Rabobank Had to Make Money Somehow (by Manipulating Libor)
Rabobank Groep announced today that it is settling a bunch of investigations with a bunch of regulators over its manipulation of the London interbank offered rate, which is a thing that happens sometimes to European banks. The etiquette now is that I go through the documents and pick out some funny quotes, and while Rabo's Libor manipulators were maybe more buttoned-down than some of their competitors, they still got the job done:
As another example, on the morning of Monday, August 13, 2007, another U.S. Dollar Trader ("Trader-2") messaged Submitter-1: "High 3s and 6s pis today mate (esp 6mths!!) if u would be so kind.. Gotta make money somehow!" Submitter-1 replied: "cool.."
Later that same day, Trader-2 messaged Submitter-1: "where do u see 6m LIBOR tomorrow pls?" Submitter-1 wrote back: "where do you like to see it, is more the question?" Trader-2 responded: "HAAA , trying to price a deal, could make/lose a tidy amount of cash!!"
On November 29, 2006, Trader-2 again wrote to Submitter-3: "Hi mate, low 1s high 3s LIBOR pls!!! Dont tell [Trader-1] haa haaaaaaa. Sold the market today doooooohhhh!"
Trader-5 responded: "Wud be nice if you could put 0.90% for 1mth cheers." Submitter-4 wrote back: "Sure no prob. I'll probably get a few phone calls but no worries mate!" Trader-5 replied: "If you may get a few phone calls then put 0.88% then." Submitter-4 responded: "Don't worry mate - there's bigger crooks in the market than us guys!"
Trader-2 again wrote to Submitter-3: "Appreciate 3s go down, but a high 3s today would be nice... cheers chief." Submitter-3 wrote back: "I am fast turning into your LIBOR bitch!!!!" Trader-2 replied: "Just friendly encouragement that's all , appreciate the help." Submitter-3 wrote back: "No worries mate , glad to help ....We just stuffed ourselves with good ol pie , mash n licker !!"
Or, I mean, you get the idea, and you probably did a while back, but I find it oddly hard to stop. These Libor settlements are among other things a wonderful corpus of idiomatic international finance English; I hope some ambitious English teachers abroad will mine them for lessons the way that Brazilian school does with Twitter.
I suppose there is other data to be mined from these settlements though. For one thing, as we get more fine-grained data about what Libor manipulation looked like, it seems ... well, both more pervasive and less bad than you might have expected. (Also less grammatical!) My sense has always been that Libor was basically set up to be robust to the sort of manipulation that is documented here. Yes, Rabobank had billions of dollars of swaps that would increase in value if it submitted Libors that were too high. But UBS, say, had billions of dollars of swaps that would increase in value if it submitted Libors that were too low. Since everyone was manipulating Libor to benefit their derivatives positions, and since those derivatives positions were by definition zero-sum, the manipulation would more or less cancel itself out.*
That's hard to prove empirically because ... well, the Libor settlements so far are not laid out in a convenient format for empirical study.** You can find the odd two-sided market, though; here is Rabo:
After January 2009, when Rabobank's Yen LIBOR submitters moved to Utrecht, Trader-5 would frequently send an entire slate of LIBOR rates to the submitters for them to input. For instance, on February 25, 2009, Trader-5 wrote to Submitter-6 with the subject line "libors" and wrote: "Could you set libors for today as below please?": "1m 0.39;" "2m 0.60;" "3m 0.65;" "4m 0.71;" "5m 0.76;" "6m 0.80."
Meanwhile, in another part of town, here is noted alleged Libor manipulator Tom Hayes at UBS:
So while Hayes was trying to keep one- and three-month yen Libor down on the 25th, Rabo was keeping it up, with submissions above the previous day's fixing (and above the ultimate fixing that day too).*** The system worked!?
To be fair, there are also instances in which multiple banks wanted to move the same Libor in the same direction on the same day, and in which multiple banks (and brokers) colluded together to move Libor. Which makes sense: Part of what is driving the Libor investigations is just that there was a conspiracy, with some traders and interdealer brokers trying to persuade multiple banks to manipulate their Libors the same way. So the cases that have been brought so far have a tendency to reinforce rather than cancel each others' manipulations.
