Too Little, Too Late, and Too Timid: The Federal Response to the Foreclosure Crisis at the Five-Year Mark

Abstract

The primary federal policy responses to the foreclosure crisis, thus far, include programs to reduce foreclosures and efforts to mitigate the impacts of foreclosures on communities. This paper reviews policy responses between 2007 and 2012. While there is less information at this point on the outcomes of mitigation polices, the overall federal response is thus far lacking. The programs pale in comparison with the challenges they are intended to solve and suffer from other program design and implementation problems. Foreclosure prevention efforts, in particular, are faulted for being too reliant on marginal incentive payments, for failing to include a key policy, bankruptcy modification, which would have encouraged lenders to modify loans more aggressively, and for not sanctioning servicers more aggressively for poor performance and/or noncompliance. The overall federal response is also characterized as moving too slowly in some cases and being too captive to the policy preferences of the financial services industry.

Keywords:

• foreclosure
• subprime crisis
• mortgage crisis

Acknowledgements

I thank Frank Alexander, Kevin Byers, Phillip Comeau, Sarah Greenberg, Alan Mallach, Ira Rheingold, and Alan M. White for their comments on an earlier version of this paper. I would also like to thank Kathe Newman, Tom Sanchez, and the anonymous referees for their comments on a previous draft. Any errors in this paper—as well as all opinions—remain entirely my responsibility.

Notes

^1. Loan modifications received by NFMC clients receiving loan modifications resulted in substantially lower loan payments than would have been received without counseling. The Urban Institute researchers estimate that NFMC clients receiving loan modifications in the first two program years reduced their mortgage payments by $267 per month more than they would have without NFMC counseling. The sustainability of modifications was also greater for NFMC clients than for other borrowers.

^2. Twelve-month redefault rates would be expected to be substantially higher than these rates.

^3. By allowing homeowners to refinance into lower cost mortgages, HARP would also result in providing borrowers with more disposable income. Thus the program also had a goal of stimulating the economy more broadly. Arguably, this goal became more important as the economic crisis worsened.

^4. In the first step of the modification, the loan servicer rolls any unpaid interest and fees into the outstanding loan balance. Then, the servicer reduces the interest rate on the loan down to as low as 2% (with half of the loss in interest being absorbed by the federal government). If the 31% monthly debt-to-income ratio has still not been reached, the term of the loan can be extended to as long as 40 years. If the 31% ratio has still not been reached, the servicer, at its option (and sometimes requiring the permission of the lender/investor) can defer the due date of some of the principal on the loan (called “forbearance”).

^5. White (2011) collected data from private-label mortgage backed securities that showed loss severity rates ranging from 40% to more than 80% with a median of approximately 64%. Fannie Mae's quarterly reporting showed REO sale prices as a share of unpaid principal balances averaged approximately 56% in 2009 and 2010 (Fannie Mae, 2011). Because unpaid principal balance does not include interest and fees, loss severity ratios will be greater than 44% on such loans, and most likely approach something on the order of 50%.

^6. A short sale occurs when a borrower sells her house for less than the amount owned on the outstanding mortgage(s) and the lender(s) agrees to release the mortgage on the property for this amount. Deeds-in-lieu of foreclosure occur when borrowers sign over their property to the lender in exchange for the mortgage being released.

^7. Table 1 also does not include non-HAMP GSE activity because this activity was not designed or initiated by the Obama administration but by the GSEs and/or their independent federal regulator, whose director predated the Obama administration. According to the Financial Crisis Inquiry Commission (2011), lenders report having independently approved 3.4 million “proprietary” loan modifications of various kinds. However, the nature of these modifications is not entirely clear, and many have involved actual increases in monthly payments.

^8. HARP assisted over 810,000 borrowers as of May 2011 in refinancing loans with loan-to-value ratios between 81% and 125%. However, approximately 93% of these loans had loan-to-value ratios between 80% and 105%, the original HARP guidelines. This means that fewer than 60,000 HARP borrowers were significantly underwater.

^9. HARP activity is not included in Figure 3 because most HARP borrowers are unlikely to have been serious risks for foreclosure since the GSEs did not relax other lending guidelines (including credit history) significantly as a part of the program. The focus here is on the non-HARP activity.

^10. The estimated number of foreclosure starts was derived as follows. First, the number of foreclosure starts was taken from the Mortgage Bankers Association National Delinquency Survey. These totals were grossed up by 1.156 because the Mortgage Bankers Association estimates that its survey of captures only 85–88% of the mortgage market (Mortgage Bankers Association, 2011). Then, because mortgages on non-owner occupied properties are generally not eligible for MHA programs, the estimate must be adjusted for owner occupancy. Bocian et al. (2010) estimate that 82% of foreclosures from 2007 to 2009 were on owner-occupied homes. This factor is used to adjust the estimate of foreclosure starts downward to provide an estimate of foreclosure starts on owner-occupied homes.

^11. Changes to the PRA program in 2012 made sizable increases to the incentives for principal reductions, essentially tripling them. This increase may affect the future participation of the GSEs in the PRA program, although as of May 2012, the FHFA continues its policy of not participating in the program.

^12. Requirements for starting trials changed significantly in June 2010 because full verification of income was not required uniformly before this date. As a result, conversions to permanent modifications increased significantly after this date (U.S. Department of Treasury, 2011a).

^13. There is considerable debate over how pooling and servicing agreements and the structures of private label securitization structures impede substantial loan modifications. Thompson (2009) is among those who argue that securitization is not a significant impediment to modification, while Levitin and Twomey (2010) argue that securitization structures can effectively impede modifications. Moreover, Agarwal et al. (2010) find that securitized loans are significantly and substantially less likely to be modified than loans held in portfolio.

^14. The 18-month deadline to obligate NSP1 funds did not mean that all funds had to be expended within that timeframe. According to HUD regulations, “obligation” means “the amounts of orders placed, contracts awarded, goods and services received, and similar transactions during a given period that will require payment by the grantee (or subrecipient) during the same or a future period” (U.S. Department of Housing and Urban Development, 2010b). Obligations must be for specific NSP activities that can be linked to a specific address or household. Thus, a subcontract to a third party that would later purchase specific properties would not be considered an obligation.

^15. Grantees' 18-month periods did not all end on October 1, 2010. The deadline depended on when their grant agreements with HUD were executed. However, most periods ended around October 2011.

^16. One source of national estimates of foreclosed properties is the financial blog Calculated Risk (2011), which relies on estimates from the housing economist Tom Lawler and have been generally consistent with some other estimates. Calculated Risk's estimates include properties owned by the GSEs, HUD/FHA, private label securities, and FDIC insured banks. Properties owned by credit unions, non-FDIC-insured lenders, and some smaller federal agencies are not included. Thus, Calculated Risk estimates that the data represent “at least 90%” of all foreclosed properties. To account for the incomplete coverage, I multiply the Calculated Risk totals by 1.11 to come up with the range.

^17. As NSP1 was being implemented in the spring of 2010, the definitions of “foreclosed” and “abandoned” properties were broadened to make the funds more flexible and responsive. For example, the definition of foreclosed properties was broadened to include properties for which a mortgage is more than 60 days delinquent as well as tax delinquent properties (U.S. Department of Housing and Urban Development, 2010c).