Early performance of the 2006 vintage of subprime loans has been notably weaker than that of recent prior subprime vintages. The increase in early payment defaults (EPDs) has squeezed originators' liquidity. What was behind this rise in EPDs, and how will it affect originations going forward?
The rise in early payment defaults (EPDs) of subprime mortgages--particularly those originated in 2006--is something of a bellwether for the weakness that overtook the mortgage industry in the past year. For participants in the market, the question is how far-reaching the problem is and what is likely ahead. [??] Among the more notable--and perhaps not entirely unexpected--aftershocks of the rise in EPDs is the extra measure of earnings pressure, as well as the corresponding liquidity squeeze, for lenders whose business models entailed selling loans (to third parties) with the promise to buy those loans back in the event of an EPD.
To gauge the severity of the problem, it helps to understand the steep trajectory of serious delinquencies (60-plus days late, in foreclosure or real estate-owned [REO]) in the 2006 vintage relative to recent prior years' vintages.
If we look at securitized loans, we can see that by the fourth month after issuance (typically six months after origination)--which is generally a good proxy period for measurement of EPDs--the level of seriously delinquent loans in the 2006 vintage subprime securitizations is at 4 percent (see Figure 1). This is nearly double that for the year earlier (2 percent) and markedly higher than the 2003 and 2004 vintages (at 1.5 percent).
Losses for the subprime 2006 vintage totaled nearly 0.2 percent by the end of the first year. While not yet very high in absolute terms, loss levels are nonetheless more than double those of the 2002-2005 vintages (see Figure 2). With a relatively high percentage of loans currently in foreclosure proceedings, this trend of increasingly weaker performance may carry into the future.
By way of comparison, 60-days-plus delinquencies for alter-native-A loans in the 2006 vintage securitizations were more than 1 percent by month four, again roughly double those for the 2003-2005 vintages but far lower than delinquencies for the 2006 subprime securitized loans (see Figures 1 and 3). Early losses for alt-A deals, however, continue to be rare (see Figure 4).
ARMs vs. FRMs
The weakening of early performance in 2006 was more pronounced for adjustable-rate mortgage (ARM) products than for fixed-rate mortgage (FRM) products (see Figures 5 and 6). Average credit scores for ARMs have improved over the past several years, but many of the loans themselves incorporate a number of risky characteristics, including reduced or no documentation and high combined loan-to-value (CLTV) ratios (see Figure 7).
[FIGURE 1 OMITTED]
Looking at selected credit trends, we observe a substantial rise in loans originated under reduced-documentation or no-documentation programs as a percentage of total ARM loans for the 2003-2006 period. The percentage of first-lien ARM loans originated with a simultaneous second-lien loan rose as well. Weighted average credit scores for...