Why We Still Need Securitization

The NYT's op-ed page is positively infested by econobloggers today: there's not only myself, but also a certain Princeton economics professor whose blog you might have noticed. His column today echoes the big Simon Johnson essay in the Atlantic, bemoaning the decades-long rise in the power of the financial sector, which still seems to have US governments both red and blue entirely captive:

Mr. Summers needs to get out more. Quite a few economists have reconsidered their favorable opinion of capital markets and asset trading in the light of the current crisis.
But it has become increasingly clear over the past few days that top officials in the Obama administration are still in the grip of the market mystique. They still believe in the magic of the financial marketplace and in the prowess of the wizards who perform that magic.

I'm sympathetic to this view: financial intermediation, in its various forms, should never account for 41% of total corporate profits, as it did this decade. That's closer to a Ponzi scheme than it is to value creation. But I disagree with the distinguished professor on the subject of securitization.

The key promise of securitization -- that it would make the financial system more robust by spreading risk more widely -- turned out to be a lie. Banks used securitization to increase their risk, not reduce it, and in the process they made the economy more, not less, vulnerable to financial disruption...
I don't think the Obama administration can bring securitization back to life, and I don't believe it should try.

I have a feeling that, contra Krugman, securitization really does hold out a lot of promise for the banking system going forwards. Much of the current crisis stems from the fear that too-big-to-fail institutions are in fact insolvent. And while one can talk about interconnectedness as being just as important as size, size still matters: if you have a trillion-dollar balance sheet, you're too big to fail. And so any financial technology which can reduce the size of a bank's balance sheet is a good thing.

Securitization -- real securitization -- is a great way of moving assets off a bank's balance sheet. The problem is that starting in the late 1990s, with JP Morgan's Bistro deal, banks stopped taking their assets, bundling them up into securitized bonds, and selling them off. Instead, they kept the assets on their balance sheets, and sold off synthetic CDOs which, according to their models, perfectly hedged the credit risk of the assets in question. Thus, instead of getting smaller, banks' balance sheets got much, much bigger. And when the models proved to be flawed, the write-downs that the banks needed to take on their "super-senior" assets (the ones which supposedly had no credit risk at all) amounted to hundreds of billions of dollars.

CDOs are not eligible to be bought as part of Treasury's bank bailout plan. Instead, the plan calls for investors to buy the actual underlying assets, thereby reducing the size of the banks' balance sheets. This is a good thing. And any sustainable recovery will be predicated on the existence of more such deals: risk being moved from the books of people who don't want it, onto the books of people who do. The problem with a lot of what looked like securitization over the past decade was that many banks thought that they'd sold off all their risk, when in fact they hadn't. The way to solve that problem is to move back to securitization 1.0, not to abolish securitization entirely.