View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

United States Government Accountability Office

GAO

Report to Congressional Committees

March 2011

TROUBLED ASSET
RELIEF PROGRAM
Treasury Continues to
Face Implementation
Challenges and Data
Weaknesses in Its
Making Home
Affordable Program

GAO-11-288

March 2011

TROUBLED ASSET RELIEF PROGRAM
Accountability • Integrity • Reliability

Treasury Continues to Face Implementation
Challenges and Data Weaknesses in Its Making Home
Affordable Program
Highlights of GAO-11-288, a report to
congressional committees

Why GAO Did This Study

What GAO Found

Two years after the Department of
the Treasury (Treasury) first made
available up to $50 billion for the
Making Home Affordable (MHA)
program, foreclosure rates remain at
historically high levels. Treasury
recently introduced several new
programs intended to further help
homeowners. This report examines
(1) the status of three of these new
programs, (2) characteristics of
homeowners with first-lien
modifications from the Home
Affordable Modification Program
(HAMP), and (3) the outcomes for
borrowers who were denied or fell
out of first-lien modifications. To
address these questions, GAO
analyzed data from Treasury and six
large MHA servicers.

The implementation of Treasury’s programs to reduce or eliminate secondlien mortgages, encourage the use of short sales or deeds-in-lieu, and
stimulate the forgiveness of principal has been slow and limited activity has
been reported to date (see table). This slow pace is attributed in part to
several implementation challenges. For example, servicers told GAO that the
start of the second-lien modification program had been slow due to problems
with the database Treasury required them to use to identify potentially eligible
loans. Additionally, borrowers may not be aware of their potential eligibility
for the program. While Treasury recently revised its guidelines to allow
servicers to bypass the database for certain loans, servicers could do more to
alert HAMP first-lien modification borrowers about the new second-lien
program. Implementation of the foreclosure alternatives program has also
been slow due to program restrictions, such as the requirement that
borrowers be evaluated for a first-lien modification even if they have already
identified a potential buyer for a short sale. Although Treasury has recently
taken action to address some of these concerns, the potential effects of its
changes remain unclear.

What GAO Recommends
GAO recommends that Treasury
require servicers to advise borrowers
to contact servicers about secondlien modifications and ensure that
servicers demonstrate the capacity to
successfully implement Treasury’s
new programs. GAO also
recommends that Treasury consider
methods to better capture outcomes
for borrowers denied or canceled
from HAMP first-lien modifications.
Treasury acknowledged challenges
faced by servicers in implementing
the program, but felt that certain
criticisms of MHA were unwarranted.
However, we continue to believe that
further action is needed to better
ensure the effectiveness of these
programs.

In addition, Treasury has not fully incorporated into its new programs key
lessons from its first-lien modification program. For example, it has not
obtained all required documentation to demonstrate that servicers have the
capacity to successfully implement the newer programs. As a result, servicers’
ability to effectively offer troubled homeowners second-lien modifications,
foreclosure alternatives, and principal reductions is unclear. Finally, Treasury
has not implemented GAO’s June 2010 recommendation that it establish goals
and effective performance measures for these programs. Without
performance measures and goals, Treasury will not be able to effectively
assess the outcomes of these programs.
Activity Under the Second-lien, Foreclosure Alternative, and Principal Reduction Programs as
of December 31, 2010
Date
announced
March 2009

Implementation
date
March 2010

Funding
allocation
Nearly $133
million

Reported activity as
of December 31, 2010
$2.9 million in
incentives paid

Home Affordable
Foreclosure
Alternatives

March 2009

April 5, 2010

$4.1 billion

$9.5 million in
incentives paid

Principal Reduction
Alternative

March 2010

October 1, 2010

$2.0 billion

Activity not yet
a
reported

Program
Second-lien
Modification

Source: Treasury.
a

PRA incentives are paid on an annual basis contingent upon successful performance of the modified
mortgage during the preceding 12 months.

View GAO-11-288 or key components. For
more information, contact Mathew J. Scirè at
(202) 512-8678 or sciremj@gao.gov.
United States Government Accountability Office

Highlights of GAO-11-288 (continued)

Treasury’s data provide important insights into the
characteristics of borrowers participating in the HAMP
first-lien modification program, but data were
sometimes missing or questionable. As shown in the
figure, more homeowners have been denied or canceled
from HAMP trial loan modifications than have received
permanent modifications. To understand which
borrowers HAMP has been able to help, GAO looked at
Treasury’s data on borrowers in HAMP trial and
permanent modifications. These data showed that
HAMP borrowers had reduced income and high debt,
but the reliability and integrity of some of Treasury’s
information was questionable. For example, Treasury’s
data on borrowers’ loan-to-value ratios at the time of
modification ranged from 0 to 999, with 1 percent of
TARP-funded active permanent modifications reporting
ratios over 400 percent. In addition, race and ethnicity
data were not available for a significant portion of
borrowers. Treasury said that it was refining and
strengthening data quality checks and that the data have
improved and will continue to improve over time.
Treasury’s success in improving the quality and
completeness of HAMP data will be critical to its ability
to evaluate program results and achieve the goals of
preserving homeownership and protecting home values.

While it appears that most borrowers who were
denied or canceled from HAMP first-lien trial
modifications have been able to avoid
foreclosure to date, weaknesses in how Treasury
requires servicers to report data make it difficult
to understand what ultimately happens to these
borrowers. First, Treasury’s system for reporting
outcomes requires servicers to place borrowers
in only one category, even when borrowers are
being evaluated for several possible outcomes,
with proprietary modifications reported first. As
a result, the proportion of borrowers reported
receiving proprietary modifications is likely to
be overstated relative to other possible
outcomes, such as foreclosure starts. Further,
Treasury does not require servicers to
distinguish between completed and pending
actions, so that some reported outcomes may
not be clear. Without more accurate information
on the outcomes of borrowers who are denied
HAMP modifications, have them canceled, or
redefault, Treasury’s ability to determine
whether further action is needed to assist
struggling homeowners is diminished.

Number of Active and Canceled Trial and Permanent Modifications through January 2011

Trial modifications canceled
68,114 Permanent modifications canceled
Active trial modifications

145,260

740,240
539,493

Active permanent modifications

Source: Treasury.

United States Government Accountability Office

Contents

Letter

1
Background
Implementation of Treasury’s Newer Housing Programs Has Been
Slow and Capacity of Servicers to Carry Out These Programs
Remains Unclear, Raising Uncertainty About the Potential
Impact of These Programs
Treasury Has Some Data on the Characteristics of Borrowers in
HAMP’s First-Lien Program, but Data Were Sometimes Missing
or Questionable
Most Borrowers Denied or Canceled from Trial Modifications
Appear to Have Avoided Foreclosure To Date, but Weaknesses
in Treasury’s Data Collection Limit its Ability to Understand the
Outcomes of These Borrowers
Conclusions
Recommendations for Executive Action
Agency Comments and Our Evaluation

4

32
47
50
51

Appendix I

Scope and Methodology

56

Appendix II

Description of GAO’s Econometric Analysis of HAMP
Trial Loans Modifications Cancellations
59

Appendix III

Comments from the Department of the Treasury

68

Appendix IV

GAO Contacts and Staff Acknowledgments

77

11

25

Tables
Table 1: Selected Outcomes of Borrowers who Had a Canceled
HAMP Trial Modification by Servicer, through August 31,
2010
Table 2: Terms of Selected Proprietary Modification Programs
Compared to HAMP
Table 3: Summary Statistics of Variables Used in Regression

Page i

39
44
60

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Table 4: Probabilistic Estimates of HAMP Trial Loan Modification
Cancellation Rates

62

Figures
Figure 1: National Default and Foreclosure Trends from Calendar
Years 1979-2010
Figure 2: GSE and Non-GSE HAMP Trial and Permanent
Modifications Made and Canceled Each Month, through
January 2011
Figure 3: Timeline of 2MP, HAFA, and PRA Guidance
Figure 4: Estimated Decrease/Increase in Likelihood of
Cancellation of HAMP Trial Modification by Borrower and
Loan Characteristics
Figure 5: Outcomes of Borrowers Denied a HAMP Trial
Modification, through August 31, 2010 (Six large MHA
servicers)
Figure 6: Outcomes of Borrowers who Had a Canceled HAMP Trial
Modification, through August 31, 2010 (Six large MHA
servicers)
Figure 7: Outcomes of Borrowers Who Redefaulted on a HAMP
Permanent Modification, through August 31, 2010 (Six
large MHA servicers)
Figure 8: Number of Proprietary and HAMP Modifications Started
Each Month, January through December 2010
Figure 9: Estimated Change in Likelihood of Cancellation of HAMP
Trial Loan Modification by Servicer, Delinquency Status
Before Modification
Figure 10: Estimated Change in Likelihood of Cancellation of
HAMP Trial Loan Modification by State

