The manipulation was not especially limited. Photographer: Chris Ratcliffe/Bloomberg
The manipulation was not especially limited. Photographer: Chris Ratcliffe/Bloomberg

A lot of banks manipulated the London interbank offered rate, and those banks have been settling with regulators, and the settlements are all pretty similar. So kudos to Lloyds Banking Group for actually being original in its Libor settlements -- with the U.K. Financial Conduct Authority, the U.S. Commodity Futures Trading Commission, the Bank of England and the U.S. Department of Justice -- announced today.

I mean, the broad outlines are familiar. It would really be news if Lloyds manipulated Libor with proper spelling and grammar, but, nope. When Rabobank Libor submitters wanted an "obseenly high" Libor, Lloyds's submitters were only too "happy to ablige."1 More quotes here; they are dumb and embarrassing and, at this point, unsurprising.

But there are two innovations here. One is that Lloyds -- and Bank of Scotland,2 which it acquired in 2009 -- went to the trouble of actually manipulating Libor. I mean, most Libor manipulation is hardly worthy of the name "market manipulation." Normally the way you manipulate a market is, you buy a lot of stuff, to push up the price of that stuff, because then you'll be able to sell even more of that stuff at the new higher price. If you have a derivative contract to sell 1,000 widgets at the closing price on Thursday, you should go buy a hundred widgets in the minute before the close on Thursday, and as inefficiently as possible: wasting an extra dollar on the last widget makes you an extra $1,000 on your derivative.

You don't need to bother with that with Libor. Libor is an interest rate based on a poll: Someone calls up a bunch of banks, asks them what they think Libor should be, and takes a trimmed average of their answers. To manipulate that, all you need to do is say the number that you want on the phone. If you're getting one of those phone calls, the manipulation is free. (Until the fines I mean.)

Lloyds did the whole lying-on-the-phone thing, of course, but also did regular market-manipulating stuff to make money on its forward rate agreements:3

Lloyds Traders on the Money Market Desks engaged in at least three schemes from September 2006 to December 2006 to “force LIBOR”. The Lloyds Traders entered into FRAs and then bid aggressively in the cash market. Whilst this was at times when the bank needed to take in cash, the primary motivation was in order to influence upwards the one month GBP LIBOR submissions of other LIBOR Panel Banks and therefore increase one month GBP LIBOR rates, which would benefit those FRAs.

Notice the dates. In 2006, before the financial crisis, Libor was perceived not as an imaginary number, but as an actual interest rate set by market conditions: It was the interest rate for unsecured interbank borrowing. So to manipulate it, of course Lloyds actually had to do some market activity instead of just lying on the phone. Later everyone figured out the easier method, but in those innocent days of 2006, Lloyds adorably thought it had to spend money to make money.4

The other innovation is in what Lloyds manipulated. Lloyds and Bank of Scotland manipulated various Libors, of course, both for the profit-seeking purpose of optimizing their derivatives positions, and for the embarrassment-avoiding purpose of not admitting to the market that they couldn't borrow money cheaply. But also: They couldn't borrow money cheaply. Fortunately for them the Bank of England came through with a solution in the form of its Special Liquidity Scheme, which ran from April 2008 through June 2011. And then, being manipulators, Lloyds and Bank of Scotland manipulated that.

Here's the idea. If you're a big Libor-panel bank, you can borrow money at Libor, without posting collateral, just based on your own unsecured credit. Or that is the theory of Libor, anyway. But by 2008 everyone knew that banks couldn't borrow at Libor, so that was just imaginary. Instead banks could borrow money from the repo markets, posting assets as collateral. But that also wasn't true, because the banks' assets were a lot of garbage, and nobody wanted to fund garbage. So the Bank of England introduced the SLS, under which "banks and building societies could, for a fee, swap mortgage-backed and other securities that had temporarily become illiquid" -- that is, garbage -- "for UK Treasury Bills, for a period of up to three years." Then they could take the Treasury bills and post them as collateral to borrow in the repo markets, so they could actually get money.

