REQUIRED READING: There are many unanswered questions that arose from the subprime mortgage crisis and subsequent foreclosure surge – the most crucial one being, ‘How can we prevent such a historical downturn from happening again in the future?’
Too many industry observers have already decided that the best way to prevent a similar crisis in the future is through extreme risk management. Under this scenario, a lower FICO score equals higher default risk. The results of this approach, at least in the period following the 2008 crash, have been a general tightening of credit and a reduction in the number of home loans being made – which brings us to another problem as opposed to a panacea.
But are other strategies available that will ensure the correct balance between quality and quantity? For starters, we may need to think outside of that proverbial box. There are numbers out there suggesting that our industry may be overreliant on formulas and FICO scores, to its own detriment. We may, in fact, be shrinking our own market while excluding potential customers who are, indeed, good investments.
For those who are not familiar with the industry's history, here is a quick recap of how we got to where we are. The FICO scoring system, which was created by Bill Fair and Earl Isaac in 1958, is generally based upon five factors: payment history, indebtedness, extent of credit history, kinds of credit utilized, and acquisition of new credit. It bears mention here that such a formula is clearly more consistent, faster and less expensive to employ than a more "judgmental" system (i.e., one that uses less standardization, and requires far more personalized attention).
At the height of the mortgage boom, most lenders simply did not have the time or resources to employ a more "judgmental" process to each case without risking the loss of an impatient potential borrower. Accordingly, that process – which required more painstaking time and energy expenditures – was reserved for larger and more complex loans.
Using the FICO formula, therefore, it seemed reasonable to assume that the less reliable borrowers had lower FICO scores and, thus, would be more likely to fall into default. To a degree, this theory has proven true for decades. But as we sift through the debris left by the massive default spike of 2007 and 2008, it would appear that the FICO theory has been turned on its head.
According to Yuliya Demyanyk, senior economist with the Federal Reserve Bank of Cleveland, there is compelling evidence that FICO scores did not correspond to the increased default rates of 2006 and 2007. In her article, "Did Credit Scores Predict the Subprime Crisis?" – published in the October 2008 edition of The Regional Economist – Demyanyk wrote that the subprime market meltdown rewrote the laws of FICO logic.
"Throughout the boom and subsequent collapse of the subprime mortgage market, borrowers with low FICO credit scores were more likely to miss their mortgage payments or default on their loans," Demyanyk stated. "However, as the data show, borrowers who took out subprime loans in 2006 and 2007 had higher or similar FICO scores – not lower – than borrowers who took out their mortgages in earlier years."
The VA way
So here is our new puzzle: If we cannot rely on FICO scores to make satisfactory judgments, what strategy can we pursue? One possible approach might require taking a new look at one of the most efficient federal government housing offerings: the mortgage program administered by the U.S. Department of Veterans Affairs (VA).
In spite of a history of facilitating mortgages for borrowers with generally lower FICO scores, the VA program boasts one of the lowest default rates in the nation. Even during the default crisis, the VA program did not create the tumult experienced in the wider lending environment. The Mortgage Bankers Association reported in 2009 that VA loan borrowers had a 2.6% default rate, compared to 3.4% for prime borrowers. This holds true even though the program generally qualifies more borrowers with significantly lower FICO scores.
What did the VA program know that the rest of the industry missed? A deeper examination of how the VA mortgage is evaluated and underwritten demonstrates that the risk assessment process – and not simply the borrower – needs to be reconsidered.
In general, the approval process for VA mortgages usually judges potential borrowers with a much lower degree of harshness when than does the process for other types of home loans. According to the Federal Housing Finance Agency, the average credit score for a VA borrower in 2009 was about 700, whereas the average credit score for all borrowers was 750. This would seem to fly in the face of traditional thinking on the accuracy of FICO scores in predicting default.
So how does the VA achieve these positive results? The fundamental difference between a conventional FICO-based scoring model and the VA credit assessment boils down to a single factor: The VA loan considers the residual or disposable income of a potential borrower, whereas a FICO-score-driven approach focuses on payment history.
Yes, there is merit to both approaches. A borrower's disposable income today may not be what it will be tomorrow, and his or her credit history cannot and should not be ignored.
Nonetheless, the most relied-upon metric for measuring a borrower's ability to repay should be based almost completely on that person's history, with almost no regard for disposable income – which is perhaps the single most important factor in repayment. VA mortgages – while they do not ignore the past financial behavior of borrowers – do not omit this most important of determinations.
Of course, the VA mortgage – not unlike other mortgage programs administered by the federal government – was created to serve a very specific demographic. Its underwriting model was designed to ensure that fiscally responsible veterans were given every opportunity to secure home financing in repayment for their service.
As we sift through the residue of the housing bust of the last few years, questions about how we choose to underwrite our mortgage loans begin to surface. The VA mortgage program approach may not be the perfect solution for the wider mortgage banking industry, especially in view of the regulatory changes placed on the industry in the past couple of years. However, it does highlight the stark reality that many responsible borrowers do not fit neatly into a metric bucket – a fact that many underwriters, originators and regulators have conveniently forgotten.
Andrew Peters is CEO of First Guaranty Mortgage Corp., headquartered in McLean, Va. He can be reached at (703) 356-9200.