Page ii

5

10
21

28

37

38

40
41

65
66

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Abbreviations
2MP
DTI
FDIC
FHA
GSE
HAFA
HAMP
HPO
HUD
IR/2
LTV
LPS
MHA
MHA-C
MLTV
NPV
OCC
OFS
OTS
PRA
SIGTARP
SPA
TARP

Second-Lien Modification Program
debt-to-income ratio
Federal Deposit Insurance Corporation
Federal Housing Administration
government-sponsored enterprise
Home Affordable Foreclosure Alternatives Program
Home Affordable Modification Program
Homeownership Preservation Office
Department of Housing and Urban Development
Investor Reporting/2
loan-to-value
Lender Processing Services
Making Home Affordable
Making Home Affordable-Compliance
mark-to-market loan-to-value
net present value
Office of the Comptroller of the Currency
Office of Financial Stability
Office of Thrift Supervision
Principal Reduction Alternative
Office of the Special Inspector General for TARP
Servicer Participation Agreement
Troubled Asset Relief Program

This is a work of the U.S. government and is not subject to copyright protection in the
United States. The published product may be reproduced and distributed in its entirety
without further permission from GAO. However, because this work may contain
copyrighted images or other material, permission from the copyright holder may be
necessary if you wish to reproduce this material separately.

Page iii

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

United States Government Accountability Office
Washington, DC 20548

March 17, 2011
Congressional Committees
Since the Department of the Treasury (Treasury) first announced the
framework for its Making Home Affordable (MHA) program over 2 years
ago, the number of homeowners facing potential foreclosure has remained
at historically high levels. The Emergency Economic Stabilization Act of
2008, which authorized Treasury to establish the $700 billion Troubled
Asset Relief Program (TARP), was intended to, among other things,
preserve homeownership and protect home values. 1 In February 2009,
Treasury announced that up to $50 billion in TARP funds was allocated to
help struggling homeowners avoid potential foreclosure. The key
component under MHA, the Home Affordable Modification Program
(HAMP), offered modifications on first-lien mortgages to reduce
borrowers’ monthly mortgage payments to affordable levels, avoid
foreclosure, and keep their homes. Since HAMP’s inception, concerns
have been raised that the program was not reaching the expected number
of homeowners. In two prior reports, we looked at the implementation of
the HAMP first-lien modification program and noted that Treasury faced
challenges in implementing the program and made several
recommendations intended to address these challenges. 2 In addition, the

1

Pub. L. No. 110-343, 122 Stat. 3765 (2008), codified at 12 U.S.C. §§ 5201 et seq. The Helping
Families Save Their Homes Act of 2009, Pub. L. No. 111-22, Div. A, 123 Stat. 1632 (2009),
amended the act to reduce the maximum allowable amount of outstanding troubled assets
under the act by almost $1.3 billion, from $700 billion to $698.741 billion. The Dodd-Frank
Wall Street Reform and Consumer Protection Act, Pub. L. No. 111-203, 124 Stat. 1376 (2010)
(1) reduced Treasury’s authority to purchase or insure troubled assets to a maximum of
$475 billion and (2) prohibited Treasury, under the Emergency Economic Stabilization Act
of 2008, from incurring any additional obligations for a program or initiative, unless the
program or initiative had already been introduced prior to June 25, 2010.

2

GAO is required to report at least every 60 days on findings resulting from, among other
things, oversight of TARP’s performance in meeting the purposes of the act, the financial
condition and internal controls of TARP, the characteristics of both asset purchases and
the disposition of assets acquired, the efficiency of TARP’s operations in using
appropriated funds, and TARP’s compliance with applicable laws and regulations 12 U.S.C.
§ 5226(a). Under this statutory mandate, we have reported on Treasury’s use of TARP
funds to preserve homeownership and protect home values. See GAO, Troubled Asset
Relief Program: Treasury Actions Needed to Make the Home Affordable Modification
Program More Transparent and Accountable, GAO-09-837 (Washington, D.C: July 23,
2009) and GAO, Troubled Asset Relief Program: Further Actions Needed to Fully and
Equitably Implement Foreclosure Mitigation Programs, GAO-10-634 (Washington, D.C:
June 24, 2010).

Page 1

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Special Inspector General for TARP (SIGTARP) and the Congressional
Oversight Panel have issued several reports containing various
recommendations to Treasury intended to improve the transparency,
accountability, and effectiveness of MHA. 3
Questions continue to be raised about the extent to which the first-lien
program has effectively reached struggling homeowners and reduced
avoidable foreclosures. For example, more homeowners have been denied
or canceled from HAMP first-lien trial loan modifications than have
received permanent modifications to date, raising questions about which
homeowners HAMP has been able to help and how best to meet the needs
of homeowners struggling to avoid foreclosure. Treasury has begun
implementing several other TARP-funded programs for struggling
homeowners under the MHA program, including the Second-Lien
Modification Program (2MP), the Principal Reduction Alternatives (PRA)
program for borrowers who owe more on their mortgages than the value
of their homes, and the Home Affordable Foreclosure Alternatives (HAFA)
program for those who are not successful in HAMP modifications. 4 All are
funded by the $50 billion originally allocated for MHA, which has since
been reduced to $45.6 billion for all TARP-funded housing programs, and
further reduced to $29.9 billion for MHA programs (with the remainder of
the balance being allocated to the HFA Hardest-Hit Fund and the FHA
Short Refinance option). Because of concerns about the effectiveness of
these newer TARP-funded programs, this report examines the extent to
which these programs have been successful at reaching struggling
homeowners. To understand the extent to which Treasury has been able
to assess who has been reached by HAMP and what additional actions may
be needed to help struggling homeowners, we also examined the
characteristics of homeowners who have been assisted by the HAMP firstlien modification program and the outcomes of borrowers who did not
complete HAMP trial or permanent modifications. We also have ongoing

For example, see Congressional Oversight Panel, April Oversight Report: Evaluating
Progress on TARP Foreclosure Mitigation Programs (Washington, D.C., Apr. 14, 2010)
and Office of the Special Inspector General for the Troubled Asset Relief Program, Factors
Affecting Implementation of the Home Affordable Modification Program, SIGTARP-103

005 (Washington, D.C., Mar. 25, 2010).
4

Treasury has also put in place the Federal Housing Administration (FHA)-HAMP, Rural
Development-HAMP, the FHA Short Refinance Option, the Housing Finance Agency
Innovation Fund for the Hardest-Hit Markets, and the Home Affordable Unemployment
Program. Information on the progress made by these TARP-funded programs in stemming
avoidable foreclosures will be discussed in a future report.

Page 2

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

work looking at the broader federal response to the foreclosure crisis,
which encompasses both TARP and non-TARP funded efforts intended to
mitigate the impact of foreclosures on homeowners.
More specifically, this report examines (1) the status of Treasury’s secondlien modification, principal reduction, and foreclosure alternatives
programs; (2) the characteristics of homeowners who HAMP has been
able to help under the first-lien modification program; and (3) the
outcomes for borrowers who were denied or fell out of HAMP trial or
permanent first-lien modifications.
To address these questions, we obtained information from and spoke with
six large MHA servicers who collectively represented about 74 percent of
the TARP funds allocated to servicers participating in the program. In
addition, we reviewed MHA program documentation that Treasury issued,
including supplemental directives for the second-lien modification,
principal reduction, and foreclosure alternatives programs. In addition, we
spoke with members of a trade association who represented both
residential mortgage loan investors and servicers, and one who represents
private mortgage loan insurers. We also analyzed loan level data from
Treasury’s HAMP database, which included data reported by servicers on
borrowers evaluated for HAMP participation through September 30, 2010,
to analyze the characteristics of borrowers who received HAMP, were
canceled from HAMP trial modifications, or redefaulted from permanent
HAMP modifications. To understand the outcomes of borrowers who were
denied or canceled from HAMP, we requested and obtained data from
each of the six servicers noted above. Finally, we conducted a Web-based
survey of housing counselors through NeighborWorks, which funds a
national network of housing counselors to obtain their perspectives of the
HAMP program. 5 We coordinated our work with other oversight entities
that TARP created—the Congressional Oversight Panel, the Office of the
Special Inspector General for TARP, and the Financial Oversight Stability
Board.
We conducted this performance audit from July 2010 through March 2011
in accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain

5
This report does not contain all the results from our survey of housing counselors. The
survey and a more complete tabulation of the results will be discussed in more detail in an
upcoming report.

