There’s nothing surprising, exactly, about this chart that Fitch sent out today, but it’s still sort of stark:
Once there was a land where bank debt was AA, AAA if it was particularly good or A if it was particularly dicey. Now AA is the new AAA and BBB is commonplace. The idea of risk-free unsecured lending to banks, implicit in things like Libor discounting, is over.
Right? I don’t entirely understand this proposal by House Republican John Campbell to require banks to “hold substantially more capital,” though the gist is basically that there’s a move to require banks to do more of their funding via long-term holdco debt. Here is a puzzling summary:
Campbell’s bill would require banks with at least $50 billion in assets to hold an additional layer of capital in the form of subordinated long-term bonds totaling at least 15 percent of consolidated assets. If an institution were to fail, the long-term bondholders would be guaranteed at no more than 80 percent of the face value of the debt.1 As a result, banks would face pressure to reduce their balance sheets.
But the gist is clear enough: it’s part of the “living will” movement to make it easier to kill a bank painlessly. Holdco unsecured debt is a lot easier to wipe out than, like, overnight repo funding, and requiring lots of it, and making the holdco the place to start any bank bailout, will have the effect of “convert[ing] senior unsecured debt into contingent capital”:
It is far more likely that holding company debt will become the resolution buffer, and therefore that minimum levels of such debt will be mandated. What such a level might be is unclear, but there could be plans to introduce legislation in the US calling for banks with more than $50 billion in assets to have subordinated long-term debt equivalent to 15% of assets.
This is sort of a neat plan: it’s a way to force banks to “raise more capital” without actually raising that much more capital. Holdco debt isn’t “capital,” in the ordinary sense of the word, for one thing2; for another, a lot of big banks already have long-term funding pretty close to 15% of assets:
But it’s not nothing, is it? The risk-free bank liabilities idea is fading but not quite gone; Moody’s last big round of bank downgrades nonetheless included 1-2 notches of uplift, even at the holdco level, from expected government support. Explicit government rules distinguishing holdco debt that really really really won’t be bailed out we promise this time, from systemically important obligations that aren’t government guaranteed but y’know nudge nudge wink wink, might make bank credit pricing, and ratings, more transparent.
A key cause of too-big-to-fail is, if you expect that your investment in a financial institution will be money-good, and you’re politically connected and/or systemically important and/or sympathetic/widow/orphan, then, in a crisis, what is the government gonna do? Not pay you? You had an expectation, and based on past practice that expectation was reasonable, so you kind of gotta get paid. Changing that expectation, even by sort of arbitrary measures like just saying “your type of debt is no longer too big to fail,” goes at least part of the way toward fixing the problem.
1. What? Guaranteed by whom? Right now your bank holdco debt is (1) 0% explicitly guaranteed by the government, (2) 100% x P implicitly guaranteed by the government, where P = probability of bailout that preserves holdco debt, and (3) explicitly guaranteed by the bank so long as it has money left over after paying senior claims. Making it 80% explicitly guaranteed by the government makes it more senior than it was before. Capping payouts by the bank make it … just confusing basically.
2. Random aside re: capital: Matthew Kleinman has an interesting argument for the Admati & Hellwig claim that banks should have way, way more capital (like, 25% of assets), with a chart that I find kind of amusing:
Basically: no matter what the assets of the banking system, its equity stays the same. It levers up and down over time but doesn’t add equity. Kleinman seems to accept “because bankers are dicks” as the explanation here; my toy explanation is “because the size of the banking system reflects primarily demand for bank liabilities and so of course those liabilities fluctuate 1-to-1 with its size.” That can’t be right – it’s a toy explanation – but, I mean, neither can “bankers are dicks.” Anyway, worth pondering.