We talked the other day about municipalities and the Libor shenanigans. Quick recap:
1) Municipalities wanted long-term fixed-rate debt.
2) They got it indirectly by selling long-term floating-rate debt and buying interest rate swaps from banks.
3) At first, this was cheaper than issuing fixed-rate debt.*
4) Later, though, sometimes it turned out to be more expensive than having issued fixed-rate debt, or at least more expensive than it should have been, because municipalities pay a floating rate based on weekly reset auctions of their debt and that rate tends to track an interest rate called the Sifma swap rate,** while they receive a floating rate based on a percentage of Libor, and in 2007-2008 those rates diverged in weird ways.
5) Specifically, banks messed with Libor.
6) You can imagine tons of derivatives counterparties who could get screwed without politicians getting that worked up about it, but poor beleaguered Nassau County is not one of them.
Anyway an informed reader wrote in with some comments, of which this was my favorite:
There is an active swap market for Sifma based swaps. What this means is that all issuers had an option between choosing a % of Libor or Sifma as the floating leg. Since the underlying bonds trade at approximately Sifma, the Sifma based swaps have much, much less basis risk than Libor based swaps. The issuers have to pay for this smaller basis risk with a higher interest rate on the fixed leg of the swap. So all of those issuers that chose Libor based swaps had the option to minimize their basis risk, but instead decided to go with the "lower cost" Libor option. But by choosing Libor, they accepted that basis risk. So sure, if Libor was manipulated enough to cause them harm, they have some right to bitch, but if they hadn't been so aggressive when they originally entered into the swap, they could have avoided the mess all together.
Endorsed! Municipalities were hosed by the basis widening between Sifma and Libor, yes, but that's because (that basis was manipulated and also) they chose to get paid for taking that basis risk.
Sort of. Imagine the banks coming in to pitch a bond deal. Ten pages of flowcharts explaining how a swap works that you quickly skip over because who wants to talk about how a swap works?, then this:
Which do you do? As the reader notes:
So if you are a CFO of a municipality, why go through all the pain of describing the mechanics of a synthetic fixed rate structure to your board/city council/whatever if you're only going to pick up the marginal benefit between a Sifma based structure and fixed rate bonds. You might as well accept the greater risks and lower fixed rate leg of a percentage-of-Libor swap.
If you put a bunch of numbers on a page, and highlight them, and tell a client he gets to pick which one to pay, he will pick the lowest one. This is not a hard and fast rule but it's a good rule of thumb for, y'know, municipalities. The weird thing is that the bank is actually indifferent as to which one you choose: the Sifma swap market is less liquid than the Libor one, but still hedgeable, and the bank is just as happy to get 4.95% on a Sifma swap and lay it off at 4.90% as they are to get 4.75% on a Libor swap and lay it off at 4.75%.*** Moving not storage etc. The banks weren't pitching these Libor structures because they were more profitable, or easier to manipulate, than Sifma-based swaps. They were pitching them because they saved their clients money. Until the banks started manipulating them.
There's no real moral to this story except maybe that, while financial markets aren't necessarily efficient in all the right ways, they're efficient in one memorable way: if you pay less for X than Y (or get paid more for X than Y), it is because X has some risks that Y doesn't have.**** I've been fascinated, for instance, by the story of the Spanish bank hybrids: basically these banks sold zillions of dollars of very subordinated debt to widows and orphans, who thought that these 7% instruments were somehow as safe as nothing-yielding bank deposits. Turns out, no! Why did they think that? Maybe because the banks told them? Maybe because they didn't care? Seven is higher than zero, after all.
It's no accident that these stories involve retail depositors and municipalities rather than Facebook and GE. There will always be suckers and they will always buy the thing that is too good to be true, but there's no iron law that banks should always sell it to them. My imaginary termsheet above might present a harder choice if the Libor 4.70% number was circled in dripping blood-red with an oversized scrawl saying "if you do this we will manipulate Libor and you will lose money and you will DIE!!!"
Of course the banks didn't know that then, but I bet they're kicking themselves now for not pointing it out, or at least spending an hour walking every county commissioner through the concept of basis risk. Every time a banking product blows up in the public eye, banks really do go back and add a lot of very pointed shouty disclaimers to make sure that, the next time it blows up, they don't get blamed for it.***** But they can't give clear prominent understandable-to-the-county-commissioners warnings of every possible problem, meaning that they'll always bear some residual risk that the basis between the performance that they promise and the performance they achieve will shift, and they'll get pilloried and sued for it. Fortunately, markets being efficient and all, they're well paid for taking that risk.
* Later, it was often also cheaper than having issued fixed-rate debt, but more expensive than having issued floating-rate debt and not having done the swap, so, y'know, FRAAAUUUUD, whatever.
** In my previous post I sort of implied, like a lot of media coverage, that muni bonds float based on Sifma rates: i.e. like a floating-rate corporate bond whose contract specifies an interest rate of 3m Libor + 200, municipal bonds would specify a rate of Sifma + 20 or whatever. This sometimes happens now but is basically an anomaly, especially pre-2008: most municipal variable rate bonds "float" by being re-auctioned each week, and then paying a coupon based on the clearing yield of the reset auction. But since munis tend to be wrapped and semi-interchangeable they tend to float at around Sifma.
*** Except, of course, if the Sifma-swap-structure isn't cheap enough to persuade the municipality to do a variable rate bond plus swap, and they just do a fixed-rate bond. The bank is indifferent to Sifma vs. Libor swap, but swap >>> no swap.
**** There is a corollary of news market efficiency that if you buy X and those risks materialize the press will get very upset on your behalf, whereas if you buy Y and the risks of X don't materialize no one will be lining up to congratulate you for your prudence.
***** If you're all "why don't they just make a product that doesn't blow up?" you're doing it wrong.