SearchUser loginNavigationTeam AgonistThoughtfulAbu Aardvark GlobalaliasBruce TimelyMixed Bag of Candy: Who's onlineSyndicate |
The Risks of "Spreading the Risk"My last corporate job was with an insurer. We used to write very large policies, and I regularly saw checks for hundreds of thousands or even millions of dollars cross my desk. But most of that money wasn't going to my employer. We underwrote the policies, sure. We issued them in our name, yup. When policyholders had a problem they called us, aye. But 90% of most of those policies was re-insured. When a policyholder made a claim, 90% of the millions of dollars we paid out came from other insurance companies, not from us. Of course, since we also were re-insuring other companies policies, when a claim was made in most of the other majors in our industry, well, we were chipping in for those too. Not 90%, mind, because there was a syndicate, but still a decent chunk. The reason this was done was because of the law of large numbers. With more insurance policies effectively on the book, but for lesser amounts of risk, we were less likely to see large ups and downs in our claims due to random luck. (For a simple example, if you flip a coin once, it's hard to predict whether it'll come down heads or tails. How many heads and tails will it come down if you flip it a 100 times? A thousand? Ten thousand? The larger the number of times, the more likely you are to get closer to the number you would expect based on probability.) But here's the downside. When you do this you make one very large risk pool. If something does go wrong, perhaps because you're all using the same risk tables, the same basic underwriting guidelines and so on (you have to, or the others won't take on the risk), you're all on the hook. Instead of the contagion just hurting one insurer who screwed up by having a bad risk model, or lousy underwriting,or bad investment experience, or fraud, or what have you, it hits everyone. Now the theory is that "everyone" together is stronger than any individual company. But it's like going to a monoculture in agriculture. If everyone's growing different crops, and within crops different strains of a crop, when a disease hits, or a pest, or even a drought, or a rainy season, some of those strains or crops will prove more hardy, and may well survive. When you have a monoculture, the disease or pest can rip through it all and destroy everything. This is especially true if you are, in effect, constantly sharing seeds and mature plants -- not only are you vulnerable, but you're automatically passing on the disease. To use the modern phrase, there's no firewall. By all standing together you may all die together. By all having the same business model, the same financial models, the same underwriting models, the same investment models, you've made it so that any financial contagion is going to hit you all, hard, and will spread easily amongst you. You're all vulnerable. Because of this I've always been very suspicious of the idea that securitization and insurance and re-insurance between financial institutions actually reduced risk. What it did, it seemed to me, was make small disasters much less common, and make having a big catastrophe much more likely. We're now getting a test-out of whether the theory that "spreading the risk" was such a hot idea. The men who lived through the Great Depression thought it was a bad idea. That's why the Glass-Steagall Act, for example, made consumer banks and investment banks stay apart. Its why banks couldn't sell insurance. It's why brokers weren't banks. It's also why banks didn't used to be able to buy "default insurance" to reduce their outstanding loan value so they needed a smaller reserve -- because, at the end of the day, that didn't actually reduce the amount of risk, it spread it. (And when push comes to shove, if you issued the loan, and everyone down the chain goes bankrupt, odds are high it'll end up in your lap.) Bigger isn't always better. More interconnected isn't always better. Spreading the risk sometimes just spreads the risk. And the monoculture of business models, where everyone was doing the same things and using the same risk and profitability assumptions, not only doesn't reduce risk, it increases it. The older model, where each entity was responsible for its own risk, and where if it failed, it failed, was a better one. There was some socialization of risk both for businesses (banks through the Fed, for example) and for individuals. But there was also an attempt to keep businesses out of each others' pockets and an attempt to split up different types of business so that a failure in one area didn't become a general financial sector collapse. It's time to stop pretending that the people who saw the 20's and the Great Depression were fools and that we are so much smarter than them. The system they put in place led to the greatest period of prosperity the US has ever seen. Perhaps it needed modifications, but it didn't need the sort of wholesale repeal we've witnessed. Wisdom, it is said, is learning from other people's mistakes. It's too late to be wise, but perhaps, just perhaps, we might learn from our own? Ian Welsh November 9, 2007 - 12:00pm
( categories: Economics | The Markets )
|
![]() |