Why would you want to be a Libor bank? It's unpaid work, in that every day you need to get on the phone to Reuters and say where you can borrow across 150 tenors and currencies in which you mostly don't borrow, which one assumes taxes the imagination. And there's potential badness, though the specifics vary over time: in 2008 the badness was "if we submit a high Libor people will think we can't fund ourselves and there will be a run"; in 2012 the badness is "if we submit a bad Libor we'll go to jail."
So why do it? Some part of the answer is a certain kind of public-spiritedness: you're JPMorgan, you trade a zillion dollars of interest-rate swaps, the swaps reference Libor, there needs to be a Libor, you and your friends who trade swaps get together and agree to make a Libor so that you can have a swaps market, and then you keep making that Libor every day. We'll go be cynical in a minute, but let's take sec to appreciate that this is a real thing and surely the primary motivator behind Libor. The financial industry is actually good, in a way that say the cell-phone mapping industry is not, at getting together and solving coordination problems via committee in a way that creates new markets for everyone to benefit from. Everyone on the committee, I mean.
The other part of the answer is of course "once you've set up this swaps market, you can manipulate it to your profit." This to some extent counteracts the preceding public-spiritedness - your zillion-dollar swaps market depends on your benchmark being reliable, so if you manipulate it you screw things up for everyone - but in meaningful ways it doesn't. For one thing every market consists of traders cheating each other within controlled bounds; it's just about plausible to think that Libor submitters thought of themselves like the bond trader who tells his customer that he just can't afford to part with these bonds for less than 102 when in fact he picked them up at 85, and that they figured that their cheating Libor down would be balanced by some other similarly dodgy character at another bank cheating Libor up, creating a system that is deceptive in every particular but trustworthy in aggregate.1 For another thing, you have a world and can go look at it, and empirical investigation will tell you that swaps volume was untroubled by actual Libor cheating through 2010, and doesn't seem especially troubled (stripping out effects of economic contraction, concerns on counterparty credit risk, increased clearing and collateral requirements, netting down, etc.) by the exposure of Libor cheating in 2012. So go ahead and cheat.
So there you are chugging along being a Libor panel bank, submitting Libors that are honest enough to foster a robust swaps market, but dishonest enough to foster robust profits for you in that market, and bang, this:
2.53 Accordingly, while it is recognised that there are arguments both for and against amending the current criminal regime, the Review recommends that Section 397 of FSMA is amended to cover manipulation of LIBOR, so that the FSA is able to investigate and prosecute such conduct.
That's from the Wheatley Review; Martin Wheatley runs Britain's FSA and, since Libor is the London IBOR, his view on who goes to jail for doing what with it carries a lot of weight. A bunch of careless IMers will be happy to note that Wheatley is not especially convinced that Libor manipulation is a crime now, under UK law, so the manipulators should be in good shape as long as they stay out of America.
But let's talk motivation. Once you submitted Libor even though nebulous badness lurked outside of it, because (1) that badness was quite nebulous (after all, submitting fake Libors apparently wasn't criminal, so the badness was of the form of public embarrassment) and (2) you could make money on it. Now the badness is (will be) crystallized (and jail-related) and the path to profitable fakery is narrower. (At a minimum, you can't do it over IM.) So why keep submitting Libors?
That's the meat of the Wheatley report. As they put it:
5.24 Given the global importance of LIBOR, the Wheatley Review also believes that the international community stands to gain from the continuing participation of major banks in the LIBOR panels. Therefore, the Wheatley Review asks that international authorities engage with relevant institutions to encourage continuing participation in the LIBOR panels.
"Encourage" it how? Wheatley has ideas and they are good; the function of the report is to (1) threaten jail for submitting bad Libors and then (2) make it easier to avoid submitting bad Libors. Like, first of all: cut out all the Libors that are entirely fake. Here are some Libors colored by degree of fakery:
So that's uselessness of input; here is meaninglessness of output:
So Wheatley recommends abolishing AUD, CAD, DKK, NZD and SEK Libors, all kinds, and all Libors everywhere other than 12, 9, 6, and <=3-month tenors (and suggests rethinking everything other than 1, 3 and 12-month), since nobody uses them ever so they're just an opportunity to get caught committing pointless fraud.1a As he puts it, "Should this recommendation be implemented in full, the number of LIBOR benchmarks published daily could be reduced from 150 to 20. These remaining rates could more easily be supported by trade data, and moreover, are heavily used by market participants."
