Customers queue to enter a Northern Rock bank branch in Bromley, in south-east London, in this 2007 file photo. Photograph by Ben Stansall/AFP via Getty Images
Customers queue to enter a Northern Rock bank branch in Bromley, in south-east London, in this 2007 file photo. Photograph by Ben Stansall/AFP via Getty Images

The other evening, I was interviewed by the charming David von Drehle for an event at the Kansas City Public Library. He noted that I had jeopardized my libertarian card by supporting the bank bailouts, a position that I still hold. We discussed that a bit, and when it came time for the Q&A, a fellow in the audience asked why, if I thought massive banking crises sort of had to be fixed by bailouts, I didn't also support deep reform that would put a stop to the crises in the first place.

I responded that I didn’t think such a reform was possible. And though I tried to explain why, I think it might also be worth laying out here.

The fundamental fact of a banking crisis, which is different from a crisis in any other industry, is that if people believe a financial institution to be bankrupt, it actually is bankrupt. As Arnold Kling puts it, banks exist to reconcile the desire of households to lend short and borrow long -- we want to have bank accounts we can empty at any time we want, but we want mortgage loans that carry fixed payments and last for decades. In financial parlance, the bank accounts are liquid -- it’s easy to turn them into cash -- and the mortgages are illiquid; if I want to get my money out, I have to find someone who wants to buy a mortgage on your house.

This creates a vulnerability at the heart of modern banks: If too many depositors try to turn their bank accounts into cash at the same time, the bank will be insolvent, because it cannot liquidate the mortgages fast enough to pay the depositors. And when are you most likely to get a flood of people trying to empty their accounts? When they think the bank is insolvent. So people who fear that their bank will go bust can actually produce that result, even if the bank is perfectly sound. Worse, if that bank fails, customers at other banks may panic, triggering a nationwide collapse.

This is unique to finance. That's why I didn’t support the bailout of General Motors Co.: As I put it at the time, if GM went bankrupt, my mini was still going to start the next morning when I put the key in the ignition. And the best way to stop a bank panic seems to be throwing money into the system. So I'm a great fan of the Federal Deposit Insurance Corporation. Ultimately, shoveling money into the banking system, almost indiscriminately, was the right thing for the government to do -- especially since, as it turns out, we made money on those investments.

It is impossible for me to read the tragic literature of the Great Depression, with its vast armies of unemployed looking hopelessly for jobs that weren’t there, and think that we should have engineered a repeat performance "pour encourager les autres." Bank bailouts are undoubtedly bad, but the alternative seems undeniably worse.

So why not build a banking system where this can’t happen? Well, alright. How shall we do that? Are we going to forbid the structure of modern banks (“fractional reserve banking,” for those who are interested in the technical term).

Leave aside the irony that outlawing a practice engaged in by millions of upstanding Americans is not, at first blush, a particularly libertarian solution. I don’t even think we could if we wanted to. Lending money at interest is illegal in countries that run by Islamic law, which has given rise to a very creative industry that structures bank transactions to just-barely-sorta look like partnerships or purchases. At least to my unprofessional eye, some of these practices look a lot like the repo market, which is where Lehman Brothers got itself into so much trouble doing essentially the same thing that bankers do: borrowing short-term money to fund longer-term investments.

I’ve heard arguments that you could fix this by allowing banks to freeze withdrawals in the event of a liquidity crisis. But I’m unpersuaded because during the Great Depression many places did freeze withdrawals. It doesn’t seem to have improved matters. At best, it solved the problem of the banks by creating a problem in the other credit-consuming sectors of the economy, like merchants and manufacturers who sold their goods on credit. Anyone who has worked in small business knows that a sound firm can still go under if a couple of big clients announce that they’re not going to pay. Or, for that matter, if you just have no orders, because suddenly no one has any cash.

I guess in theory we could go to a system where all transactions are in cash, up front, and bank accounts are safety deposit boxes that you pay rent on. The first problem with this idea is that people don’t want this and would get very upset. But the larger issue is that, however much I hate bank bailouts, I have a hard time thinking that it would actually be a better world if no one could, say, check into the hospital unless they had several hundred thousand dollars to deposit in case something went wrong.

Well, of course you don’t mean that … but then, where do you draw the line? I think you’ll find that banking, and credit booms, will creep back into your economy one way or another, no matter how hard you try to squelch them.

That emphatically does not mean I think there’s nothing we can do to make the banking system safer; like practically everyone else who writes about economics, I believe in higher capital requirements for banks, better resolution mechanisms for those that fail and more prudential regulation of how they can invest. What it does mean is that I don’t think these steps will ultimately make another crisis impossible. I just think they make the inevitable less expensive.