Are Short Sales Really That Bad?
My colleague Joanne Gaskin wrote a great post about the impact to the FICO® Score from short sales and other mortgage stress-related events. One of the questions we get asked most often is whether it remains appropriate for the scoring model to treat a short sale in a manner similar to a foreclosure. Critics assert that since short sales do not cost the bank as much money as foreclosures, the penalty to a credit score should be less, commensurate with the financial impact on a lender. Some also suggest that the borrower’s willingness to work with the lender should have a positive effect on his/her credit risk.
Another argument for revisiting the scoring model’s treatment of these mortgage stress events is that the mortgage crisis was unprecedented. Consumers, who would have otherwise paid responsibly, were now making decisions that theoretically were not representative of their true risk.
To test these arguments, we conducted a study isolating more recent occurrences of mortgage stress events. By studying the subsequent performance of these borrowers on all accounts, we determined the credit risk associated with their mortgage events. Looking at data from October 2009 to October 2011, we were able to verify that short sales and other events of recent mortgage distress continue to represent a high degree of risk. These results closely match earlier studies of the risk associated with short sales and other events of mortgage stress.
As the graph below shows, short sales remain extremely risky. However, foreclosures have a bad rate of 72.0% while short sales have a better bad rate of 55.1%. Should that lead to less punitive treatment for short sales?
While it is true that short sales represent slightly better risk than foreclosures, they do not perform well enough to merit a more positive treatment in the FICO® Score. Here’s why. In the population we studied, one out of every two borrowers who experienced a short sale went on to default on another account within two years. That is exceptionally high risk. Additionally, the overwhelming majority of consumers with short sales have some other evidence of mortgage delinquency.
From a weighting perspective, all these mortgage events – short sale, foreclosure, deed in lieu – fall into the same heavyweight class, because they correlate with exceptional riskiness. They aren’t alone in that class either. Based on the data, consumers with short sales perform no better than consumers who have a severe delinquency (90+ days past due), a collection, or a derogatory public record (e.g., bankruptcy, tax lien, etc.) on file.
By comparison, only about one in every 50 borrowers with a score in the high 700s will default on one of their credit obligations. This strong separation of goods from bads is what makes FICO® Scores so useful.
The analysis above looks at the performance on any trade line, including mortgage trade lines. But what if we examined the consumers’ subsequent payment behavior on bankcards alone, a credit obligation far removed from their mortgage obligations?
Figure 2 indicates that there remains a strong link between poor bankcard payment behavior and consumers who experience a mortgage stress event. In fact, for all of the events in question, the bad rates observed are at least twice as risky when compared to the total population.
We’re also often asked about the risk impact of loan modifications. Since the credit reporting codes representing loan modifications were just introduced in 2010, at the time the research was conducted, not enough time had passed to provide a 24-month performance window for analysis, which is the length of the performance window used for FICO® Score redevelopments.
We will continue to monitor risk associated with these types of events and will update the FICO® credit scoring models appropriately. For instance, if we see empirical evidence demonstrating that consumers with a reported loan modification perform substantially worse in repaying their credit obligations, future models will be modified to appropriately classify those codes. As I’ve said on this blog before, we regularly conduct this type of research to identify changing risk patterns, so we can continuously produce the most consistently predictive scores possible.