the lawyers for the banks will gain the most, with lawyers for the plaintiffs coming in a close second, and the plaintiffs and banks wondering how they got sucked into spending so much on legal fees. I could be wrong on the lawsuit side, but even the evil side of me, has trouble figuring out how to make a strong case with big potential payouts.