Why Lending Standards Did Not Fall

Last week I wrote about new research from St. Louis Federal Reserve economists which, contrary to conventional wisdom, found very little evidence that subprime lending standards loosened in recent years.

But Emre Deliveli points out that another recent IMF paper finds the exact opposite. Which one is closer to the truth? To investigate, let's look at what each paper means by "lending standards".

The St. Louis Fed paper, written by Geetesh Bhardwaj and Rajdeep Sengupta, looked at FICO scores of subprime borrowers and found that even as the volume of no-doc, or low-documentation loans rose -- implying that lenders were more willing to gamble on borrowers -- the FICO scores of these borrowers also rose. That suggests originators tried to compensate for riskier loans by demanding stronger credit histories.

The IMF paper, written by Giovanni Dell'Ariccia, Deniz Igan, and Luc Laeven, says that evidence for a decline in lending standards are evident in the correlation between higher subprime mortgage originations, higher loan-to-income ratios rose, and lower loan denial rates. In other words, subprime lenders became more willing to lend to people with smaller incomes relative to loan size, causing a decline in denial rates. Meanwhile, the researchers found that this dynamic was absent from the prime market.

Here's what The Economist said about the paper in May:

The authors blame what economists call "asymmetric information". Prime borrowers have long, publicly available credit histories; subprime borrowers do not. That might seem like a good reason to be warier of subprime borrowers, but it did not work that way. "The next person in the queue is seen as just another profitable opportunity, rather than someone who has been rejected by another lender," says Mr Dell'Ariccia.

Ironically, this explanation is actually why I think the St. Louis Fed paper looks to be the stronger of the two since that paper can explain the results we see in the IMF research. Someone who applies for a prime loan is likely to have a better credit history than the average subprime borrower. If home-buyers are getting denied for prime loans, it's logical that some of these buyers would try to get subprime loans instead, with the idea that once housing prices continued their ascent, the buyers could refinance and swap their subprime mortgages for prime loans. That could explain why subprime loan volume increased, denial rates declined, and loan-to-income ratios ballooned since a potentially prime borrower is probably buying a bigger house than a subprime borrower.

The St. Louis Fed paper is also "cleaner" in the sense that it uses far fewer statistical manipulations to get at its results. That by itself doesn't mean it's a better paper, but when you add more complexity to your calculations, there are more chances to make mistakes.

If you've gotten this far, you might be wondering why we even care about lower lending standards. The central issue is that many economists and central banks believe the cause of the subprime crisis is rooted in what's called the "originate-to-distribute" model of lending, where the entity that actually makes the loan to a borrower then turns around and sells the loan to various third parties. The problem with this model, many economists think, is that it reduces the incentives to keep risks in check since the originator can now "divorce" themselves from the consequences of making a bad loan.

The findings of the IMF paper fall in line with this criticism of the "originate-to-distribute" model while the St. Louis Fed paper's results suggest that this view may be too simplistic -- the first empirical evidence of a concern raised by some other economists.

In a lengthy paper presented at the Fed's Jackson Hole Symposium in August, Yale economist Gary Gorton showed that mortgage originators do not pass along all risks after they originate or sell a loan. Here is one example:

Originators of mortgages retain significant interests in the mortgages they originate due to servicing rights and retained interests. Mortgage servicing rights are valuable, and retained interest are also significant. When loans are sold in the secondary market, the mortgage servicing rights that are created are typically not sold.

In fact, the originate-to-distribute model has been around for decades, so what was so special about subprime mortgages that would've made it fail? Nothing, says Gorton. The problem with how subprime loans were constructed was that they were overly sensitive to home price changes:

What does explain the performance of...vintage subprime mortgages? House price...depreciation, is the biggest single factor explaining defaults...

The idea that there is a moral hazard due to the alleged ability of orginators to sell loans without fear of recourse...also assumes that the buyers of these loans are irrational. That may be, but the irrationality, it turns out, had to with the belief that house prices would not fall.

Comments

[ETFnerd:]

if you are arguing that lending standards did not fall during the housing bubble...

...put down your crack pipe and slowly walk away.

2008 Oct 21 12:58 PM

[AlexS:]

My sixth grade teacher, Daniel E. Rile used to say, "Figures never lie, but liars figure." I always keep that in mind when I read a report based on statistics.

2008 Oct 21 01:04 PM

[User 278265:]

This article looks like pro-Democrat propaganda.

2008 Oct 21 01:28 PM

[koz:]

I dont think he is arguing anything, just siting a couple of studies. Interesting concept. I doubt anyone will ever be able to tell for sure what caused this but its probably a lot of linked things. I just hope are government doesn't do the typical knee jerk reaction and do something stupid. I am now paying 30 dollars an airline ticket for crappy airport security. ugh... Oh and i remeber that sarbaines oxley thing, just another over regualtion example with no real benifit except to open up the door for europe to compete with us

2008 Oct 22 02:31 AM

[John Preston:]

In some cases, subprime lenders actually had higher average FICO scores than did Fannie and Freddie....and certainly higher than FHA, during the period from 2001-2007.

Investors and speculators found it easy to pay a rate premium when they were getting a yield premium...through rampant appreciation....

That said, the fact that lending standards expanded, and have now contracted, is at the core of the issue for lenders...this arbitrary, greed/fear driven cycle of accordion-like expand-contract-expand is manipulation...not market action...

2008 Oct 22 09:15 AM