The traditional method of measuring mortgage denial rates is fundamentally flawed and can produce mistaken conclusions about access to credit. Evaluating not just the quantity but also the quality of mortgage applicants can give a clearer picture of actual denial rates.
A recent National Mortgage News column, "Narrative of the 'Fog-a-Mirror Finance' Era May Need Reexamining," uses the traditional measure of mortgage denial rates to conclude that "denials were an even larger percentage of the total [applications] a decade ago than they are now," and therefore, perceived wisdom of the loose underwriting standards before the housing crisis may need rethinking. This is exactly the wrong conclusion.
Instead, an analysis of denials that takes into account the mix of applicants shows the apparent spike in denials during the boom actually reflects a spike in lower-credit applicants. It's much harder to obtain a mortgage today than it was during the bubble years.
The traditional mortgage denial rate is calculated by dividing the number of denied mortgages by the total number of mortgage applications. For example, if 100 people apply for a mortgage and 30 are denied, the traditional mortgage denial rate is 30%. But this measurement reveals nothing about the mix of applicants, a critical variable in the denial decision.
Let's say 40 of the 100 applicants in this scenario are applying for non-risky products and have near-perfect credit characteristics (FICO scores greater than 700, loan-to-value ratios less than 78%, and debt-to-income ratios less than 30%). The traditional measure of denial rates has no way to account for the fact that those high-quality applicants are virtually guaranteed to be approved.
Indeed, as the article suggests, the traditional denial rate of owner-occupied, purchase mortgage applications did rise during the boom, peaked at 18% in 2006, and then dropped as the bottom fell out of the market in 2008. Since 2011, it has remained at the historic low of 14%. But this trend obscures the fact that the share of less-than-perfectly-qualified applicants rose during the boom years and fell during the crash.
To address this flaw in the traditional denial rate metric, the Urban Institute's Housing Finance Policy Center has developed a new measure of denial rates that considers only those applicants who stand some chance of being denied a mortgage.
When this methodology is applied to the above scenario of 100 loans where 40 applicants have almost no chance of being denied, and 30 of the remaining 60 applicants are denied the adjusted denial rate is 50%, instead of the 30% derived from the traditional measurement.
Note that within the group of less-than-perfect credit applicants, individuals may have different credit profiles. This is a variable that's not controlled for in the metric. Still, the adjusted rate produces a better measure of credit tightness because it reveals how willing lenders are to originate loans to applicants that pose some risk.
Considering only the denial rate of applicants with less than "near-perfect" credit consistently reflects mortgage market trends and reveals some key points.
Denial rates have been significantly underestimated. Applicants with less-than-perfect credit faced a 43% denial rate in 2013 more than three times higher than the 14% estimated by the traditional calculation.
Today's credit box is also much tighter than in the bubble years. Denial rates for applicants with less-than-perfect credit have risen substantially, from 25% in 2004 to 43% in 2013.
For applicants with less-than-perfect credit, the gap in denial rates between whites and minorities has narrowed. The traditional denial rate calculation shows that blacks and Hispanics are denied mortgages at significantly higher rates than whites. However, this calculation does not account for the differences in the share across racial groups of applicants with less-than-perfect credit. But taking these differences into account shows a narrower and declining gap between whites and minorities.
Lastly, it's always been easier to qualify for an FHA loan than a conventional loan, even after the housing crisis. Using the traditional denial measure, it looks like post-crisis consumers are more often denied FHA loans than conventional loans. In reality, this difference is all about the applicant pool. FHA lending is still the channel by which lower-credit profile applicants are most likely to get mortgages.
Ellen Seidman is senior fellow and Wei Li is senior research associate at the Urban Institute. Seidman previously served as director of the Office of Thrift Supervision and concurrently served on the board of directors of the Federal Deposit Insurance Corp.