«Barriers to Federal Home Mortgage Modification Efforts During the Financial Crisis Patricia A. McCoy August 2010 MF10-6 Paper originally presented at ...»
Another group, consisting of about 35% of PSAs, limits modifications to 5% of the loan pool (measured by the loan amount or number of loans). PSAs often contain one or more other restrictions on loan modifications. Examples include mandatory trial modification periods, use of specific resolution procedures, caps on interest rate reductions, restrictions on the types of eligible loans, and limits on the number of modifications in any one year.77 For the 90% or so of private-label securitizations that allow loan modifications to some degree, it is unclear whether limits on those modifications have become binding. The Boston Fed study expressed doubts on this score after finding minor differences at most in the rates of loan modifications for non-conforming loans held by securitized trusts versus those held in portfolio.78 Similarly, a Berkeley survey of PSAs concluded that “large-scale modification programs may be undertaken without violating the plain terms of PSAs in most cases.”79 Even for securitizations that prohibit loan modifications outright or cap them at 5%, some of those 72 Credit Suisse (2007), at 6; Hunt (2009), at 7. Among other things, this has the salutary benefit of allowing servicers to contact borrowers before any payments are missed to determine the borrower’s ability to handle the new payments and, if not, to explore other options.
73 Credit Suisse (2007), at 6; Hunt (2009), at 7.
74 See generally Eggert (2007).
75 Credit Suisse (2007); Hunt (2009), at 6.
76 Hunt (2009), at 7-9. In general, any change in the principal balance, the interest rate, or the final maturity will constitute a “material” modification. Hunt (2009), at 7.
77 Credit Suisse (2007), at 6-7, 20; Kiff and Klyuev (2009), at 11.
78 Adelino et al. (2009), at 13-18 & tbl. 5. See also Thompson (2009), at 6.
79 Hunt (2009), at 10.
22 PSAs have been amended to allow more modifications.80 In addition, credit rating agencies no longer count modified loans that are current 12 months after modification against the 5% cap where one exists.81 Taken together, these findings raise questions whether the limitations in PSAs explain the low level of permanent loan modifications, at least for the 90% of securitized trusts that allow modifications.
Calculating Net Present Value Before a servicer may modify a loan hold by a private-label securitized trust, it must first determine that modification will maximize the net present value of the loan, relative to foreclosure. PSAs normally require servicers to maximize the recovery for the benefit of the investors in the trust as a whole. Servicers implement this requirement by choosing the higher net present value (NPV), as between a loan workout and a foreclosure.
Although the NPV requirement sounds formulaic, PSAs give servicers of private-label RMBS a high degree of latitude in how to calculate net present value. This discretion allows the servicer to pick the likely sales price from a foreclosure, the discount rate to apply to the reduced revenue stream from a loan modification, and the likelihood that the borrower will redefault.
Investors rarely monitor these choices or question them.82 As a result, for many distressed loans, servicers can produce a NPV calculation to support either a loan modification or foreclosure.
There are three key factors in servicers’ NPV determinations. First, servicers are concerned about the risk that a modified loan may quickly redefault. Second, some distressed loans self-cure. Finally, servicers worry about moral hazard, i.e., the concern that loan modifications will encourage other borrowers who are current to default on their loans as well.
Redefault Rates The likelihood that a borrower will redefault on a loan following a workout will directly affect the net present value calculus. As redefault becomes more likely, servicers will prefer to initiate foreclosure immediately instead of delaying the inevitable. This has the enticing effect as well of accelerating the servicer’s final payout. Housing price movements will also affect the 80 Most PSAs allow caps on loan modifications to be waived upon consent by a rating agency or a bond insurer; only a few require investor approval. Kiff and Klyuev (2009), at 11.
81 Thompson (2009), at 6-7.
82 Cordell et al. (2008), at 18; Thompson (2009), at 6-9. See also Kiff and Klyuev (2009), at 8 n.10 (discussing market forces affecting choice of discount rates). For GSE loan pools, this discretion is more limited. Fannie Mae and Freddie Mac require servicers to use standardized software to calculate NPV. Cordell et al. (2008), at 18; Credit Suisse (2007), at 6-7, 20.