In exchange for this service, banks had to pay the Bank of England a fee equal to the spread between the three-month repo rate and the three-month Libor rate, floored at 20 basis points. That math makes the scheminess of the scheme obvious:

  • Libor = Repo Rate + Spread
  • If you use the SLS, you're borrowing at Repo Rate and paying max(Spread, 20 basis points) to the Bank of England.
  • So you're borrowing at Libor or a little higher, and giving up your mortgage-backed securities as collateral.
  • Obviously no bank would prefer that to just borrowing unsecured at Libor, if it could actually borrow unsecured at Libor.
  • But, y'know. "Libor -- the rate of interest at which big banks don't lend to each other."

Some banks, especially Lloyds and Bank of Scotland, relied heavily on the scheme, meaning that it was much cheaper funding for them than they could get elsewhere.5 And of course they wanted to pay as little as possible, which they did by manipulating it.

Now to minimize the Libor/repo spread that the banks paid, they could:

  • manipulate Libor down, and/or
  • manipulate the repo rate up.

Lloyds and Bank of Scotland were on the repo rate setting panel, and for about a year and a half in 2008-2009, they manipulated the repo rate up on days (once every three months) when the SLS borrowings reset. This, according to the Bank of England and the FCA, cost the Bank of England 7.76 million pounds (total on all banks' borrowings -- which, remember, were all indexed to the rate that Lloyds was manipulating). That is not very much money on a program that swapped some 185 billion pounds of bills; it works out to about 0.3 basis points per year.6

But they were caught, and today Lloyds paid back that 7.76 million pounds to the Bank of England. "The Bank believes that this payment fully compensates it for all losses which it may have suffered."

Now let's go through this again slowly.

  • The banks paid the Bank of England the SLS Spread.
  • SLS Spread = Libor minus Repo Rate.
  • The banks manipulated the Repo Rate up, reducing the SLS Spread and costing the BoE money.
  • Costing it about 0.3 basis points a year, in fact.
  • Lloyds paid back the BoE for that manipulation.
  • "The Bank believes that this payment fully compensates it for all losses which it may have suffered."

How else could Lloyds have manipulated the SLS Spread, which remember equals Libor minus the repo rate? Well, for instance, if Lloyds manipulated Libor down, then that would also have the effect of reducing the SLS spread. Is it ... is it just possible that Lloyds was manipulating Libor down in 2008 and 2009?

Oh right yes obviously that is exactly what happened and it's extensively documented in the FCA order.7 Also it's the thing that everyone knows about Libor manipulation: Most (all?) banks, starting with the innovators at Barclays, were lowering their Libor submissions at the same time and for the same reasons: to avoid spooking the market by admitting that they couldn't borrow cheaply in the interbank market. It's hard to know exactly how big the effect was, but one biased (plaintiffs'-expert) view is that "Libor rates were probably 25 basis points to 35 basis points lower than they should have been from late 2007 through early 2009." Or you know around one hundred times as much as the effect of Lloyds manipulating the repo rate.

But the Bank of England only got them for the repo rate manipulation, not for the Libor manipulation. Weird, right?

There's an obvious explanation. Yes, Libor manipulation cost the Bank of England money by reducing the spread it made on its Special Liquidity Scheme. But the Bank of England knew that going in. It knew that banks couldn't borrow at Libor, because the whole point of the SLS was to charge banks at least Libor for secured borrowing, which would be ridiculous if Libor was actually a viable unsecured borrowing rate. The Bank of England was happy to accept a fee that was lower than it should have been, because Libor was lower than it should have been. That was just part of the deal, and it would be silly for the BoE now to go back and demand more money just because Libor was manipulated down.

All the BoE asked in exchange for its below-market financing was that the banks show a little gratitude and not manipulate the other interest rate that affected its fees. And Lloyds couldn't even manage that.

1 Actually that Rabo Yen Libor submitter wanted that high Libor for his "little ... [racial epithet redacted] friend in tokyo" and oh boy don't you want to know what the epithet was?

Also I'm fond of this one:

The Authority has identified communications involving Lloyds Trader A and certain Brokers in 2007 in which they discuss the manipulation of Lloyds Bank’s GBP LIBOR submissions. For example, on 17 August 2007, Lloyds Trader A called Broker A saying, “I ain’t got any 3s fixings mate. I’ve got no fixings today. So I can do my LIBORs wherever I f****** want to put them, mate.”