This addresses the sheer workload of Libor submitters - they have to submit many fewer rates, and the rates they submit require less imaginative effort. It also, and more importantly, addresses the riskiness of Libor submitting: if you can point to an actual transaction (interbank or, somewhat less Liborishly, wholesale) to support your Libor submission, you're less likely to have it second-guessed and end up going to jail. And Wheatley is sensitive more generally to that risk; suggestions like an industry-wide code of conduct, recordkeeping of Libor-supporting transactions, and auditor review of Libor submission processes are at least as much about protecting banks from second-guessing as they are about protecting counterparties from manipulation.2
So that's all nice! But still, the Libor guidelines are a bit daunting. The recordkeeping and codes of conduct and auditing, sure, but also the irreducible fictitiousness of some of it. The Libor submission guidelines3 ask submitters to "use their experience of the inter-bank deposit market and its relationships with other markets," sets out a "hierarchy of transaction types," and allows adjustment based on "interpolation or extrapolation from available data" and "non-representative transactions." If you're actually getting unsecured inter-bank deposits in the relevant currency and tenor, you're great. (You were before, too.) If not, judgment still applies; actually even if you're getting unsecured inter-bank deposits you have to wonder if they're "non-representative." There are parameters and procedures and safeguards, but ultimately you're still in a world of "ehhh, figure out the right Libor and write it down." Now with official criminal penalties.
I like it! The Wheatley Review doesn't "fix" Libor in the sense of making it transaction-based and immune to manipulation. Not because that's impossible, but because it's hard, and there's no great reason that Wheatley should be the one doing it. Libor is a public good that is useful for making swaps out of; a regulator-imposed thing might or might not be useful for making swaps out of. It's hard for Wheatley to know, and he doesn't have that much incentive to find out.
You know who does? JPMorgan. And every other bank that trades swaps and makes floating-rate loans. The Wheatley Review doesn't blow up the $300 trillion of contracts referencing Libor; it just gently nudges banks away from them.4 Now they know that judgment-based Libors will be subject to a lot of scrutiny and criminal penalties, so they have every incentive to come up with a better system that avoids jail risk for them - but that also is efficient and trustworthy enough for the market to adopt.5 Wheatley's Chapter 6 gives an overview of potential transaction-based benchmarks, but why should he make the decision between them? That's, after all, what the market - or, at least, a committee of banks - is good at.
1.Of course this doesn't work in 2008ish when everyone's incentives go the same way: manipulate Libor down so as not to look desperate. In existential crises, your market-integrity concerns have less purchase.
1a.Update! (1) Thoughts on legitimacy of "1a" footnote reference? (2) Simon Hinrichsen writes: "In Denmark, we have the CIBOR (Copenhagen IBOR), which underpins DKK 420bn worth of mortgages and other household loans. In fact, it has been under scrutiny over the summer. Surprisingly, it hasn't been a big story other places than in Denmark. It has, like Libor, been said to be too high, enriching the banks and costing consumers. The Danish FSA has just said that they will create 'Cita' based on actual transactions and implement something that looks like the Wheatley review."
2.Other benefits include a recommendation (paragraph 5.16) that individual Libor submissions not be published for at least 3 months, reducing the important risk that a true but too-high Libor submission will cause an immediate run on a shaky bank.
3.Here they are, from Box 4.B:
LIBOR submissions should be determined based upon the following hierarchy of transaction types. Submitters should use their experience of the inter-bank deposit market and its relationships with other markets to develop their LIBOR submission. Greatest emphasis should be placed on transactions undertaken by the contributing bank.
1 Contributing banks’ transactions in:
- the unsecured inter-bank deposit market;
- other unsecured deposit markets, including but not limited to, certificates of deposit and commercial paper; and
- other related markets, including but not limited to, overnight index swaps, repurchase agreements, foreign exchange forwards, interest rate futures and options and central bank operations.
2 Contributing banks’ observations of third party transactions in the same markets.
3 Quotes by third parties offered to contributing banks in the same markets.
4 In the absence of transaction data relating to a specific LIBOR benchmark, expert judgement should be used to determine a submission.
Submissions may also include adjustments in consideration of other variables, to ensure the submission is representative of and consistent with the market for inter-bank deposits. In particular, the information obtained above may be adjusted by application of the following considerations:
- Proximity of transactions to time of submission and the impact of market events between transactions and submission time;
- Techniques for interpolation or extrapolation from available data;
- Changes relative credit standing of the contributor banks and other market participants; and
- Non-representative transactions.
4.And gets a start on documentation fixes:
5.34 Therefore, the Wheatley Review recommends that industry bodies that publish standardised legal documentation in relation to contracts referencing LIBOR, as well as LIBOR users, should develop robust contingency procedures to take effect in the event of longer-term disruption to the publication of LIBOR. Specifically, new contingency provisions should be designed to function without reliance on submissions from LIBOR panel banks. Chapter 6 explores alternative benchmark rates that could be used in such a contingency.
5.Like, one churlish fantasy approach is to say "fine, we'll stop quoting Libor or doing swaps based on Libor, we'll quote a new rate called AntiLibor, with the same fraudy procedures as old Libor, and build a swaps market around that." Right? But that's prevented in part by Wheatley's suggestion in paragraph 2.54 that "The scope of the offence could be limited to false or misleading statements in relation to the setting of a specified benchmark, or relevant financial instruments or relevant agreements as currently defined in FSMA," and in part by the fact that no customer would really want to move from the trustworthy-ish regulated Libor to the totally untrustworthy unregulated Libor. So you need to offer them something with a compelling case for trustworthiness.