Page 3

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

sufficient, appropriate evidence to provide a reasonable basis for our
findings and conclusions based on our audit objectives. We believe that
the evidence obtained provides a reasonable basis for our findings and
conclusions based on the audit objectives. For additional information on
our scope and methodology, see appendix I.

Background

Although default rates (loans 90 days or more past due) fell from an alltime high of 5.09 percent at the end of the fourth quarter of 2009 to 3.94
percent at the end of the fourth quarter of 2010 (a nearly 23 percent drop
over the course of a year), the percentage of loans in foreclosure rose to
equal the highest level in recent history at 4.63 percent (fig.1). 6 The
increase in foreclosure inventory during the latter part of 2010 may be due
to issues surrounding foreclosure processing and procedures that resulted
in various foreclosure moratorium initiatives. In addition, the percentage
of loans that newly entered the foreclosure process in the fourth quarter of
2010 remained high at 1.27 percent, compared to 0.42 percent in the first
quarter of 2005.

6
The primary source of information on the status of mortgage loans was the Mortgage
Bankers Association’s (MBA) quarterly National Delinquency Survey, which was estimated
to represent about 88 percent of the mortgage market in the fourth quarter of 2010.

Page 4

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Figure 1: National Default and Foreclosure Trends from Calendar Years 1979-2010
Q2 2005–Q4 2010

Percentage

Percentage

6

6

5

5

4

4

3

3

2

2

1

1

0

0

HAMP
program

Default

Foreclosure starts

10
20

09
20

08
20

07
20

06
20

05

Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4

20

19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
20
09
20
10

Q1 1979–Q4 2010

Foreclosure inventory

Source: GAO analysis of MBA data

As we reported in December 2008, Treasury has established an Office of
Homeownership Preservation within the Office of Financial Stability
(OFS), which administers TARP, to address the issues of preserving
homeownership and protecting home values. 7 On February 18, 2009,
Treasury announced the broad outline of the MHA program. The largest
component of MHA was the HAMP first-lien modification program, which
was intended to help eligible homeowners stay in their homes and avoid
potential foreclosure. Treasury intended that up to $75 billion would be
committed to MHA ($50 billion under TARP and $25 billion from Fannie
Mae and Freddie Mac) to prevent avoidable foreclosures for up to 3 to 4
million borrowers who were struggling to pay their mortgages. According
to Treasury officials, up to $50 billion in TARP funds were to be used to
encourage the modification of mortgages that financial institutions owned
and held in their portfolios (whole loans) and mortgages held in private-

7

GAO, Troubled Asset Relief Program: Additional Actions Needed to Better Ensure
Integrity, Accountability, and Transparency, GAO-09-161 (Washington, D.C.: Dec. 2,
2008).

Page 5

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

label securitization trusts. 8 Fannie Mae and Freddie Mac together were
expected to provide up to an additional $25 billion from their own balance
sheets to encourage servicers and borrowers to modify or refinance loans
that those two Government Sponsored Enterprises (GSE) guaranteed. 9
Only financial institutions that voluntarily signed a Commitment to
Purchase Financial Instrument and Servicer Participation Agreement
(SPA) with respect to their non-GSE loans are eligible to receive TARP
financial incentives under the MHA program.
HAMP first-lien modifications are available to qualified borrowers who
occupied their properties as their primary residence, who had taken out
their loans on or before January 1, 2009, and whose first-lien mortgage
payment was more than 31 percent of their gross monthly income
(calculated using the front-end debt-to-income ratio (DTI)). 10 Only singlefamily properties (one-four units) with mortgages no greater than $729,750
for a one-unit property were eligible. 11
The HAMP first-lien modification program has four main features:
1. Cost sharing. Mortgage holders/investors are required to take the first
loss in reducing the borrower’s monthly payments to no more than 38
percent of the borrower’s income. For non-GSE loans, Treasury then

8

Loans held in private-label securitization trusts include loans not securitized by Fannie
Mae or Freddie Mac, and not insured or guaranteed by the Department of Housing and
Urban Development’s (HUD) FHA, the Department of Veterans Affairs (VA), or the U.S.
Department of Agriculture’s (USDA) Rural Housing Loan Program. Loans guaranteed by
HUD’s Federal Housing Administration (FHA) and the USDA Rural Housing Service are
eligible for TARP incentives when modified under requirements issued by those agencies.
The $50 billion was intended to be used for loan modifications and other foreclosure
prevention activities.
9

Any funds that Treasury provides to the GSEs Fannie Mae and Freddie Mac under the
preferred stock purchase agreements will, like TARP programs, be funded through the
issuance of public debt. Treasury will also issue public debt to cover any losses that the
GSEs incur because of the additional $25 billion they provide, as long as the GSEs have
liabilities that exceed assets.
10

The front-end DTI ratio used for the HAMP program is the percentage of a borrower’s
gross monthly income required to pay the borrower’s monthly housing expense which is
comprised of mortgage principal, interest, taxes, insurance, and if applicable,
condominium, co-operative, or homeowners’ association dues.

11

Unpaid principal balance limits (prior to modification) are $729,750 for a one-unit
building; $934,200 for a two-unit building; $1,129,250 for a three-unit building; and
$1,403,400 for a four-unit building.

Page 6

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

uses TARP funds to match further reductions on a dollar-for-dollar
basis, down to the target of 31 percent of the borrower’s gross monthly
income. The modified monthly payment is fixed for 5 years or until the
loan is paid off, whichever is earlier, as long as the borrower remains
in good standing with the program. After 5 years, investors no longer
receive payments for cost sharing, and the borrower’s interest rate
may increase by 1 percent a year to a cap that equals the Freddie Mac
rate for 30-year fixed rate loans as of the date that the modification
agreement was prepared. The borrower’s payment would increase to
accommodate the increase in the interest rate, but the interest rate and
monthly payments would then be fixed for the remainder of the loan.
2. Standardized net present value (NPV) model. The NPV model
compares expected cash flows from a modified loan to the same loan
with no modification, using certain assumptions. If the expected
investor cash flow with a modification is greater than the expected
cash flow without a modification, the loan servicer is required to
modify the loan. According to Treasury, the NPV model increases
mortgage investors’ confidence that modifications under HAMP are in
their best financial interests and helps ensure that borrowers are
treated consistently under the program by providing an externally
derived objective standard for all loan servicers to follow.
3. Standardized waterfall. Servicers must follow a sequential
modification process to reduce payments to as close to 31 percent of
gross monthly income as possible. Servicers must first capitalize
accrued interest and certain expenses paid to third parties and add this
amount to the loan balance (principal) amount. Next, the interest rate
must be reduced in increments of one-eighth of 1 percent until the 31
percent DTI target is reached, but servicers may not reduce interest
rates below 2 percent. If the interest rate reduction does not result in a
DTI ratio of 31 percent, servicers must then extend the maturity and/or
amortization period of the loan in 1-month increments up to 40 years.
Finally, if the target DTI ratio is still not reached, the servicer must
forbear, or defer, principal until the payment is reduced to the 31
percent target. Servicers may also forgive mortgage principal at any

Page 7

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

step of the process to achieve the target monthly payment ratio of 31
percent, provided that the investor allows principal reduction. 12
4. Incentive payment structure. Treasury uses TARP funds to provide
both one-time and ongoing incentives (“pay-for-success”) for up to 5
years to non-GSE loan servicers, mortgage investors, and borrowers.
These incentives are designed to increase the likelihood that the
program will produce successful modifications over the long term and
help cover the servicers’ and investors’ costs for making the
modifications.
Borrowers must also demonstrate their ability to pay the modified amount
by successfully completing a trial period of at least 90 days before a loan is
permanently modified and any government payments are made under
HAMP. Treasury has entered into agreements with Fannie Mae and
Freddie Mac to act as its financial agents for MHA. With respect to Freddie
Mac, these responsibilities are carried out by a separate division of that
entity. Fannie Mae serves as the MHA program administrator and is
responsible for developing and administering program operations
including registering servicers and executing participation agreements
with and collecting data from them, as well as providing ongoing servicer
training and support. Within Freddie Mac, the MHA-Compliance (MHA-C)
team is the MHA compliance agent and is responsible for assessing
servicers’ compliance with non-GSE program guidelines, including
conducting on-site and remote servicer loan file reviews and audits.
Initially, only servicers who signed a SPA prior to December 31, 2009, were
eligible to participate in MHA. Subsequently, the Secretary of the Treasury
exercised the authority granted under the Emergency Economic
Stabilization Act of 2008 to extend TARP’s obligation authority to October
3, 2010, which allowed servicers to continue to sign SPAs to participate in
MHA until that time. As of December 31, 2010, there were a total of 143

12

The principal forbearance amount is noninterest-bearing and nonamortizing and cannot
accrue interest under the HAMP guidelines or be amortized over the loan term. Rather, the
amount of principal forbearance will result in a balloon payment fully due and payable
upon the borrower’s transfer of the property, payoff of the interest-bearing unpaid principal
balance, or maturity of the mortgage loan. If, in order to reach the target DTI ratio, the
investor will be required to forbear more than 30 percent of the unpaid principal balance,
or an amount of principal necessary to reach 100 percent of the mark-to-market loan-tovalue ratio (MLTV), the servicer may, but is not required to modify the loan.

