Why AIG was in the CDS Business

Gari has a question about AIG, in the comments:

Why did a well-capitalised insurance company with plenty of long-dated liabilities decide that it was a better use of shareholders' and policyholders' cash to write protection against debt instruments rather than owning them? I think part of the reason was the anomaly in the pricing of CDS compared to bond yields that the CPDO tried to exploit. I don't know whether I'm just raging against accounting arbitrage here, but doesn't any examination of systemic risk and CDS need to focus on this question?

You don't need to know about CPDOs or about the spread between CDS and bond yields in order to answer this question. It's much, much simpler than that. The fact is that AIG didn't use "shareholders' and policyholders' cash to write protection against debt instruments". If it wanted to buy bonds, then it would have needed to come up with some cash to do so. But writing protection, by contrast, was a way of receiving money, not spending it.

When AIG wrote protection on CDOs and the like, it got insurance premiums in return, and considered those premiums to essentially be free money, since (according to AIG's own models, and those of the ratings agencies) the chances of those CDOs defaulting were essentially zero.

Now, of course, it's clear that those insurance contracts constitute an enormous contingent liability for AIG -- one so big that without government help the company would have gone bust. But at the time, no one at AIG was worried about that, so busy were they raking in the dollars insuring CDOs which they were positive would never suffer any losses.

AIG's biggest mistake was in failing to realize that this business couldn't scale in the way that most insurance does scale. Most insurance does scale: if you insure a house against fire, for instance, it's easy to lose much more money than was paid in insurance premiums. But if you insure houses across the country against fire, you'd need a nationwide conflagration in order to lose lots of money.

The CDO market doesn't work like that, however. The reason AIG's models said the CDOs couldn't suffer any losses was that house prices don't fall in all areas of the country simultaneously. Since AIG was only insuring the last-loss CDO tranches, investors with lower-rated tranches took the risk that prices in Florida, or Arizona, or California might fall. AIG would only lose money if prices fell in all those states at once -- which is, of course, exactly what happened.

But AIG never stopped to think that the event which would precipitate a payout on one CDO was exactly the same event which would precipitate a payout on all the other CDOs as well. AIG could easily afford any given CDS contract. What it couldn't afford was lots of CDS contracts -- because with CDS, unlike with most insurance, there was no safety in numbers, only more danger.

The investors in CPDOs, at least, put their money up front, and looked to make their relatively modest profits slowly, over time. They lost their money, but at least they had their money before they lost it. At AIG, the financial products group booked its profits immediately, without spending any money at all. When their losses arrived, the firm had to scramble to find the cash, since it had never allocated much in the way of capital to the group.

Insurers are always happy to take your money. But spending money on insurance is always fraught. You've spent your money up front, and now you hope that if the thing you're insuring against comes to pass, the insurance company will do the right thing and pay out. Your big fear is that they won't, either because they think they've found a reason to reject the claim, or because they've gone bust.

That's why insurers need to be very highly regulated. If they weren't, anybody could set themselves up as an insurer, take in lots of premiums, and then simply disappear. But that's also why AIG was writing protection on bonds rather than buying bonds outright. Under the insurance model, you can rake in your premiums and provision very little capital against them, so long as you wow your regulator with enough whiz-bang models saying that you'll never need to pay out on those policies. If you buy a bond, by contrast, the seller wants cash up front. And where's the fun in that?