23 attractiveness of immediately going to foreclosure. In the recent climate of falling home prices, any delay in an inevitable foreclosure would depress the ultimate sales price at the sheriff’s sale.
All of these dynamics discourage loan workouts.83 Concerns about redefault are not hypothetical. Studies consistently show that redefault rates for many recent loan workouts have been relatively high. The Boston Fed reported, for example, that the redefault rate84 for loan workouts of all stripes was 40% to 50% during the first 6 months following workout.85 Similarly, OCC and OTS reported that 42.7% of loans modified in the first 3 months of 2009 were at least 60 days delinquent 6 months later.86 That said, the right workout technique can substantially lower the risk of redefault. As discussed, many loan workouts during the crisis involved capitalizing arrears, which had the effect of increasing borrowers’ monthly mortgage payments. For borrowers with cash-flow problems – especially problems that were indefinite or permanent in nature – such workouts were doomed to failure.
In contrast, evidence consistently shows that loan modifications that lower monthly payments, either by reducing interest, reducing principal, or extending the maturity date, have substantially lower redefault rates. The Boston Fed reported that redefault rates for those types of modifications dropped to 20% to 40%.87 The OCC and OTS reported similarly improved rates for modifications that decreased payments in 2008-2009; in contrast, the 6-month redefault rate for workouts increasing payments over that period was 53.6%. The agencies further found that the bigger the decrease in payments, the lower the redefault rate.88 Researchers at the University of North Carolina found that lowering monthly payments reduced the likelihood of redefault by 10% to 18%, compared to workouts that raised them. Principal write-downs had the lowest default rate of all, probably because they lowered monthly payments while reducing or eliminating negative equity.89 83 Adelino et al. (2009), at 21.
84 Defined as loans that were 60 days delinquent, 90 days delinquent, in foreclosure, or in real estate owned within 6 months after modification.
85 Adelino et al. (2009), at 19 & tbl. 8. See also Cordell et al. (2008), at 23.
86 OCC and OTS (2009), at 30. See also Fitch Ratings (2009b), at 9-10.
87 Adelino et al. (2009), at 19 & tbl. 8; Quercia and Ding (2009), at 175-177, 184-190. See also Mason (2009), at 35 (interest rate modifications lower 9-month redefault rates more than 10% compared to capitalizing arrears or principal modifications).
88 OCC and OTS (2009), at 36-37. See also Fitch Ratings (2009b), at 10-12.
89 Quercia and Ding (2009), at 188-190.
24 Cure Rates The cure rate also affects the net present value calculation. This is the rate at which seriously delinquent borrowers resume payments on their own. Cure rates can retard loan workouts by giving servicers hope that borrowers will repay with no further intervention, either by making up their arrears or paying off their loans in full. In many instances of cure, the servicer’s recovery is higher than it would be from lowering the monthly payment by modifying the loan.
Up through 2006, the cure rates on distressed mortgages were substantial. From 2000 through 2006, prime loans had an average cure rate of 45%; for subprime loans, the average cure rate was 19.4%. During the financial crisis, however, cure rates took a nosedive. By 2009, cure rates had plummeted to 6.6% for prime loans and 5.3% for subprime loans.90 The tight market for refinance loans and the rising tide of underwater mortgages helped explain this decline. The sharp fall in cure rates may be another reason why servicers have been more amenable to modifications and other types of workouts, at least of the temporary type.
Moral Hazard and Strategic Default Finally, servicers, lenders, and investors have understandable fears that loan modifications will induce other borrowers who have the wherewithal to pay their mortgages to strategically default (or threaten to default) to negotiate lower loan payments. Strategic defaults have become a special concern now that so many borrowers have underwater mortgages.
Fears of strategic default make servicers especially reluctant to reduce principal. To the extent servicers write down principal, they often insist on a short sale that requires the borrower to move out instead of a partial charge off that keeps the borrower in the home. This punitive effect is designed to discourage strategic defaults.91 Tranche Warfare In the debate about the low incidence of permanent modifications, commentators have also pointed to the role of potential fear of litigation from modifying loans. This colloquy stems from the fact that a servicer’s decision to choose loan modification over foreclosure affects different tranches differently. Modifications that return loans to performing status benefit the junior tranche 90 Fitch Ratings (2009a).
