Some suggestions on how to improve Libor…
The first is from the Economist, which compares Libor with the problems facing an art gallery or museum — price discovery in many markets is a tricky process.
Essentially, the British Bankers’ Association (BBA), and its troubled Libor, is in a sticky position as benchmarks for rates are required every day. Rain or shine, in conditions of ample liquidity or in clammed up markets, for rarely traded currencies and maturities too.
Given that potentially unrealistic prices may be submitted by panel members — for whatever reason — how might a commitment value be elicited in order to ensure true price information and ward against collusion? The Economist takes a page out of auction theory as they make this suggestion (our emphasis):
[Some Libor] traders had, in effect, formed a “bidding ring”, analogous to a sort of cartel that is familiar to observers of auctions…
A better LIBOR system would be based on actual data as far as possible: not using any market data just because some are missing was never a good idea. If the BBA needs estimates to fill the gaps, it should learn a simple lesson from auctions: you have to stick to your bid. False bidding in an auction is penalised... [more on this below].
The FT has noted that the idea to base the new rates on a combination of actual trades and binding estimates already has appeal among policy circles in Europe. Banks would have to be willing to enter into transactions at their quoted rate, rather than just calling it an estimate.
We would note, however, that there are downsides to basing a rate on transactions due to liquidity issues which open the game up to an entirely different type of manipulation — to simplify more than a little bit, think the Irish music charts or, more pertinently, a situation where a rate is formed in an illiquid market by just one trade. What would that represent exactly?
Thus, on to the still required estimates…
Once some “commitment-value” has been introduced into the process, the system should focus on the weaknesses that auction cartels are known to have.
One suggestion, put forward recently enough by Rosa Abrantes-Metz of NYU’s Stern School of Business, is as simple as allowing the entry of outsiders by allowing other lenders such as money-market funds to submit estimates too. A “more the merrier” approach to the problem.
And/or the BBA could follow a process suggested by Paul Klemperer of Oxford University to the Bank of England, which alongside the US Treasury was trying to push cash into illiquid markets by buying up dodgy collateral and needed prices for assets in an illiquid market. A type of proxy auction:
In Mr Klemperer’s “product mix” auction, bidders submit detailed bids, which include both the prices they would pay and quantities they would accept for a range of goods. Because bids are simultaneous and are never revealed, bidders cannot learn from one another, making collusion harder. Since the auctions are of the many-winner financial type, a knockout system, as in the stamp bidding ring, is unlikely.
Having received a set of bids for different goods, at various prices and quantities, the auctioneer in Mr Klemperer’s set-up then conducts a proxy auction on bidders’ behalf to see who should get what, and what the price should be. Because nothing is revealed to the bidders and they know they cannot influence this process, their best bet is to tell the truth. What is more, since the auctioneer has price information for a range of quantities, it is possible to see how prices change as supply does.
(The Economist argues that the banks are already considering opting out of the Libor-setting process, so the BBA could end up running an auction with no bidders. However, we are sure someone could find a way to persuade them.)
That brings us to a recent idea put forward in the FT. It comes from Frank Partnoy and boils down to this: if you submit the lowest bid, you trade it and the same applies at the highest level.
More specifically (and, again, with our emphasis):
First, regulators should co-ordinate globally to adopt rules requiring each major bank to submit Libor estimates for each currency and maturity. In place of the BBA, a new entity – call it the Libor Trust – would collect these estimates. The Libor Trust would determine the lowest and highest submissions in each category, as is done now.
The teeth of the new regulation would be a rule requiring the bank that submitted the lowest Libor estimate to lend a significant amount of money, say $1bn, to the Libor Trust at its submitted low rate. Conversely, the bank submitting the highest Libor estimate would be required to borrow the same amount from the Libor Trust, in the relevant currency for the specified period of time, at its submitted high rate.
The Libor Trust would collect money from the low bidder, lend it to the high bidder and manage the payments over time. If several banks submitted the same estimates, they would equally share the lending or borrowing. Thus, regulators could create a real and instantaneous consequence for any bank that submitted a false estimate.
This quite practical solution would, in effect, penalise outliers and ultimately corral submitters into a narrower, truthier range — or so it is hoped.
Other benchmark-setting processes, that are already in existence, require the entire group to eat what they quote. For example, the central clearing of credit default swaps hasn’t been around long, but a system to keep members honest is already in place. ICE operates “Random Firm Trade dates” when participants trade at the levels they have submitted. There were 30 such days in 2010 for Eligible iTraxx Index Series. The same also now occurs for Eligible Single Name CDS products.
Might such a system work for Libor? If banks were subject to random trading days when submissions were tested the low-risk incentive to game the system would surely be greatly lessened.
However, there are obvious problems. One, is that since, as mentioned, quite a few currencies and maturities that are not traded very often are included in Libor, test-trades would be difficult/impossible to conduct… let alone at random.
And the same thing applies more broadly in a frozen market — in times of real stress Libor becomes a joke. There is nothing being traded so the estimates are all you have and there is no way to test them, except against each other which doesn’t exactly engender trust. No change there then.
(One possible solution might involve a health warning that depends on the quality of the tested submissions which, if they fell below a certain level, forced contracts to rely on their fallback calculations.)
Of course, we could just be made to go the way of ratings in a Dodd-Frank world, and be shown the door. That would leave an opening for it to be replaced by… something. Objections and suggestions are welcome in the space below.
Related links:
How auction theory can help improve the system for setting LIBOR – Economist
Make banks pay if they cheat on Libor – FT
After Libor – the search for a new benchmark – FT