mathbabe

Another way of getting a rough idea of whether the subsidy is big or small is to ask, how hard did they fight to keep it? I’m this case, I think the answer’s pretty clear: we went through the worst financial crisis we’ve had since the 30s, and thanks to Wall St lobbying & members of the revolving door brigade, we’ve basically done nothing substantial to prevent another too-big-to-fail crisis. That’s insane, both from a common sense, business, and basic Econ perspective. Even someone like Volcker thinks its insane. So whatever subsidy the banks are getting from this, it must be pretty damn big. If a model tells you the subsidy is pretty small, it’s time to take another look at the model’s assumptions. It’s still possible that the model is right and that Wall St fought tooth and nail for some other reason. But it’s pretty unlikely.

LikeLike

One of the problems with a term like “taxpayer subsidy” is that it’s vague: leaving aside whether “taxpayer” is the right subject (banks pay taxes, no? So does Jamie Dimon – should we exclude them from the calculus?) I think you need a more precise null hypothesis about the form a subsidy might take before you can realistically calculate its scale.

Presumably the subsidy you have in mind is something approximating a promise by the government (taxpayers) to banks’ creditors that if a bank becomes insolvent, their obligations will be funded by the state. It probably doesn’t matter if the promise is explicit or implicit, but what does matter is whether creditors believe that the promise exists and whether it will be honored.

My central issue with Bloomberg’s original editorial was that it treated all creditors as homogenous. They’re not, which is probably why the authors of the paper that Bloomberg sourced took good care not to assume that they are, and solved accordingly. Banks typically fund their asset base via a wide range of borrowings: deposits, overnight repos, short and long term wholesale debt, various hybrids…. and also retained earnings and equity. So if there is a taxpayer subsidy to banks, presumably it enables the banks to pay less than they otherwise might on one or more of these sources of funds?

Bloomberg insists that it’s valid to look at deposit prices in the context of a government guarantee, and has another paper that supports its position. The problem with the second paper is that it is quite explicit about NOT being able to quantify exactly what portion of the interest rate differentials it posits is due to government support, and it’s easy to come up with other explanations for differences in reported deposit rates. All in, I don’t see conclusive evidence that depositors are willing to lend money to TBTF banks at rates that are substantially lower than those at which they lend to smaller institutions.

Hybrid securities are probably a fruitful place to look at risk-pricing, because they combine elements of equity and debt and are priced to reflect both. The problem that you run into when you start looking at them is that they’re fairly idiosyncratic instruments, and it isn’t clear whether you could use these to compare the perceived risk among different issuers or whether their functionality would be limited to a longitudinal examination of a particular issuer.

When you take out the portions of banks’ activities that are funded via some form of collateral – such as repos – and are therefore priced according to the value of that collateral, you’re left with straight bonds, and this is where I’d focus if I wanted to model the value of a government guarantee – presumably, this is where the value of a subsidy principally accrues, given the issues described above. I’d also note that this is where the authors of Bloomberg’s source paper focused their attention, and re-air my original complaint about Bloomberg’s unwarranted extension of their findings to other areas of the balance sheet.

Rather than the debt itself, though, I’d probably look at CDS if I wanted to examine the perceived credit of a given bank. There are problems with this, because it’s been widely documented that you don’t need an insurable interest to purchase a CDS, but I think this best addresses the point I raised at the outset – what matters is firstly which lenders believe their loan has been guaranteed AND whether that guarantee will actually pay out. I haven’t looked at the data in a systematic way (still less modeled it), but IF there is an expectation that the government will make large bank bondholders whole if it fails in a way that they won’t for a small bank’s creditors, that differential would presumably be reflected in the cost of insuring those banks’ debt.

You should be able to control for a lot of the factors that drive pricing, and what you’re left with is the value that debt holders place on a guarantee from taxpayers.

Having said all that, you’ve actually asked (or implied) a slightly different question: perhaps I’m wrong, but it sounds as though you’re wondering what the cost to taxpayers is…. and I’m REALLY not sure how to approach that. Taxpayers have a contingent liability, not an explicit cost. You would think that the price of that contingent liability would be reflected in the overall cost of government debt, right? But that’s gone nowhere but down, so we’re left with Matt Levine’s conclusion: that actually, the TBTF banks are subsidizing everyone else.

Tricky, no?

LikeLike

Gordon,

You are brilliantly illustrating how we can delve into the details and use obfuscation to justify what we wanted to believe in the first place.

You are right. The question Mathbabe/Bloomberg is not perfectly formulated. And, as you show, we can get into abstruse questions of how to define and measure it.

Here are some broader, simpler and more important questions:
Is there a widespread perception in the market that the mega-banks will be bailed out if they need it?
Is that widespread perception undermining market discipline that would ordinary constrain/monitor the risks the megabanks can take?
Is that widespread perception distorting the market in other major ways?
Have the megabanks avoided prosecution because of fear of “horrible collateral consequences”?
Did all of these factors cost the US taxpayer billions of dollars in the past crisis?
Did all of these factors harm the US (and world) economy?
is there any reason to believe the problem has been addressed?
is there any evidence that the existing government/bankers/regulators will address it in the near future?

Considering that my answers to these questions are: Yes, Yes, Yes, Yes, Yes Yes No and NO that leads me to the last two questions:

What can we work to make sure that it will be addressed soon?

Two things have made me more optimistic that this effort has promise. First, there is a broad consensus that this is a serious problem. For instance, two Democratic Senators (Warren and Merkley) and two Republican Senators (Vitter and Corker) made a major issue of this yesterday when Bernanke testified before the Banking Committee. This has also been raised by current and former Fed officials, other bank regulators, bank analysts, libertarian economists, occupy Wall Street people, people on the street and nearly everyone who thinks seriously about the issue. I follow this closely and their aren’t many who disagree who are not employed by the megabanks in Wall Street or in Washington.

Also, the problem is easier than I believed until two weeks ago. Read “The Bankers’ New Clothes”.

LikeLike