Tuesday, March 18, 2014

It’s Utterly Mad



For centuries, scientists have searched for ways to mix different materials to create gold. In 1995, David Li, a thirty-something math whiz from rural China, was doing something similar with loans. Li was trying to figure out how to mix risky loans together to get risk-free ones.

Surprisingly, his great insight came from death. Li knew about the “broken heart” problem, in which people die more quickly after their spouses die. Li saw an analogy to loan defaults. When one borrower defaulted, others were more likely to default. Not everyone defaulted at the same time, but the defaults were correlated – they moved together to some degree. Li used the same math that statisticians used to model how people reacted when their spouses died to model how different loans reacted when one of them “died,” or defaulted. Li told the Wall Street Journal, “Suddenly I thought that the problem I was trying to solve was exactly like the problem these guys were trying to solve. Default is like the death of a company, so we should model this the same way we model human life.”

According to the math, huge amounts of risk disappeared when you pooled risky assets together in a CDO. The key assumption was that although some loans might default at the same time, not all of them would default simultaneously. For example, if you assumed the chances of two-thirds of the loans defaulting at the same time were close to zero, you could split the CDO into a risky piece (which would bear the first losses when loans in the pool defaulted) and a safer piece (which would not lose any money unless more than one-third of the loans defaulted).  Then, the safer piece would be rated AAA.

The CDO that Allan Sloan describes in Inside Job was based on exactly this assumption. The banks and rating agencies assumed that, although some of the mortgage loans in the pool might default at the same time, the likelihood of more than one-third defaulting together was basically zero. In other words, they assumed the correlation was low.

Historically, this correlation had been low, especially as housing prices rose. But what would happen if the nature of the loans changed (they were made to borrowers with bad credit who put virtually no money down), and then housing prices fell? Even a slight decline in housing prices would pull borrowers underwater, meaning the amount they had borrowed was more than the value of their houses. Then, the correlation would be high. Everyone would default.

The experts who put together subprime CDOs vastly underestimated the correlation of defaults. 
  1. Why might they have done this? 
  2. Was it an innocent mistake, which surprised the banks and rating agencies as much as it surprised most investors? Or was it an intentional ruse, which generated phantom profits and bonuses, even as it sowed the seeds of financial destruction?
  3. How, exactly, was it “mad”?

No comments:

Post a Comment