February 28, 2013 4 Comments
So after commenting on some commentary about the purported $83 billion TBTF subsidy yesterday, I noticed that the discussion had already gotten way ahead of me. This should teach me to write half-cocked before getting up to speed by reading all the material. (Note: it won’t.)
To recap, Bloomberg wrote an editorial claiming that the TBTF subsidy amounts to something like $83 billion. Matt Levine wrote a piece detailing how that $83 billion is cobbled together, which (IMHO) showed that the $83 billion result is almost-certainly incorrect. That’s because along the way he showed that by having made other reasonable choices of some parameters and things, one could have come up with a value of $3.7 billion for that subsidy. Or, negative $16 billion for that matter.
The point, I assume, wasn’t to seriously put forth those other numbers as alternative models. The point was to show how the 83 result came from making certain arbitrary and not-entirely-defensible choices as to what to value and how to value it - other defensible choices for which, importantly, would have led to wildly different results. That means at best these numbers – 83, 3.7, -16, or 4 trillion for that matter – are all ‘correct’ up to some multiplier(s), and we just don’t know what those multipliers are. If you’re a ‘modeler’ you might think this means we’re getting somewhere. If you’re me (wait, just realized: am I a ‘modeler’ too? maybe!), you don’t have the patience for this so you just parse ‘correct up to a multiplier’ as ‘wrong’ and consider this all as symptomatic of the fact that this is a Large Calculation. (Also, if you’re me, my condolences.)
But I certainly do think that the value of TBTF is X and that X>0. My point is that I just don’t know how much value there is in all these calculations and ‘models’ whose resulting estimate of X has error bars that, if fairly accounted for, would likely turn out to be multiples of X. In the interest of illustrating this by blowing yet more smoke in the eyes of people who really want to try to answer this question and really want to put a hard estimate, any estimate, for X onto their Occupy Wall Street leaflets and placards, let me just mention some other factors that I think these models are omitting entirely.
- Decompression. While it’s presumably true that ‘TBTF’ helps, say, JP Morgan, if only by contrast it must be hurting any and all banks that aren’t considered TBTF. Right? I mean, here’s an industry. The US Government has come in and carved out a subset of the businesses in this industry, the bigshots, and declared their implicit backstop (or actually, it’s explicit now right?) for them. Surely this makes things harder for anyone not in that special subset to compete. Just like a government tax-credit subsidy for Wal-Mart hurts Mom & Pop, a government TBTF subsidy for JP – if there is one - must be hurting SmallTown Bank. However much it lowers spreads for the TBTFers, by comparison it must be harming the relative spreads of the non-TBTFers – leading to decompression. To the latter, therefore, that is a cost not a subsidy. And if the question is “How much does the TBTF ‘subsidy’ help banks?” where by “banks” you mean, like, “banks”, as in, all banks - then surely you need to include that cost in your accounting. No one seems to be doing so. Doing so might actually show TBTF to be a net cost not a subsidy. Who knows. (After all, all else equal I certainly don’t think government distortion of markets in this way makes people wealthier overall.)
- What about the protection value itself? My commenter Dave reminded me of this point when he put forth his alternative ‘model’ valuing TBTF at $4 trillion. Of course, he neglects to account for the fact that such a drawdown is, let’s just say, a low-probability event. But his point is not 0% valid. All the commentary has been focused only on how TBTF affects banks’ funding cost. But surely that’s not the only effect of TBTF. Surely there is also the contingent liability itself: the fact that, while these banks can take minor day-to-day losess (or even London Whale-sized losses) without the government stepping in, if indeed such a bank had such a loss that it threatened to play a “Fail” card, the government would summon a “Too Big To!”, the tapping of which would then – in that event – tangibly cost the government $X00 billion of mana. Another way to put this is to view the government as providing senior or super-senior (super-duper-senior?) credit protection to TBTF banks on, like, all of their risky activities (loans, trading, counterparty, repos, lines of credit, legal exposure to being sued for manipulating LIBOR or hiding drug money or helping tax evaders, fat-finger trades erroneously booked backwards, whatever else). If those activities’ losses never exceed some (unknown to me and you and everyone else) threshold, the government pays nothing; but if they did, then if TBTF means anything it means the government would be on the hook for some actual cashflows. This is a credit protection or insurance the government is providing, and it is worth something by itself, independently of how it affects bank credit spreads. How much is it worth? I’m not sure where banks could source equivalent super-senior tranche protection at the moment (there’s also the fact that USG is the only credible seller of this particular super-senior protection, which presumably means they could have overcharged for it if they were to charge market prices instead of giving it away for free). Golly, this is hard to say. So I shall just buttstimate** that it is worth something like 3-55bps on, like, the sum of the notionals and implied notionals of all banks’ liabilities or possible losses on anything. Note: I don’t know what that is at all but note: whatever it is, it’s actually a lot, and probably way, way more than $83 billion.
So in conclusion, if you take the above two bullets to heart, and do the same mental back-of-envelope addition of them that I’ve just done just now, I have probably helped convinced you as much as I am convinced that the TBTF subsidy is worth something between, oh let’s say, -$500 billion and +$1 trillion. I hope this knowledge helps you in your endeavors and protests.
‘Modeling’ is fun!
**buttstimate (v.): to estimate, by pulling a number out of your butt.