Page 8

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

active servicers. 13 Through January 2011, $29.9 billion in TARP funds had
been committed to these servicers for modification of non-GSE loans. 14
Based on the MHA Servicer Performance Report through January 2011,
nearly 1.8 million HAMP trial modifications had been offered to borrowers
of GSE and non-GSE loans as of the end of January 2011, and nearly 1.5
million of these had begun HAMP trial modifications. 15 Of the trial
modifications begun, approximately 145,000 were in active trial
modifications, roughly 539,000 were in active permanent modifications,
roughly 740,000 trial modifications had been canceled, and roughly 68,000
permanent modifications had been canceled. Recently, the number of new
trial and permanent modifications started each month has declined (fig. 2).
As of December 31, 2010, $1 billion in TARP funds had been disbursed for
TARP-funded housing programs, of which $840 million was disbursed for
HAMP-related activity.

13

The GSEs have directed all of their approximately 2,000 servicers to implement parallel
HAMP programs on first-lien mortgages owned or guaranteed by the GSEs.

14

The balance of the difference between this amount and the $45.6 billion allocated to
housing programs was allocated to the FHA Short Refinance Program and the HFA
Hardest-Hit Fund.

15

Roughly 46 percent of borrowers who were either in trial or permanent modifications as
of September 30, 2010, had non-GSE loans and, therefore, fell under the TARP-funded
portion of HAMP.

Page 9

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Figure 2: GSE and Non-GSE HAMP Trial and Permanent Modifications Made and Canceled Each Month, through January 2011
Modifications in thousands
200
Treasury announces
goal of 500,000
trials by
November 1,
2009

150

Start of Treasury's
Conversion Campaign

100

50

0
May June
and prior

July

Aug.

Sept.

Oct.

Nov.

Dec.

2009

Jan.

Feb.

Mar.

2010

Apr.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

Jan.
2011

Trials started
Trials canceled
Permanents started
Permanents canceled
Source: GAO analysis of Treasury data.

Page 10

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Implementation of
Treasury’s Newer
Housing Programs
Has Been Slow and
Capacity of Servicers
to Carry Out These •
Programs Remains
Unclear, Raising
Uncertainty About the
Potential Impact of
These Programs

Treasury has recently implemented programs to reduce or eliminate
payments on second-lien mortgages, provide incentives for the use of
short sales or deeds-in-lieu as alternatives to foreclosure, and provide
incentives for the forgiveness of principal for borrowers whose homes are
worth significantly less than their mortgage balances. However, as of
December 2010, reported activity under these three programs had been
limited. 16

•

HAFA was announced in March 2009 and had disbursed $9.5 million out of
$4.1 billion allocated to the program by the end of December 2010.
Restrictive program requirements—for example, that borrowers be
evaluated for a HAMP first-lien modification before being evaluated for
HAFA, appear to have limited program activity to date. Treasury has taken
steps to revise program guidelines, but it remains to be seen the extent to
which these actions will result in increased program activity.

•

PRA was announced in March 2010 and Treasury had not reported activity
as of December 2010 for this $2 billion program. Mortgage investors and
others have cited concerns that the voluntary nature of the program and
transparency issues, including concerns about the extent of reporting on
PRA activity, may limit the extent to which servicers implement PRA.
Treasury has not yet implemented our June 2010 recommendation that it
report activity under PRA, including the extent to which servicers
determined that principal reduction was beneficial to investors but did not
offer it, to ensure transparency in the implementation of this program
feature across servicers.

2MP was announced in March 2009, and had disbursed $2.9 million out of
nearly $133 million allocated to the program by the end of December 2010.
In part, the limited activity appears to be the result of problems that
servicers have experienced using the database that Treasury required to
identify second-lien mortgages eligible for modification. Treasury has
taken some steps to address these challenges, but could take further
action to ensure that borrowers are aware of their potential eligibility for
the program.

16

In commenting on a draft of this report, Treasury noted that participating servicers were
matching HAMP first liens with second liens in their portfolio, and by December 31, 2010,
had generated over 200,000 matches, which they were in the process of modifying.

Page 11

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Further, Treasury has not incorporated key lessons learned from
implementation challenges it faced with the first-lien program. 17 Similar to
the first-lien modification program, Treasury has not established effective
performance measures for these three programs, including goals for the
number of borrowers it expects to help. As a result, determining the
progress and success of these programs in preserving homeownership and
protecting home values will be difficult.

Challenges in Matching
First- and Second-Lien
Mortgage Data and
Potential Lack of
Awareness of the Program
Have Slowed
Implementation of the
Second-Lien Modification
Program

Under 2MP, Treasury provides incentives for second-lien holders to
modify or extinguish a second-lien mortgage when a HAMP modification
has been initiated on the first-lien mortgage for the same property.
Treasury requires servicers who agree to participate in the 2MP program
to offer to modify the borrower’s second lien according to a defined
protocol when the borrower’s first lien is modified under HAMP. That
protocol provides for a lump-sum payment from Treasury in exchange for
full extinguishment of the second lien or a reduced lump-sum payment for
a partial extinguishment and modification of the borrower’s remaining
second lien. The modification steps for 2MP are similar to those for HAMP
first-lien modifications, with the interest rate generally reduced to 1
percent and the loan term generally extended to match the term of the
HAMP-modified first lien. In addition, if the HAMP modification on the
first lien included principal forgiveness, the 2MP modification must forgive
principal in the same proportion. Servicers were required to sign specific
agreements to participate in 2MP. As of November 2010, 17 servicers were
participating in the program, covering nearly two-thirds of the second-lien
mortgage market.
According to Treasury, 2MP is needed to create a comprehensive solution
for borrowers struggling to make their mortgage payments, but Treasury
officials we interviewed told us that the pace of 2MP modifications had
been slow. Of the six servicers we contacted, five had signed 2MP
participation agreements and represented the majority of potential second
liens covered by servicers participating in the program. 18 Only one of these
five servicers had begun 2MP modifications as of the date we collected
information from these servicers—over 18 months after the program was

17

See GAO-10-634 for a discussion of the implementation challenges associated with the
HAMP first-lien modification program.
18

The remaining servicer told us that it had not signed a 2MP participation agreement since
second liens represented only a small portion of the loans it serviced.

Page 12

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

first announced by Treasury. This servicer reported that it had started
1,334 second-lien modifications. As of January 2011, Treasury had not yet
begun reporting activity under 2MP. According to servicers and Treasury
officials, the primary reason for the slow implementation of 2MP has been
challenges in obtaining accurate matches of first and second liens from the
data vendor required by Treasury. Treasury’s 2MP guidelines specify that
in order for a second lien to be modified under 2MP, the corresponding
first lien must first have been modified under the HAMP first-lien
modification program. Fannie Mae, as the MHA program administrator,
has contracted with a mortgage loan data vendor—Lender Processing
Services (LPS)—to develop a database that would inform second-lien
servicers when the corresponding first lien had been modified under
HAMP. LPS was also the data vendor used by Fannie Mae to process the
loan level data reported by servicers for the HAMP first-lien program.
Under 2MP, participating servicers agree to provide LPS with information
regarding all eligible second liens they serviced. LPS, in turn, provides
participating 2MP servicers with data on second liens that have had the
borrowers’ corresponding first-lien mortgages modified under the HAMP
program. However, the five participating 2MP servicers we spoke with all
expressed concerns about the completeness or accuracy of LPS’ data. In
particular, they noted that differences in the spelling of addresses—for
example, in abbreviations or spacing—could prevent LPS from finding
matches between first and second liens. Additionally, another servicer
reported that first-lien data could be incorrectly reported in LPS—for
example, in one case, a borrower was incorrectly reported as not in good
standing and, subsequently, was reported as canceled from HAMP. This
mistake prevented the borrower’s first and second liens from being
matched, even though the borrower was in good standing and eligible for
2MP. Treasury has also acknowledged that an inability to identify first- and
second-lien matches poses a potential risk to the successful
implementation of 2MP.
Initial 2MP guidelines stated that servicers could not offer a second-lien
modification without a confirmation of a match from LPS, even if they
serviced both first and second liens on the same property and, thus, would
know if the first lien had been modified under HAMP. In November 2010
Treasury provided updated program guidance that revised the match
requirement if servicers serviced both the first and second lien on a
property. According to these updated guidelines, servicers can offer a 2MP
modification when they identify a first- and second-lien match within their
own portfolio or if they have evidence of the existence of a corresponding
first lien, even if the LPS database has not identified it. While this change
may enable more 2MP modifications, Treasury did not release this

