With financial regulation, there’s a problem, akin to the problem of seeing the forest for the trees, which I will call the problem of the weeds and the swamp. In an environment as ugly and overgrown as that, it’s hard to tell what’s a noxious invader, and what’s a healthy and important part of the ecosystem. Which means there’s always the temptation to stop working so hard at managing the swamp, and simply let it go to seed.

Take section 716 the “push-out” provision of Dodd-Frank, which requires banks to trade uncleared derivatives out of uninsured subsidiaries rather than mingling them with the FDIC-insured parent bank. The provision was a response to one driver of the financial crisis – the fact that large banks (prominently Citigroup, who lobbied heavily and repeatedly for the provision’s repeal) held huge positions in senior risk associated with pools of mortgages that they had securitized (or, even more dangerously, in pools of asset-backed securities based in turn on such pools of mortgages).

Because of their seniority, the risk associated with these positions was treated as nugatory by the ratings agencies and, consequently, by the capital rules that govern such banks. The underlying mortgages would have had to fail at historically-unprecedented rates, and the recovery values for the mortgaged properties would have to be unprecedentedly low, for the pools to lose enough value to create a liability to the bank. So the banks were allowed to say that this simply wouldn’t happen, and hold capital accordingly.

And then, of course, it happened – and the banks had to be bailed out.

The push-out provision forced FDIC-insured banks to put such activities in separately-capitalized subsidiaries. This was supposed to prevent the catastrophic failure of such a swaps dealer from bringing down the insured bank, requiring a bailout.

There are several problems with this theory, however. First of all, AIG was structured in exactly this fashion. And AIG required one of the largest bailouts, notwithstanding that it wasn’t FDIC insured – indeed, it wasn’t a bank at all. Second, if I understand correctly, the uninsured bank can still use highly customized swaps for hedging. This is what Goldman, for example, did with their senior risk – they hedged it. With AIG. One of the scandals of the AIG bailout is precisely that it was done on terms that made Goldman whole, even though, logically, Goldman should have known that if it ever needed to make a claim on its “hedges” with AIG, AIG would never have been able to pay (since such a claim would imply that there had been extraordinary, historic losses in the mortgage market, and such extraordinary, historic losses would surely have already wiped AIG out, as indeed they did).

My gut feeling is that the provision was a form-over-substance attack on the real problem, and one that I don’t think would have been particularly efficacious in the run up to a 2008-style crisis. So perhaps we shouldn’t be so worried by the prospect of its being scrapped.

Except that I am worried.

From my perspective, we’ve never grasped the nettle of what we want our banking system to do. Classically, banks turn savings into capital, and classically they do it pretty directly, by taking deposits and lending to businesses and consumers. Banks are supposed to be good at evaluating the individual risks associated with these loans, as well as constructing portfolios of sufficient diversification to minimize the risk of them getting any of those individual evaluations wrong. Bank regulation is supposed to keep banks from pushing the envelope on either front in the interests of higher profit.

In our modern world, most of what banks do is intermediate in more complicated ways. They don’t do a whole lot of evaluation of individual loan decisions; instead, they are expert at aggregating and repackaging financial risk. The upside to this is that a much larger world of borrowers can access a much larger world of lenders, which should push down the cost of lending generally, and thereby facilitate growth. The downside, which has only manifested itself over time, is a financial system with an ever-increasing aversion to real risk (because nobody involved in large institutions can really evaluate it, and small institutions cannot achieve economies of scale to compete effectively with the large ones), coupled with an ever-increasing appetite for complexity that provides inherent reflexive self-justification (you need a really complex and well-paid operation to manage that complexity, so finance can never reliably be shrunk as a percentage of the economy).

I can’t help but believe that financial regulation and innovation in capital rules has facilitated the above development. In a world where hedge funds do more and more of the direct lending to small businesses – the kind of activity that used to be bread-and-butter for FDIC-insured banks – whatever financial regulation is doing, it isn’t forcing insured banks to return to their “proper” core economic function. And the kinds of activities that 716 tries to push out into uninsured subsidiaries are precisely the kinds of activities where much of the real risk lies in the tail of the distribution – the place where crises come from.

The repeal of 716 matters not because “push-out” was such a successful or important reform, but because it shows how the terms of debate have changed. We’re now debating whether it’s “time” to weaken financial reform, and let the banks operate more “efficiently,” when financial reform never actually achieved its core goal of taming finance’s role in the economy.

The core question is not whether uncleared derivatives are too risky to live in an insured bank, but whether we’re creating incentives to take risk or to hide it. Banks are supposed to do the former. They have a lot of natural incentives to do the latter. Financial regulation needs to be exceptionally vigilant about sniffing out that kind of behavior and punishing it. Encouraging banks to stuff those risks in an uninsured subsidiary that could then, in a crisis, bring down an insured bank doesn’t strike me as an especially auspicious solution. Scrapping that solution because it isn’t capital efficient, and replacing it with nothing, is even more alarming, because it implies that no solution is necessary. Bankers and politicians agree: the problem no longer exists. Which is exactly when we should start worrying that it’s about to smack the economy in the face yet again.