The point here seems to be that Trader A would normally submit a Libor that would help his derivatives positions. But that particular day he had no derivatives positions that would be helped or hurt by his Libor submission. So he called up an interdealer broker to say, basically, that he was open to being bribed to submit a Libor that would help someone else's derivatives positions. Efficient!

2 It's Bank of Scotland to the FCA, but HBOS plc to the CFTC. (Technically HBOS seems to be the holding company of the Bank of Scotland; HBOS in turn is owned by Lloyds.)

3 Bank of Scotland had a similar idea:

Bank of Scotland had a portfolio of assets (typically loans made to third parties) that were referenced to GBP LIBOR and re-fixed on a rolling basis for the term of the asset, which process was described as a “roll”. Bank of Scotland engaged in a scheme whereby it sought to take advantage of the fact that GBP LIBOR submitters at other LIBOR Panel Banks took into account perceived conditions in the three month GBP cash market. This scheme involved increasing Bank of Scotland’s bids for three month GBP cash in order to influence the perception of other GBP LIBOR Panel Banks, with the aim of causing them to increase their three month GBP LIBOR submissions. BoS Trader B engaged in at least one such scheme in November and December 2006.

4 Perhaps I exaggerate slightly:

The Traders were prepared to stand by their bids, but to avoid this they engaged in these schemes during periods of market illiquidity reasoning that “you’ve got to do it when people can’t lend” and therefore the likelihood of their bids being fulfilled was lessened.

You don't want to spend much money to make money.

5 The Bank of England loaned out a total of 185 billion pounds of bills in the program, with Lloyds and Bank of Scotland apparently being about half of the scheme:

Institutions were able to swap assets under the SLS until January 2009. Through the SLS, the Bank of England lent Treasury Bills with a face value totalling £185 billion. ...

The total SLS fees paid to the Bank of England by all participating firms was £2.6 billion. The Firms were amongst the largest users of SLS. During the period of the Firms’ participation in the SLS, between April 2008 and June 2011, Lloyds Bank paid £394 million and Bank of Scotland paid £884 million in SLS fees to the Bank of England.

Lloyds and Bank of Scotland combined for about half the fees, so figure they borrowed half the bills, or around 90 billion pounds or so. On the other hand a 20 basis point spread for 3 years on 90 billion pounds only comes to 540 million pounds, or less than half of their actual fees, though sometimes the spread was over 20 basis points.

6 That is, 7.76 million divided by 185 billion equals 0.000042 (0.42 basis points). Spread that over the 16 months or so of their manipulation and you get about 0.026 basis points a month, or about 0.31 per year.

7 Actually it's just documented for Bank of Scotland, not Lloyds. Bank of Scotland was in particularly grim shape for borrowing in the interbank market, which is part of why it's now owned by Lloyds:

During 2008, there was concern at Bank of Scotland about making LIBOR submissions away from the pack because Bank of Scotland management wanted to avoid negative media comments and market perception about the bank’s creditworthiness. On 6 May 2008, BoS Senior Manager A emailed BoS Senior Managers B and C about USD LIBOR submissions: “It will be readily apparent that in the current environment no bank can be seen to be an outlier. The submissions of all banks are published and we could not afford to be significantly away from the pack.” ...

On 26 September 2008, BoS Manager B instructed BoS Trader C to lower Bank of Scotland’s USD LIBOR submissions into line with other LIBOR Panel Banks. In a Bloomberg message to a trader at another bank, BoS Trader C said “ive been pressured by senior management to bring my rates down into line with everyone else.” Accordingly, on 26 September 2008 Bank of Scotland’s three month USD LIBOR submission fell to the same level as the highest other submission, albeit it remained 44 basis points above the published LIBOR rate.

Over the course of October 2008 Bank of Scotland’s three month USD LIBOR submissions fell into line with the level of the LIBOR fix and remained within the pack of submissions until the merger with Lloyds Bank on 19 January 2000.

To contact the writer of this article: Matt Levine at mlevine51@bloomberg.net.

To contact the editor responsible for this article: Zara Kessler at zkessler@bloomberg.net.