Page 13

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

guidance until after participating servicers had already begun
implementing 2MP, more than a year after the program’s guidelines were
first announced in August 2009.
If they do not service both liens, second-lien servicers must rely on LPS for
matching data or obtain sufficient documentation of the HAMP first-lien
modification to identify the match. If the matching data provided by LPS is
not accurate, it is possible that eligible borrowers will not receive secondlien modifications. Treasury noted that there are no standard data
definitions in the servicing industry, making it difficult to match these data
across servicers. To address some of the concerns about inaccurate and
incomplete matches, Treasury officials told us they worked with LPS to
change the matching protocols. Now LPS provides 2MP servicers with a
list of confirmed address matches and a separate list of probable matches
based only on loan number and zip code. Treasury told us that it would
issue additional guidance for handling probable matches, but added that
servicers would be responsible for confirming probable matches with LPS.
Treasury does not require first-lien servicers to check credit reports to
determine if borrowers whose first liens they modified also had second
liens, and if so, the identity of the second-lien servicer. One servicer noted
that credit reports did not always have complete and reliable information.
In addition, Treasury does not require first-lien servicers to inform
borrowers about their potential eligibility for the second-lien program.
Therefore, borrowers may be unaware that their second lien could be
modified and unlikely to inquire with their second-lien servicers about a
second-lien modification. Any gaps in the awareness of 2MP could
contribute to delays in modifying eligible second-lien mortgages or missed
opportunities altogether. Additionally, any delays or omissions increase
the likelihood that the borrower with an eligible second lien may not be
able to maintain the required monthly reduced payments on the modified
first- and unmodified second-lien mortgages and ultimately redefault on
their HAMP first-lien modification.

Page 14

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Treasury Has Taken Some
Recent Steps to Address
Requirements That May
Have Been Affecting
Participation in the
Foreclosure Alternatives
Program

Under HAFA, Treasury provides incentives for short sales and deeds-inlieu of foreclosure as alternatives to foreclosure for borrowers who are
unable or unwilling to complete the HAMP first-lien modification
process. 19 Borrowers are eligible for relocation assistance of $3,000 and
servicers receive a $1,500 incentive for completing a short sale or deed-inlieu of foreclosure. In addition, investors are paid up to $2,000 for allowing
short-sale proceeds to be distributed to subordinate lien holders. Servicers
who participate in the HAMP first-lien modification program are required
to evaluate certain borrowers for HAFA—those whom they cannot
approve for HAMP because, for example, they do not pass the NPV test or
have investors that prohibit modifications; those who do not accept a
HAMP trial modification; and those who default on a HAMP modification.
All six of the large MHA servicers we spoke with identified extensive
program requirements as reasons for the slow implementation of the
program, including the requirement in the initial guidance that borrowers
first be evaluated for a HAMP first-lien modification. Restrictive short-sale
requirements, and a requirement that mortgage insurers waive certain
rights may have also contributed to the limited activity under HAFA. As a
result, they said they did not expect HAFA to increase their overall
number of short sales and deeds-in-lieu. Some of the program
requirements identified by servicers as a reason for the slow
implementation of the program were recently addressed by Treasury’s
December 28, 2010, revisions to its HAFA guidelines.
•

Borrowers had to first be evaluated for HAMP. According to Treasury’s
initial guidelines, borrowers were to be evaluated for a HAMP first-lien
modification before being considered for HAFA, even borrowers who
specifically requested a short sale or deed-in-lieu rather than a
modification. As such, borrowers interested in HAFA had to submit all
income and other documentation required for a HAMP first-lien
modification. According to servicers we interviewed, this requirement was
more stringent than most proprietary short-sale requirements, and
borrowers may have had difficulty providing all of the documentation

19

Under a deed-in-lieu of foreclosure, the homeowner voluntarily conveys all ownership
interest in the home to the lender as an alternative to foreclosure proceedings. In a short
sale, a house is sold by the homeowner through a real estate agency or other means, rather
than through foreclosure, and the proceeds of the sale are less than what the homeowner
still owes on the mortgage. The lender must give permission to such a transaction and can
agree to forgive the shortfall between the loan balance and the net sale proceeds. Under
HAFA, accepting a deed-in-lieu must satisfy the borrower’s entire mortgage obligation in
addition to releasing the lien on the subject property.

Page 15

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

required. For example, one servicer told us that it evaluated borrowers for
proprietary short sales on the basis of the value of the property and the
borrower’s hardship and that income documentation was not required.
Additionally, a HAMP evaluation may add extra time to the short-sale
process. In cases where a borrower had already identified a potential
buyer before executing a short-sale agreement with the servicer, the
additional time required for a HAMP first-lien evaluation may have
dissuaded the buyer from purchasing the property.
In response to this concern, Treasury released updated HAFA guidance on
December 28, 2010, to no longer require servicers to document and verify a
borrower’s financial information to be eligible for HAFA. The updated
guidance requires servicers to notify borrowers who request a short sale
before they have been evaluated for HAMP about the availability of HAMP,
but no longer requires the servicer to complete a HAMP evaluation before
considering the borrower for HAFA, especially in circumstances where the
borrower already has a purchaser for the property. As a result, borrowers
who specifically request a short sale or deed-in-lieu can be considered for
HAFA at the start of the HAMP evaluation process, rather than having to
wait until the completion of the HAMP evaluation process. 20
•

Restrictive short-sale requirements. According to servicers we spoke
with, some HAFA short-sale requirements, such as occupancy
requirements, may have been too restrictive. Specifically, one servicer
cited as too restrictive the requirement in the initial guidelines that a
property not be vacant for more than 90 days prior to the date of the shortsale agreement, and that if it is vacant, it is because the borrower
relocated at least 100 miles away to accept new employment. To address
this concern, Treasury issued updated guidance in December 2010 which
extended the allowed vacancy period from 90 days to 12 months and
eliminated the requirement that the borrower moved to accept
employment, but added a requirement that the borrower had not
purchased other residential property within the prior 12 months. Owneroccupancy restrictions may also limit the number of HAFA short sales and
deeds-in-lieu. One servicer noted that many of the short sales it completed
outside of HAFA were for nonowner-occupied properties, which may
include second homes or commercial properties. However, HAFA offers

20

Treasury’s revised guidelines continue to require servicers to verify the borrower’s
financial hardship by obtaining a signed Hardship Affidavit or Request for Modification and
Affidavit, official documents used in the HAMP first-lien modification program.

Page 16

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

alternatives to foreclosure only for eligible loans under HAMP, which is
intended for a property serving as a borrower’s principal residence.
•

Waiving of rights by mortgage insurers to collect additional sums.
According to Treasury guidelines, “a mortgage loan does not qualify for
HAFA unless the mortgage insurer waives any right to collect additional
sums (cash contribution or a promissory note) from the borrower.” 21 Some
servicers noted that this requirement had prevented some HAFA short
sales from being completed due to difficulties in obtaining approval for
HAFA short sales from mortgage insurers. Lenders frequently require
mortgage insurance for loans that exceed 80 percent of the appraised
value of the property at the time of origination. Under a short-sale
scenario, the mortgage insurance company could be responsible for
paying the mortgage holder or investor for all or part of the losses incurred
under the short sale depending upon the coverage agreement and
proceeds from the sale.
Mortgage insurance representatives we spoke with indicated that while
they supported HAFA participation, they felt that mortgage insurers
should not have to waive their rights to collect additional sums if
borrowers had some ability to pay them. These representatives told us that
they had not seen many requests for approvals of HAFA foreclosure
alternatives, so they did not believe this requirement was a key
impediment for HAFA. However, they agreed that because servicers did
not know whether mortgage insurers would agree to waive their rights, the
requirement could make it more difficult to solicit borrowers for HAFA.
To minimize the impact of this requirement, one mortgage insurance
representative noted that his company commits to responding to servicers
within 48 hours with a decision about whether the mortgage insurance
company agrees to forego a contribution from the borrower.
We plan to continue to monitor the progress of the HAFA program,
including the impact of Treasury’s December 2010 revisions to its HAFA
guidelines as well as the other program requirements identified by
servicers as contributing to the slow implementation of the program, as
part of our ongoing oversight of the performance of TARP.

21

MHA, MHA Handbook (Washington, D.C., Dec. 2, 2010) Section 6.2.1.