You can see why these interbank conspiracies would be the easiest to catch and document: If you're just manipulating for your own account there's less reason to, like, e-mail and instant message about it.**** Though not none! One weird aspect of the Rabobank case is that several of its submitters had their own derivatives positions: They weren't just manipulating Libor to help their buddies or their banks; they were manipulating it to help their own books. And, you know, e-mailing about it (emphasis added):
In addition to accommodating swaps trader requests, certain Rabobank Yen LIBOR submitters also took their own trading positions into account. At times, submitters also had to balance competing requests when multiple Rabobank swaps traders had trading positions that conflicted with each other. For example, on Monday, March 26, 2007, Trader-4 emailed Submitter-4: "On libors, this week have a fair bit of 6mths rolling off, I am short so if you can discreetly drop your 6m by 1-2 bp without any trouble would be great - if not no probs mate" and "Im happy for you to keep your 1 mth high as both [another Yen swaps trader ("Trader-5")] and I are long a fair bit of those till Wed." Submitter-4 wrote back: "sure no prob - all my stuff is mainly 1 mth so will keep that high and drop 6's cheers."
The Department of Justice statement of facts goes into a number of instances where Rabo submitters balanced supply and demand within the bank, giving priority to the traders whose positions were bigger and/or who were nicer and better friends with the submitters. There are important lessons about market microstructure to be gleaned from those instances.
Libor remains a strange scandal psychologically. On the one hand, everyone knew, and said in recorded written conversations, that what they were doing was wrong -- saying "there's bigger crooks in the market than us guys" is a pretty clear admission that you're a crook. On the other hand, everyone acted as though what they were doing was normal and above-board. So in the Rabo case, senior management seemed to think that Libor manipulation was part of best practices:
Rabobank designated certain Traders as "Centres of Competence" Traders on the basis that those Traders had extensive experience, market knowledge and expertise in cash instruments for a particular currency, because, in part, they had overall responsibility for managing the net risk arising from all Rabobank's transactions in a particular currency.
From 2009 to November 2010, Rabobank Managers specifically instructed JPY LIBOR Submitters to seek market colour from these Centres of Competence Traders. That instruction was given to provide Submitters with the benefit of the Traders' expertise in JPY cash trading. However, this instruction also created a risk that the Traders who had significant derivatives trading positions would improperly seek to influence those JPY LIBOR Submitters. This risk was aggravated by the fact that certain JPY LIBOR Submitters took the instruction to seek market colour to be a direction to obtain and make LIBOR submissions at levels suggested by Traders, including Traders who, in practice, exploited their roles by making suggestions that directly benefitted trading positions.
And one of those Centres of Competence did this, which is weird:
In fact, although Trader-5's requests to Rabobank's Yen LIBOR submitters were frequently motivated by his own trading positions, Trader-5's communications after January 2009 were sometimes the result, at least in part, of his belief that he was Rabobank's Yen LIBOR submitter and that Rabobanlc's actual Yen LIBOR submitters, due to their lack of relevant knowledge, needed his input. Thus on numerous occasions Trader-5 communicated rates to Rabobank's Yen LIBOR submitters even though neither he nor others on his desk had a trading position related to that rate. As such, on numerous occasions, Trader-5's communications were, at least in part, not motivated by a desire to benefit swaps traders' trading positions.
He just wanted them to get Libor right! Where "right" meant, um, whatever you think it means (accurately reflecting market borrowing rates or whatever) when he had no position, and helpful to his position when he had a position. You get the sense that he really believed that that's how Libor was supposed to be set.
* I mean, to some approximation. The net interest rate swap positions in the world need to sum to zero, but that's not true of the net interest rate swap positions of Libor submitter banks. And of course some manipulators just tried harder.
** Some of that is just a formatting problem that could be fixed by a diligent reporter, blogger or intern sitting down and building a searchable and sortable database of all of the documented cases of Libor manipulation in all the settlements, categorized by date, bank, currency, tenor, manipulation direction (up or down) and outrageous instant message/e-mail quote, if available. My crack team of interns is on the case! (That sentence is false! I did start building a spreadsheet, but my attention span and analyst resources are not what they used to be.) Some of it is not; the settlements are full of vague statements like "The following are just some examples of the numerous" etc. etc., or "These requests included: ... At least 278 JPY LIBOR requests to Broker 1," which means that any privately constructed database is unlikely to be comprehensive. Still, seems worth doing, for some value of "worth doing."
*** I see 3-month yen Libor at 0.6375 percent on 24 February 2009 and 0.63625 on the 25th. One-month is 0.3750 and 0.37625. So Rabo's submissions were above the previous day's fix in both tenors, and moved the 1-month up a touch despite Hayes's efforts, while Hayes did manage to move the 3-month down.
**** Investigators seem to find this downright unsporting, as in the case of RBS, which moved its swaps traders to sit next to its Libor submitters for a while. Eventually it stopped doing that, and, "When they were no longer in close proximity to the submitters, the traders increased their use of Bloomberg chats and instant messages to continue making requests for beneficial submissions, which were frequently accommodated." Which I'm sure the investigators appreciated.
To contact the author on this story:
Matthew S Levine at firstname.lastname@example.org