> > > > >
4. FHA Delinquency Status One Year After Closing Date
FHA Lending Patterns Nationally and in the Third District States (27 pages, 858 KB)
As the FHA portfolio has grown in market share in recent years, concerns over the FHA program’s performance and viability have increased as rising delinquencies have depleted the program’s reserve funds.15 Here we provide information on the performance of the 2006, 2007, 2008, and 2009 cohorts of the FHA portfolio using LPS data. The results are in Tables 4 through 9. Tables 4, 6, and 8 provide national results for purchase, refinance, and all loans. Tables 5, 7, and 9 provide corresponding results for the Third District states. For each loan, the performance measure is the loan status one year from the closing date.16
Table 4 provides the loan status for all FHA purchase loans one year after the closing date for the United States as a whole, for each year from 2006 to 2009. In the 2006 cohort, 87.4 percent of loans were current or had been paid off at the one-year point, while 2.8 percent were 90 days or more delinquent, and 1.1 percent were in foreclosure. Performance was substantially weaker at the one-year point for the 2007 cohort: 4.5 percent were 90 days or more delinquent, and 1.7 percent were in foreclosure. This cohort is among the worst performing in the history of the FHA program. 17 Performance of the 2008 purchase cohort is better than that of 2007 and similar to the performance of the 2006 cohort. (The percentage of loans that are at least 90 days delinquent at the one-year point is a bit lower in the 2006 cohort than in the 2008 cohort, while the percentage of loans that are current or paid off is a bit higher in the 2008 cohort than in the 2006 cohort at the one-year point.) When comparing 2008 and 2006 performance, one might expect better performance in 2008, since credit scores were higher in that year. In this regard, the similarity in cohort performance between 2006 and 2008 is somewhat surprising and suggests that credit history is not the only factor affecting performance.18
Table 4 shows a dramatic improvement in the one-year performance of FHA purchase loans originated in 2009 compared with earlier years. Given that average credit scores for borrowers in the FHA program are higher in 2009, then to the extent that credit history is an important predictor of performance, it is not surprising that performance for that year improved sharply. However, a recent analysis conducted by the FHA suggests that delinquencies in 2009 purchase loans and nonstreamlined refinances that have moved beyond the one-year point are increasing relatively rapidly.19
In the Third District states, as shown in Table 5, delinquency rates at the one-year point for the 2006, 2007, 2008, and 2009 FHA purchase cohorts follow the overall national pattern. In general, the delinquency rate for any given cohort tends to be higher than the national level in New Jersey and lower than the national level in Pennsylvania. (The 2007 to 2009 purchase cohorts in Delaware tend to perform less well than the corresponding national cohorts, but not to the same degree as in New Jersey.) Unemployment rates in New Jersey have been higher than in either Delaware or Pennsylvania in recent years and, with the onset of the housing crisis, have tracked the national average (Chart 4). Also, as shown in Table 3, FHA borrowers in New Jersey had lower mean credit scores than the other Third District states in 2006, 2007, and 2008. Performance converges for the Third District states in 2009.
Table 6 provides information on loan status at the national level one year after the closing date for FHA refinance loans in the 2006 through 2009 cohorts. The 2006 cohort performs considerably better than the 2007 and 2008 cohorts but not as well as the 2009 cohort. The 2009 cohort exhibits a sharp increase in performance over previous years for the nation as a whole, which is not surprising considering the earlier findings regarding mean borrower credit scores for the 2009 cohort.20
As in the nation as a whole, the performances of the 2007 and 2008 refinance cohorts at the one-year point were considerably weaker than those of the 2006 and 2009 cohorts in each of the Third District States (Table 7). For any given year, the New Jersey cohort did considerably worse on the performance indicators contained in Table 7 than the Pennsylvania cohort, while the Delaware cohort’s performance tended to fall in between that of New Jersey and Pennsylvania.
At the national level, the performance of refinance and purchase loans was about the same in 2006. However, from 2007 to 2009, the performance of refinance loans was substantially worse than that of purchase loans at the one-year point. In 2008, many borrowers with conventional loans refinanced through the FHA program. A significant share of these borrowers may have done so in order to seek relief from the terms of subprime loans made during the middle years of the decade; such borrowers may have been particularly likely to face problems in meeting future loan obligations. (A similar phenomenon is also likely in the 2007 cohort, when large-scale refinancing from conventional to FHA loans also occurred.) The FHA attributes part of the gap in performance between purchase and refinance loans in 2009 to poor performance of streamlined refinance loans in that cohort.