91 Thompson (2009), at 9.
25 by helping it avoid sustaining losses from foreclosure. At the same time, modifications that cut the monthly payments hurt the senior tranches by reducing their revenue stream. Instead, senior tranche holders may prefer foreclosure because the junior tranche will absorb the loss first while the senior tranche holders will often receive their principal back in full.92 These dynamics have fueled speculation that servicers avoid loan modifications to limit the risk of “tranche warfare,” i.e., lawsuits against them by tranche holders.93 There is scant evidence that tranche warfare is a real impediment to loan modifications. No investors have sued servicers to date for agreeing to loan modifications. In part that is because a sufficient number of investors must consent before suit can be filed.94 Meanwhile, servicers report that investors rarely question workouts, examine NPV calculations, or even threaten to bring a lawsuit.95 In 2008 and 2009, Congress alleviated any concern that tranche warfare might affect servicers’ psychology by enacting a safe harbor for servicers from investor suits for modifications that comport with standard industry practice or government modification programs.96 These statutory provisions did not raise the level of loan modifications in any obvious way, however, suggesting that tranche warfare is not the reason for low levels of loan modifications.
Tax Considerations To the extent that tax considerations might have played a role in loan modification determinations in the past, those considerations have been allayed. Virtually all securitized trusts are structured as pass-through entities under the Real Estate Mortgage Investment Conduit or REMIC provisions of the Internal Revenue Code to avoid double federal income tax liability. At one time, it was thought that the REMIC tax rules penalized loan modifications.97 This concern arose from the fact that REMICs must be limited to static loan pools to keep their tax-favored status. It was feared that loan modifications would destroy the static nature of the loan pool. On 92 Kiff and Klyuev (2009), at 11-12.
93 Eggert (2007).
94 Thompson (2008), at 8.
The PSAs for certain Countrywide securitizations contained a unique clause allowing Countrywide to modify up to 5% of the loan pool in dollar terms only if Countrywide bought back the modified loans from the loan pool. Hunt (2009), at 9-10. In Greenwich Financial Services Distressed Mortgage Fund 3, L.L.C. v. Countrywide Financial Corporation, No. 650474/2008 (N.Y. Sup. Ct., N.Y.
Cty., filed 2008), an investor in a securitized Countrywide trust sued Countrywide to force it to buy back any loans that it modified under a settlement agreement with state attorneys general. Despite surface appearances, the Greenwich suit is not an actual case of tranche warfare because Greenwich concurred that Countrywide had the latitude to modify the loans. Instead, Greenwich simply demanded that Countrywide buy back the loans it modified pursuant to its agreement. Thompson (2009), at 42 n.48.
95 Cordell et al. (2008), at 23.
96 HOPE for Homeowners Act of 2008, Pub. L. No. 110-289, div. A, tit. IV, § 1403 (July 30, 2008); Pub. L. No. 111-22, div. A, tit. II, § 201(b) (May 20, 2009) (codified at 15 U.S.C. § 1639a).
97 See, e.g., Thompson (2009), at 9-10. See generally Barr and Feldman (2008).
26 May 16, 2008, however, the Internal Revenue Service laid this concern to rest by ruling that loan modifications for owner-occupied homes will not endanger REMIC status so long as the loan is in default or the servicer reasonably believes there is a significant risk of default. In addition, any modification must follow a standard protocol.98 Due to this IRS ruling, tax considerations no longer pose a significant obstacle to modifications of distressed home loans.
Lessons Learned At this point, federal foreclosure prevention programs of one sort of another have been underway for nearly three years. In this section, I discuss the lessons to be learned.
The Importance Of Aggressive Early Intervention During the crisis, the federal government’s foreclosure prevention efforts got off to a painfully slow start. Only later did those efforts become more aggressive, particularly under the Obama Administration.
The federal government’s early lukewarm response was a serious mistake. Aggressive loss mitigation from the outset of the crisis might have slowed the spread of the nation’s housing woes, which instead triggered the deepest recession since the Great Depression. By stopping unnecessary foreclosures early on, home values would have stabilized more quickly, helping to halt the negative feedback loop and the widespread loss of jobs. In addition, large-scale loss mitigation would have been easier before so many people became unemployed.