Page 17

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Large MHA Servicers
Generally Have Agreed to
Offer Principal Reductions,
but Mortgage Investors
Had Concerns about
Program Design and
Transparency

PRA provides financial incentives to investors who agree to forgive
principal for borrowers whose homes are worth significantly less than the
remaining amounts owed under their first-lien mortgage loans. Treasury’s
PRA guidelines require servicers to consider principal forgiveness for any
HAMP-eligible borrowers with MLTV greater than 115 percent, using both
the standard waterfall and an alternative. 22 While servicers must consider
borrowers for principal forgiveness, they are not required to offer it, even
if the NPV value to modify the loan is higher when principal is forgiven. If
they choose to offer forgiveness, servicers must reduce the balance
borrowers owe on their mortgages in increments over 3 years, but only if
the borrowers remain current on their payments. Servicers must establish
written policies to Treasury detailing when principal forgiveness will be
offered. According to Treasury, a survey of the 20 largest servicers
indicates that 13 servicers are planning to offer principal reduction to
some extent.
Of the six servicers we spoke with, three said that they planned to offer
principal reduction under the program in all cases in which the NPV was
higher with PRA, unless investor restrictions prevented it. 23 As of October
2010, one of these three servicers had begun HAMP trial modifications
with PRA, another had begun implementation of PRA but had not yet
made trial modification offers with PRA, and the third servicer had not yet
completed implementation of the program. The three remaining servicers
we spoke with said they would limit the conditions under which they
would offer principal forgiveness under the program. One servicer offered
PRA only for adjustable-rate mortgage loans, subprime loans, and 2-year
hybrid loans, and the other had developed a “second look” process for
reviewing loans that had a higher NPV result with principal forgiveness.
This servicer reevaluated these loans using its internal estimates of default
rates and did not forgive principal unless its own estimates indicated a
higher NPV with forgiveness. As a result, only 15 to 25 percent of those

22

The alternative waterfall includes principal reduction as the second step, after
capitalization of accrued interest and certain expenses. The mark-to-market LTV is the
unpaid principal balance divided by the property value at the time of modification.

23

The NPV 4.0 model is the updated version of the NPV model that went into effect on
October 1, 2010, and incorporates PRA. The NPV 4.0 model changed several assumptions
from the prior NPV model such as the probability of default based on more recent loan
performance information. The NPV 4.0 model calculates the net present value of the
modification under the standard HAMP waterfall as well as the alternative waterfall under
PRA. The alternative waterfall includes principal reduction as the required second step for
all loans with a LTV ratio greater than 115 percent.

Page 18

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

who otherwise would have received principal forgiveness will receive it
after this “second look” process, according to this servicer. The third
servicer said it would not offer PRA for loans that had mortgage insurance,
noting that mortgage insurers typically took the first loss on a loan and the
PRA would alter that equation with the investor absorbing the full amount
of loss associated with the principal reduction.
Four of the six servicers we contacted told us that investor restrictions
against principal forgiveness would not limit their ability to offer principal
reduction. However, one servicer noted that about half the loans it
serviced had investor restrictions against principal forgiveness. Another
servicer noted that a material number of its servicing agreements with
investors prohibited principal forgiveness.
Mortgage investors we spoke with expressed concern about PRA’s design
and transparency. In particular, they expressed concern that because the
HAMP NPV model did not use an LTV that reflected both the first and
second liens (combined LTV), the model might not reflect an accurate NPV
result. That is, the NPV model might understate the likelihood of redefault
if it did not use the combined LTV. As a result, investors face the prospect
of forgiving principal without knowing the true redefault risk. Further,
although the purpose of PRA is to address negative equity, not taking the
combined LTV into account would underestimate the population of
underwater borrowers since it would not account for any associated
second liens. In addition, under PRA, servicers must forgive principal on
the second lien in the same proportion as the principal forgiven on the first
lien. However, mortgage investors expressed concern about limited
transparency into whether servicers were forgiving principal on the
second lien. Additionally, SIGTARP recommended in July 2010 that
Treasury reevaluate the voluntary nature of the program and consider
changes to ensure the consistent treatment of similarly situated
borrowers. 24 According to Treasury, servicers began reporting PRA activity
in January 2011 for trial and permanent modifications through December
31, but it is still unclear what level of program detail Treasury will publicly
report. We recommended in June 2010 that Treasury report activity under
PRA, including the extent to which servicers determined that principal
reduction was beneficial to mortgage investors but did not offer it, to
ensure transparency in the implementation of this program. Treasury
officials told us they would report PRA activity at the servicer level once

24

SIGTARP, Quarterly Report to Congress (Washington, D.C., July 21, 2010).

Page 19

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

the data were available. We plan to continue to monitor Treasury’s
reporting of PRA and other TARP-funded housing programs.

Treasury Could Do More to
Incorporate Lessons
Learned from the FirstLien Modification Program
in Implementing Newer
Programs

•

In our June 2010 report, we pointed out that it was important that Treasury
incorporate lessons learned from the challenges experienced with the
HAMP first-lien modification program into the design and implementation
of the newer MHA-funded programs. 25 In particular, we noted that it would
be important for Treasury to expeditiously develop and implement these
new programs (including 2MP, HAFA, and PRA) while also developing
sufficient program planning and implementation capacity, including
providing program policies and guidance, hiring needed staff, and ensuring
that servicers are able to meet program requirements. Treasury officials
said they solicited input from servicers and investors when designing 2MP,
PRA, and HAFA, and have begun to perform readiness reviews for these
servicers. However, servicers have cited challenges with changing
guidance under these programs. We also noted that Treasury needed to
implement appropriate risk assessments and meaningful performance
measures in accordance with standards for effective program
management. However, Treasury has not completed program-specific risk
assessments, nor has it developed performance measures to hold itself and
servicers accountable for these TARP-funded housing programs or
finalized specific actions it could take in the event servicers fail to meet
program requirements.
Program planning and implementation capacity. Treasury has provided
servicers with some guidance on the new programs, but some servicers
said that ongoing changes to the guidelines have presented challenges. In
June 2010, we noted that effective program planning included having
complete policies, guidelines, and procedures in place prior to program
implementation. 26 Treasury published initial guidance for 2MP, HAFA, and
PRA prior to the dates these programs were effective, and some servicers
indicated that implementation of these newer programs was smoother
than it was with the first-lien modification program (see fig. 3). However,
other servicers indicated that initial program guidance was unclear and
that additional guidance was issued late in the implementation process.
For example, while Treasury first announced the 2MP program in March
2009, it did not publish specific 2MP guidelines until August 2009 and then

25

GAO-10-634.

26

GAO-10-634.

Page 20

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

issued revisions to the guidelines in March 2010, the first month of official
implementation, with revisions in June 2010 and again in November 2010.
According to the servicers we contacted, ongoing program revisions
presented challenges such as needing to retrain staff and, in some cases,
delayed program implementation. Treasury officials noted that issuing
additional guidance improves the program and is often necessary as
circumstances change. Servicers also reported that while initial guidance
for PRA was issued before the effective date of the program, Treasury did
not issue guidance specific to the NPV 4.0 model until October 1, 2010, the
date PRA became effective. As a result, servicers told us that there was
insufficient time to update internal servicing systems in time to implement
PRA as of its effective date.
Figure 3: Timeline of 2MP, HAFA, and PRA Guidance

3/4: Treasury first announces
incentives to extinguish junior
liens on homes with first-lien
loans that are modified under
HAMP, as well as compensation
for completing short sales or
deeds-in-lieu.

8/13: 2MP implementation
guidance issued—requirement to
use LPS to match first and second
liens, but servicers servicing both
first and second liens do not need
to wait on LPS’ matching service to
offer 2MP modification.

2009

3/26: 2MP revised—servicers are now
required to use LPS to identify all eligible lien
matches for 2MP to offer a 2MP modification,
even in cases where the servicer services
both the first and second liens.
HAFA revised to include increased incentives
for borrowers, servicers, and investors.
Treasury announces several new housing
programs, including PRA.

10/15: Revised PRA
guidance on
consideration of 12/28: Revised
loans that were HAFA guidance
modified under
on changes in
HAMP prior to
vacancy
October 1, 2010. requirements and
timing for issuing
short sale
10/1: Net Present
agreements, with
Value model for
effective date of
PRA ready for
February 1, 2011.
servicers to use.

2010

11/30: HAFA
implementation
guidance issued,
with effective date
of April 5, 2010.

4/28: Treasury announces
additional details related
to the second-lien
modification program.

6/3: Principal Reduction
Alternative implementation
guidance issued, with
effective date of October 1,
2010.
2MP guidance on principal
forgiveness and
forbearance revised.

