There's a hedge fund hiding in there somewhere.
There's a hedge fund hiding in there somewhere.

Two things that I took away from Zions Bancorporation's run-in with the Volcker rule today are:

1. I seem to have been wrong about one important effect of the Volcker rule, and
2. Zions had a conference call in which its chief financial officer said "we got to that conclusion in part by looking at footnote 1,861," so now my dream in life is to have a bank chief financial officer cite a four-digit-numbered footnote of mine; look forward to that.

One thing I had thought about the Volcker rule is that it penalizes short-term proprietary positions held for trading, making those proprietary positions much less attractive (slash possibly less legal) than long-term proprietary positions like "loans." Hold-to-maturity positions, I'd thought, would be safe from the Volcker rule, which explicitly allows prop trades that are intended to be held for more than 60 days.

So it's weird to see Zions going the other way:

Zions anticipates that in the fourth quarter of 2013 it will reclassify all covered CDOs that currently are classified as “Held to Maturity” into “Available for Sale,” and that all covered CDOs, regardless of the accounting classification, will be adjusted to Fair Value through an Other Than Temporary Impairment non-cash charge to earnings.

Basically they had a bunch of collateralized debt obligations that they were planning to hold to maturity and so they were umm let's say being a bit fuzzy about how much those CDOs were down on a mark-to-market basis. And then along came Volcker and, bang, they had to reclassify them as shorter-term holdings and mark them (down) to market.

One should not feel too sorry for them about that. As my colleague Jonathan Weil puts it, "the accounting rules had been letting Zions maintain a fiction." I've argued that maintaining that fiction is part of the point of being a bank but you could reasonably disagree.

But why would the Volcker rule require Zions to take a long-term real-money investment and turn it into a proprietary trading position? The thing is that the Volcker rule doesn't just prohibit "prop trading," which these CDOs aren't. It also prohibits banks from investing in hedge funds, which these CDOs ... I mean, I think it would be fair to say that these CDOs aren't that either? (If they are then they are terrible hedge funds, but that tells you nothing.)

But "hedge funds" are not really a thing, so what the Volcker rule actually prohibits is investing in things that would be investment companies but for sections 3(c)(1) or 3(c)(7) of the Investment Company Act of 1940, which say that if you limit yourself to a small and/or accredited pool of investors you don't have to register your investment company as a mutual fund. That's what hedge funds use to get out of mutual-fund restrictions, but it's also what a lot of securitizations use for the same purpose.

If you like securitizations, this is probably not a thing that you like. And so in fact the people who like securitizations moaned about it quite a bit, and in the final rule there are pretty robust exemptions for securitizations of loans, which includes collateralized loan obligations and perhaps other things (presumably mortgage-backed securities would count as loans).1

But securitizations of non-loans are off limits to banks under the Volcker rule,2 and so Zions will probably have to get rid of its portfolio of bank and insurance trust preferred CDOs.

Which is not like the end of the world, or even the end of the bank trust preferred CDO market. That ended several years ago; here is a 2011 Federal Reserve working paper titled "The Trust Preferred CDO Market: From Start to (Expected) Finish," which doesn't even mention the Volcker rule. The market died through a combination of bad performance and repeal of the capital rules that made TruPS attractive for bank issuers.

It is probably too late, and too uninteresting, to save Zions' TruPS portfolio, but there does seem to be a structuring opportunity here. Stuff wants to be securitized,3 and not all of that stuff is naturally characterizable as "loans." The natural way to securitize things is to put the things into a (static) pot, slice up that pot, sell the highly-rated bits of it to banks, and sell the low-rated bits to hedge funds and other non-capital-regulated actors. That pot does not feel particularly like a hedge fund but, for Volcker rule purposes, now it is. And so banks can no longer touch it, unless it is entirely filled with loans.

So one option is to fill future pots with loans -- goodbye bank trust preferred stock, hello highly subordinated loans to banks. Another option is to slice the pot differently: Sell a pool of whole assets to a bank, say, and have the bank buy insurance from the hedge funds who'd otherwise be buying the subordinated tranches of a securitization.

I dunno. But securitization itself is largely a creature of regulatory arbitrage: a way to put a bunch of stuff worth $100 in a pot and sell slices of the pot for $102, because the slices got ratings or capital treatment or whatever that made them more attractive to buyers than the undifferentiated stuff in the pot. The apparatus that grew up around that regulatory arbitrage is pretty robust. It's hard to imagine that it would be stymied by one regulatory footnote.

1 Incidentally Zions said about its CDO revisions, "This is not something that we had anticipated nor do we think we reasonably could have anticipated based on what was on the proposed rule," which ... I mean, does not seem to be strictly true? Because it was in the proposed rule? And because lots of people anticipated it, and moaned about it? I mean, I don't want to be too critical; I certainly didn't anticipate it. But securitization-y people did.

2 Unless of course they're permitted by footnote 1861! You can find that on page 538 of the Volcker rule release. Basically there are other '40 Act exemptions that might work, particularly for registered securitizations. Always read the footnotes is the lesson here, and everywhere.

3 Does it not?