Model Hubris
If mortgage lenders were nervous about their falling loan standards, they could always take confidence in their models. Lenders believed deeply that, whatever their guts told them, their statistically based, computer-driven models could reliably determine whom they should lend to, what kinds of loans they could offer, and how much interest they should charge. The models were brilliant; using the available information on the borrower and the property, they would forecast the odds of the borrower paying regularly and on time. Fannie Mae and Freddie Mac had long been champions of these so-called automated underwriting (AU) models, and now other lenders became converts as well.
Lenders felt sure that their AU models could accurately judge a borrower's creditworthiness.
One essential input into every lender's model is a borrower's credit score.11 This is a number, typically ranging from 350 to 850, indicating the risk that a borrower will eventually face a credit problem. The lower the score, the more problem-prone the borrower is. A score below 620 generally marks a borrower as subprime. Scores are constructed from credit files information that lenders report regularly to credit bureaus. Scores reflect not only past payment history, but also how a person uses credit. For example, the more credit cards in your wallet and the closer you are to your credit limits, the greater the risk is of lending you more money and, therefore, the lower your score.
Another essential input into all mortgage-lender models is the value of the house being purchased. Lenders use old-fashioned appraisals, but they also rely on so-called automated valuation models (AVM) for this. An AVM is automated because it doesn't require an actual human being to look at the property; instead it uses statistics and computing power to estimate a home's value: recent sales of comparable homes, tax assessments, other characteristics of the property, and price trends in the immediate area. AVMs are much less costly than human appraisals and require no waiting for an appraiser to visit a home.
Lenders had little choice but to believe in their models during the housing boom. Loan applications were coming thick and fast; without models that could spit out answers instantaneously, the applications couldn't be processed quickly enough to meet the competition among lenders. Automated underwriting models, credit scores, and AVMproduced house values are, in most circumstances, effective and promising tools, but lenders were much too comfortable with the results and had come to totally rely on them.
Models are only as good as the information that goes into them, and by 2007, the information going into many models was increasingly suspect. Borrowers had figured out how to manipulate their own credit scores, not by improving their habits, but by making a few simple adjustments in how they managed their debts. For instance, some borrowers paid a fee to be added as an "authorized user" on the credit-card accounts of people with higher scores. The positive payment information from the good cardholder improved the additional user's credit score. Such transplanting of credit DNA wasn't widespread, but it shows how borrowers were able to game the system and affect lenders' models.
Our website is not responsible for the information contained by this article. Webworldarticles.com is a free articles resource thus practically any visitor can submit an article. However if you notice any copyrighted material, please contact us and we will remove the article(s) in discussion right away.
This article was sent to us by:
Gregorian Y. at
06102010
1. What is the issue of a debt recovery claim
All articles in this directory are property of their respective authors. Additionally, read our Privacy Policy
© 2010 WebWorldarticles.com - All Rights Reserved. Partners: Gunblade Saga