11/23: Revised 2MP guidance
allows servicers servicing both
first and second liens to offer a
2MP modification when they
identify a match, even if LPS
has not identified it.
12/2: Updated version of the MHA
Handbook consolidates previously
released guidance and includes
guidance for 2MP and HAFA.

Source: GAO.

Treasury has also not completed a needed workforce assessment to
determine whether it has enough staff to successfully implement the new
program. In July 2009, we recommended that Treasury place a high
priority on fully staffing vacancies in its Homeownership Preservation
Office (HPO), the office within Treasury responsible for MHA governance,

Page 21

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

and fill all necessary positions. According to Treasury officials, each
director within HPO conducts ongoing informal assessments of staffing
needs, and Treasury has recently added two positions in marketing and
communications, as well as two additional staff to address policies
regarding the borrower complaint process. In addition, two additional staff
positions to support the borrower complaint resolution process have
recently been approved by the staffing board. HPO has also named a
Deputy Chief. In addition, Treasury officials told us that Fannie Mae and
Freddie Mac, Treasury’s financial agents for MHA, had doubled the
number of staff devoted to these functions as the complexity of MHA has
increased. However, as of December 2010, Treasury had not conducted a
formal workforce assessment of HPO, despite the addition of the new
MHA programs, 2MP, HAFA, and PRA. As we noted in July 2009, given the
importance of HPO’s role in monitoring the financial agents, servicers, and
other entities involved in the $45.6 billion TARP-funded housing programs,
having enough staff with appropriate skills is essential to governing the
program effectively.
Servicers have not demonstrated full capacity to effectively carry out these
programs. Treasury has previously stated that the implementation of the
HAMP first-lien program was hindered by the lack of capacity of servicers
to implement all of the requirements of the program. According to
Treasury, Fannie Mae has conducted program-specific readiness reviews
for the top 20 large servicers for HAFA and PRA, including all 17 servicers
participating in 2MP. These reviews assess servicers’ operational
readiness, including developing key controls to support new programs,
technology readiness, training readiness, as well as staffing resources and
program processes and documentation. According to Treasury officials, 5
servicers have completed readiness reviews for 2MP, and 5 additional
servicers were scheduled to be surveyed in January 2011; 19 servicers have
completed these reviews for HAFA; and 18 servicers have completed these
reviews for PRA. According to Treasury’s summary of these reviews, a
large majority of servicers completing these readiness reviews did not
provide all documentation required to demonstrate that the key tasks
needed to support these programs were in place at the time of the review.
Of those that had complete reviews, 4 had provided all required
documents for HAFA and 3 had provided all required documents for PRA.
None of the servicers provided all required documents for 2MP. Treasury
notes that it relies on Fannie Mae to monitor program readiness and that
MHA-C reviews all programs as part of its on-site reviews. Nonetheless, it
is unclear what actions Treasury has taken to ensure that the servicers
who did not submit the required documentation have the capacity to
effectively implement the programs, making less certain the ability of

Page 22

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

these servicers to fully participate in offering troubled homeowners
second-lien modifications, principal reduction, and foreclosure
alternatives.
•

Meaningful performance measures and remedies. As we also reported in
June 2010, Treasury must establish specific and relevant performance
measures that will enable it to evaluate the program’s success against
stated goals in order to hold itself and servicers accountable for these
TARP-funded programs. While Treasury has established program
estimates of the expected funding levels for 2MP, HAFA, and PRA
programs, it has not fully developed specific and quantifiable servicerbased performance measures or benchmarks to determine the success of
2MP, HAFA, and PRA, including goals for the number of homeowners
these programs are expected to help. Treasury officials told us that they
were using the amounts of TARP funds allocated to MHA servicers to
determine estimated participation rates, but this estimate is adjusted on a
quarterly basis and according to Treasury, is not the best measure for
holding servicers accountable. Treasury officials stated that when data
became available they would assess certain aspects of program
performance—for example, they noted that Treasury planned to assess the
redefault rates of modifications that received PRA or 2MP, compared with
those that did not. However, Treasury has not set benchmarks, or goals,
for these performance measures, as we recommended in June 2010. In
addition, Treasury has not stated how it will use these assessments to hold
servicers accountable for their performance or what remedial actions it
will take in cases where individual servicers are not performing as
expected in these programs. We continue to believe that Treasury should
take steps to establish benchmarks that can be used to hold servicers
accountable for their performance.

•

Appropriate risk assessment. We previously reported that agencies must
identify the risks that could impede the success of new programs and
determine appropriate methods of mitigating these risks. In particular, we
highlighted the need for Treasury to develop appropriate controls to
mitigate those risks before the programs’ implementation dates. Although
Treasury has not systematically assessed risks at the program level,
Treasury officials told us they had identified several risks associated with
2MP, HAFA, and PRA and specified ways to mitigate these risks, and
added they were planning to begin new risk assessments in January 2011
that would be completed by June 2011. According to Treasury officials,
this new round of risk assessments will include 2MP, HAFA, and PRA, but
the programs will not be evaluated individually.

Page 23

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

In addition, Treasury has not yet fully addressed all program-specific risks.
As we have seen, Treasury has acknowledged the risk that the matching
database for 2MP may not identify all first liens modified under HAMP.
While Treasury began addressing this issue in updated guidance released
in November 2010, it cannot yet determine whether all borrowers eligible
for 2MP are being identified and considered for second-lien modifications.
Treasury has also acknowledged several potential risks with all types of
short-sale transactions, including HAFA transactions. According to
Treasury officials, these risks include those arising from sales to allied
parties, side agreements, and rapid resales. For example, Treasury officials
noted a short-sale purchaser could be inappropriately related to the
servicer, allowing the short sale to be inappropriately engineered to
generate extra compensation for one or both parties. Treasury states that
HAFA includes requirements to mitigate these risks, such as requiring
arms-length transactions. According to Treasury officials, MHA-C, the
group within Freddie Mac that acts as Treasury’s financial agent for MHA
compliance activity, is also in the process of developing compliance
procedures to address these risks. Further, Treasury has identified several
potential risks with PRA, including servicer noncompliance with PRA
requirements, moral hazard (the risk that borrowers would default on their
mortgages to receive principal reduction when they otherwise would not
have), and low program participation. According to Treasury officials,
these risks will be mitigated through regular compliance reviews, servicer
reporting of NPV results both with and without PRA, and other program
requirements. For example, to guard against moral hazard, Treasury
requires that borrowers be experiencing hardship and that servicers
forgive the principal over 3 years only if the borrower remains current on
the modified payments. However, low program participation may continue
to be a risk for PRA, despite the initial participation plans of several of the
large servicers. While Treasury officials told us they plan to monitor the
reasonableness of the extent of principal forgiveness on a servicer-specific
basis, we continue to believe that due to the voluntary nature of the
program, Treasury will need to ensure full and accurate servicer-specific
reporting of program activity for future assessments of the extent to which
servicers are offering PRA when the NPV is higher with principal
forgiveness, as we recommended in June 2010. We plan to continue to
monitor and report on Treasury’s risk assessment and control activities for
MHA programs as part of our ongoing oversight of Treasury’s use of TARP
funds to preserve homeownership and protect property values.

Page 24

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Treasury Has Some
Data on the
Characteristics of
Borrowers in HAMP’s
First-Lien Program,
but Data Were
Sometimes Missing or
Questionable

Our analysis of Treasury’s HAMP data through September 30, 2010,
indicated that borrowers who entered into trial modifications or received
permanent modifications continued to have elevated levels of debt, as
evidenced by the median back-end DTI for these two groups (55 and 57
percent, respectively). 27 Borrowers who received a trial modification
based on stated (unverified) income—a practice that Treasury no longer
permits—were the most likely to have their trial modifications canceled,
and borrowers who were the most delinquent on their mortgage payments
at the time of applying for a loan modification were the most likely to
redefault on their modifications. While the data Treasury collected from
the servicers provided these and other insights into the characteristics of
borrowers helped under the program, some data were missing and some
information was inaccurate, preventing certain types of analyses of HAMP
borrowers. For example, race and ethnicity information was not available
for a significant portion of borrowers. In addition, Treasury’s data on
borrowers’ LTV ratios at the time of modification ranged from 0 to 999,
with 1 percent of non-GSE borrowers in active permanent modifications
reporting ratios over 400 percent, implying that some borrowers who
received HAMP modifications did not have a mortgage, and others had
loan amounts more than 4 times the value of their homes. Treasury said
that it and Fannie Mae were continuing to refine and strengthen data
quality checks and that the data would improve over time.

