Federal Reserve governor Jeremy Stein has a learned talk out today about the need to take an "institutional view" of credit market dynamics. There are a lot of stylized models out there in which credit intermediation institutions basically don't exist, and the volume of credit is simply determined by fluctuations in household and firm preferences. He argues that you have to think about the way the institutions actually function and the incentives facing actors within them.
Specifically, he notes "three factors that can contribute to overheating" in credit markets.
One is "financial innovation." The second is "changes in regulation ... new loopholes ... exploited by variants on already existing instruments." The third is "a change in the economic environment that alters the risk-taking incentives ... a prolonged period of low interest rates ... 'reach for yield.'"
And fair enough. But here's some evidence that's emerged from litigation between Dexia, a large Belgian bank that lost a ton of money buying mortgage-backed securities from JPMorgan and also from two banks that JPMorgan has since acquired:
According to the court documents, an analysis for JPMorgan in September 2006 found that “nearly half of the sample pool” — or 214 loans — were “defective,” meaning they did not meet the underwriting standards. The borrowers’ incomes, the firms found, were dangerously low relative to the size of their mortgages. Another troubling report in 2006 discovered that thousands of borrowers had already fallen behind on their payments.
But JPMorgan at times dismissed the critical assessments or altered them, the documents show. Certain JPMorgan employees, including the bankers who assembled the mortgages and the due diligence managers, had the power to ignore or veto bad reviews. [...]
At Bear Stearns and Washington Mutual, employees also had the power to sanitize bad assessments. Employees at Bear Stearns were told that they were responsible for “purging all of the older reports” that showed flaws, “leaving only the final reports,” according to the court documents. [...]
Ratings agencies also did not necessarily get a complete picture of the investments, according to the court filings. An assessment of the loans in one security revealed that 24 percent of the sample was “materially defective,” the filings show. After exercising override power, a JPMorgan employee sent a report in May 2006 to a ratings agency that showed only 5.3 percent of the mortgages were defective.
Which is just to say that I think Stein should take his own point about the granularity of institutions and incentives more seriously. Whether or not this malfeasance is something that JPMorgan is ultimately legally liable for is a legal question I can't speak to, but it's clear that in economic terms we were dealing with a marketplace in which major vendors were concealing materially relevant information. The ability to get away with that kind of thing is a factor that can easily lead to overheating. After all, the actual creditworthiness of potential borrowers is a major constraint on the expansion of credit. If intermediaries believe they can get away with misstating the creditworthiness of borrowers (either through direct fraud or by covering up fraud on the part of the borrowers) then there's going to be a lending boom.
None of this is to undermine the stuff Stein does say, but there continues to be a big blind spot in the economics world to this stuff, even though there's more and more lawyering around it every day.