Banks Being Sued For Systematically Manipulating Libor Up While Also Systematically Manipulating It DownBy Matt Levine
A probably important and genuinely difficult question is: all that Libor stuff, did it affect your mortgage? Probably important in that in expectation (1) you have a mortgage and (2) honestly you don’t really care about what banks do otherwise1 so you need to know how mad to get at them. Genuinely difficult at least because it is hard to measure how much banks manipulated Libor, or even in which direction, though overall it seems to have been mostly down, implying that floating-rate mortgage borrowers paid lower rates in the aggregate than they would have in an unmanipulated world.2
But lawyers abhor a vacuum so now there is this lawsuit, in which the homeowners are suing banks for manipulating Libor up, costing them (the homeowners) money by making their adjustable-rate mortgage payments higher. Now you might think that this sits uneasily with the widespread assumption that the banks mostly manipulated Libor down, at least during the crisis, to make themselves look better, but in fact the complaint has a super-simple way of measuring homeowner screwedness that sort of refutes your objection:
61. Defendants, the banks that comprised the U. S. dollar LIBOR panel during the Class Period [January 2000 to February 2009], were motivated to manipulate and increase LIBOR on or about the first day of each month because they knew that most adjustable rate mortgages and promissory notes contained a clause establishing the first day of the month as a “Change Date” and that the new rates would be set on that day. …
64. Throughout the Class Period, the LIBOR 6 month rates on the first business day of each month are, on average, more than two basis points higher than the average LIBOR 6 month rates throughout the Class Period. Additionally, from August, 2007 through February, 2009, the LIBOR 6 month rates on the first business day of each month are, on average, more than seven and one-half basis points higher than the average LIBOR 6 month rates. Finally, the LIBOR 6 month rates on the first business day of each month are, the great majority of the time, higher than the five-day running average of the LIBOR 6 month rate surrounding the first business day submissions throughout the Class Period.
That is cool is it not? If you are reporting Libor every day, you might just report your actual cost of funds on the second Thursday in March because, y’know, who cares, why commit crimes every day. But if zillions of dollars of contracts reset on the first day of every month, you might be extra careful that day to report a fake Libor that maximizes your profits on those contracts.
Is it true? I dunno, I went to Bloomberg and got the numbers and here are some:
So! I can replicate their 2bps and 7.5bps numbers; in fact I get slightly higher differences, with their special 2007-2009 period actually showing about 8bps higher Libor on the first of the month than on any subsequent days. I can’t replicate the five-day running average thing; as far as I can tell the first-of-the-month Libor averages out to within a tenth of a basis point of the surrounding days everywhere except during that special 2007-2009 period, where it’s still within half a basis point, and where the first-of-the-month Libor is actually lower, on average, than the surrounding days.3 I think that last line gives you some useful context: in that super-intense Libor manipulating period of 2007-2009, Libor was just bouncing around a lot, because (1) there was a credit crisis and (2) it got better, so it’s not that surprising that one subsample of Libor data points was roughly one daily standard deviation away from the broader sample. That puts the evidence of manipulation in context but doesn’t really get rid of it. The first-day Libors are all around a standard deviation up, not down, from the other Libors.
So I dunno if this is true. But isn’t it interesting? My favorite thing to ponder is: if you take this at face value – banks faked Libor up on the first day of every month – and combine it with another important stylized fact about the 2007-2009 period – banks faked Libor down generally – where does that leave you? Like, we talked a while back about the municipalities suing the Libor banks for making Libor too low during roughly the same period, and we calculated a reasonable guesstimate of the amount of that manipulation at about 11-13 basis points (the municipalities claim 20bps). So if you are a homeowner with a floating-rate mortgage, do you get mad at your bank for popping Libor up 7bps on the first of every month – or do you thank them for holding it down by 11bps for most of the month, and then only popping it up by 7bps on your reset date, giving you an interest rate on your mortgage that is still 4bps too low? If you are a court, do you give homeowners damages because banks’ monthly counter-manipulation deprived them of the full benefit of banks’ daily regular manipulation? Do you instead say “haha, homeowners, now that I’ve got you here, I’m going to take money from you and give it to those municipalities who actually lost out”? Or what? I dunno.
You can ponder other things too.4 Why is Libor on the first of the month higher than on other days? Is there a non-manipulation explanation? And if so what is it? Do banks want to borrow more on the first of each month and thus push up rates? Why, and what would it mean if they did?5
Or: if it is manipulation, what does that tell you? One thing it would tell you, if true, is “banks basically make money on higher Libors,” which is why they would manipulate them up on reset dates. That in turn would tell you that if in fact the banks were manipulating Libor down generally, they were sacrificing real economic profits in order to look more solvent: sacrificing the reality of profit for the illusion. Which wouldn’t be that big of a surprise.
Banks Sued by U.S. Homeowners Over Rigging of Libor Benchmark [Bloomberg]
Complaint [SDNY via Bloomberg]
1. The “you” in this expectation is, once again, not you as in you. I mean, you you are here. In the footnotes!
2. But also because there are also weird expectations effects that: if your mortgage priced at L + 300, because people want you to pay 275bps over bank borrowing rates and they figured Libor was 25bps too low, but then Libor was rendered more honest by publicity or Wheatley or the threat of jail or whatever, then at your next reset you are paying too much on that mortgage. To me, this is a bigger deal than the monthly manipulation, but it is even less susceptible to knowledge.
3. For fun I did the same thing for 3-month Libor and got similar results.
Arguably this … tells you what? That banks weren’t manipulating 6-month Libor for mortgage resets because the same thing was happening in 3-month Libor? That banks were manipulating 3-month Libor for corporate loan resets the same way they manipulated the 6-month for mortgages? Other?
Also, just for giggles, here is a chart of (1) Libor on the first day of each month minus (2) the average of Libors over the previous ten days. I think you’d read this chart to say “Libor was mostly going up in 2005-2006, going down in 2007-2008, and being crazy in 2009,” but feel free to offer other interpretations:
4. Yet another one is: can this be right? The Cleveland Fed put out a paper last week saying “In our view, an appropriate reference rate [for home mortgages] would have (at least) two properties: It would be free from concerns about manipulation, and it would represent a low-risk benchmark interest rate much like the Libor was before the financial crisis.” In other words, the rate at which you borrow from your bank should not reflect broad banking sector credit risk. This seems perverse to me but it’s a Fed, so, thoughts?
5. A fact in the world is that banks tend to unwind repos at around quarter end, so that their quarterly financials show them as being less levered than they really are. One way to look at that fact is “banks unwind repos to fool people into thinking that they are less levered than they really are,” which is called “window dressing,” and that is a perfectly sensible view. But banks will tell you that they are unwinding the repos due to changes in customer demand, and you can treat that claim skeptically if you like, but if you believe it 100%, you are still left with an oddity, which is: why are customers unwinding repos at quarter end? And the answer is, I guess, “because they are also shrinking balance sheet at quarter end to make themselves look less levered,” meaning, I guess, that they are window dressing. Someone is window dressing and you can wander around wondering whom to blame or, just, not. There’s a quantum of shadiness in the world and you can pass it around but it’s, like, somewhere, right?