Certain Factors Increase
the Likelihood of Trial
Modification Cancellation
and Early Data Indicate
that Borrowers Who
Redefaulted from
Permanent Modifications
Were Further Into
Delinquency

According to Treasury’s HAMP data, 88,903 non-GSE borrowers were in
active HAMP trial modifications and 205,449 borrowers were in permanent
modifications as of the end of September 2010. These borrowers generally
cited a reduction in income as their primary reason for hardship when
applying for HAMP modifications.
•

Over half of borrowers cited a “curtailment of income,” such as a change
to a lower-paying job, as the primary reason they were experiencing
financial hardship (56 percent and 53 percent of those in active trial and
permanent modifications, respectively). However, only 5 percent of
borrowers in each of these groups cited unemployment as their primary
reason for hardship.

27

Back-end DTI ratio consists of items included in the front-end DTI (principal, interest,
taxes, insurance, and any homeowners’ association or condominium fees associated with
the first-lien mortgage and property) and all other monthly debt payments (installment
debts, payments on junior liens, alimony, car payments, etc.).

Page 25

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

•

Borrowers in trial and permanent modifications through September 2010
also had high levels of debt prior to modification—median front-end DTI
ratios of 45 and 46 percent, and back-end DTI ratios of 72 and 76 percent,
respectively. Even after modification, these borrowers continued to have
high debt levels (median back-end DTI ratios of 55 and 57 percent for
those in trial and permanent modifications, respectively). Treasury has
defined a high back-end DTI to be 55 percent, and has required borrowers
with total postmodification debt at this level to obtain counseling.

•

In addition, borrowers in trial and permanent modifications tended to be
“underwater,” with median mark-to-market LTV ratios of 123 percent and
128 percent, respectively.
Borrowers who were unsuccessful in HAMP modifications, either because
they were canceled from a trial modification or because they redefaulted
from permanent modifications, shared several of these characteristics,
including having high levels of debt and being “underwater” on their
mortgages. However, some characteristics appeared to increase the
likelihood that a borrower would be canceled from a trial modification.
Holding other potential factors constant, the following factors increased
the likelihood that a borrower would be canceled from a trial
modification:

•

Use of Stated Income. Borrowers who received a trial modification based
on stated income were 52 percent more likely to be canceled from trial
modifications than those who started a trial modification based on
documented income. In some cases, borrowers who received trial
modifications based on stated income were not able to or failed to provide
proof of their income or other information for conversion to permanent
modification. 28 In other cases, borrowers may have submitted the required
documentation but the servicer lost the documents. Over one-third of the
396 housing counselors who responded to our survey identified servicers
losing documentation as the most common challenge that borrowers have
faced in providing the required documentation for a permanent
modification. In December 2010, the Congressional Oversight Panel also

28
Treasury has recognized challenges with documentation as a reason for the low
conversion rate to permanent modifications and, as of June 2010, began requiring that
servicers verify borrowers’ income before placing borrowers into trial modifications.

Page 26

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

reported that Treasury has failed to hold loan servicers accountable when
they have repeatedly lost borrowers’ paperwork. 29
•

Length of Trial Period. Borrowers who were in trial modification periods
for fewer than 4 months were about 58 percent more likely to have their
trial modifications canceled than borrowers in longer trial periods. This
finding may indicate that borrowers who default on their trial
modifications will do so earlier in the process rather than later.

•

Delinquency Level at Time of Modification. Borrowers who were 60 or 90
days or more delinquent at the time of their trial modifications were 6 and
9 percent more likely to have trial modifications canceled, respectively,
compared with borrowers who were not yet delinquent at the time of their
trial modifications. Treasury has acknowledged the importance of
reaching borrowers before they are seriously delinquent by requiring
servicers to evaluate borrowers still current on their mortgages for
imminent default, but as we noted in June 2010, this group of borrowers
may be defined differently by different servicers. 30 In addition, most
borrowers who received HAMP were delinquent on their mortgages at the
time of modification—as of September 30, 2010, 83 percent of those who
had begun trial or permanent modifications were at least 60 days
delinquent on their mortgages.
According to our analysis, there were also several factors that lowered the
likelihood of trial cancellations, although the effect was generally smaller
than the factors that increased the likelihood of being canceled.

•

High MLTV Ratio. Borrowers who had high MLTV ratios (above 120
percent) were less likely to be canceled from a trial modification
compared to those with MLTV ratios at or below 80 percent. That is, loans
with a MLTV between 120 and 140 percent were 7 percent less likely to be
canceled, while loans with an MLTV of more than 140 percent were 8
percent less likely to be canceled.

•

Amount of Principal or Payment Reduction: While only about 2 percent
of borrowers had received principal forgiveness as of September 30, 2010,
borrowers who received principal forgiveness of at least 1 percent of their
total loan balance were less likely to be canceled from trial modifications,

29

Congressional Oversight Panel, December Oversight Report: A Review of Treasury’s
Foreclosure Prevention Programs (Washington, D.C., Dec. 14, 2010).

30

GAO-10-634.

Page 27

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

compared with those who did not receive principal forgiveness. In
addition, larger monthly payment reductions lowered the likelihood that a
trial modification would be canceled. For example, our analysis showed
that borrowers who received a principal and interest payment reduction of
least 10 percent were less likely to be canceled from their trial
modifications than borrowers who received a payment reduction of less
than 10 percent or who had an increase in payments.
Figure 4 illustrates the extent to which certain factors increase or
decrease likelihood of borrowers being canceled from HAMP trial
modification. See appendix II for further details on our analysis of factors
affecting the likelihood of trial modification cancellation.
Figure 4: Estimated Decrease/Increase in Likelihood of Cancellation of HAMP Trial
Modification by Borrower and Loan Characteristics
Change in likelihood of
trial modification cancellation

Characteristics
Loan had loan-to-value ratio greater than 140
percent, compared to 80 percent or less

-8%

Loan had loan-to-value ratio between 120 and 140
percent, compared to 80 percent or less

-7%

Borrower received principal forgiveness of between 1 and
50 percent of total loan balance

-6%

Borrower's principal and interest payment on loan reduced
by more than 20 percent, compared to a decrease of 10
percent or less or an increase

-5%

Borrower's principal and interest payment on loan reduced
by between 10 and 20 percent, compared to a decrease of
10 percent or less or an increase.

-5%

Borrower was 60 to 89 days days delinquent prior to trial
modification, compared to being current on mortgage
payments
Borrower was 90 or more days delinquent prior to trial
modification, compared to being current on mortgage
payments
Borrower was evaluated for trial modification based on
stated income prior to June 1, 2010)

6%

9%

52%

Borrower was in trial modification period for 4 months or
less

58%

Source: GAO analysis of Treasury data.

In addition, our initial observations of over 15,000 non-GSE borrowers
who had redefaulted from permanent HAMP modifications through

Page 28

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

September 2010 indicated that these borrowers differed from those in
active permanent modifications in several respects. Specifically, non-GSE
borrowers who redefaulted on their HAMP permanent modifications
tended to have the following characteristics
•

higher levels of delinquency at the time of trial modification evaluation
(median delinquency of 8 months compared to 5 months for those still in
active permanent modifications);

•

lower credit scores, although borrowers current on their HAMP-modified
payments also had low median credit scores (525 and 552, respectively);

•

lower median percentage of payment reduction compared with those who
were still current in their permanent modifications (24 percent compared
with 33 percent for those who were still current in their permanent
modifications); and

•

lower levels of debt before modification than borrowers who did not
redefault (median front-end DTI ratio of 41 percent prior to modification
compared to 46 percent front-end DTI ratio for those still current in their
permanent modifications)—these borrowers likely did not receive as
much of a payment reduction from the modification due to lower levels of
debt to begin with.
These results were largely consistent with information that the Federal
Deposit Insurance Corporation (FDIC) released on the performance of its
IndyMac loan modifications. For example, FDIC found that borrowers’
delinquency status prior to loan modification correlated directly with
redefault rates after modification, with a 1-year redefault rate of roughly 25
percent for borrowers who were 2 months delinquent at the time of
modification compared to a nearly 50 percent redefault rate for those who
were more than 6 months delinquent at the time of modification. 31 FDIC
also reported that the redefault rates for its IndyMac modifications
declined markedly with larger reductions in monthly payments.

31
Richard Brown, The FDIC Loan Modification Program at IndyMac Federal Savings
Bank. Presented at Mortgages and the Future of Housing Finance Conference, Washington,
D.C., Oct. 25, 2010.

Page 29

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Some Key Information on
HAMP Borrowers and
Applicants Was Missing or
Inaccurate in Treasury’s
Database

Treasury’s data on HAMP provide important information and insights on
characteristics of borrowers who are in trial and permanent modification,
who have been canceled from trial modifications, and who have
redefaulted from permanent modifications. However, Treasury’s database
contained information that was inaccurate or inconsistent, and Treasury
does not collect information on all borrowers who are denied HAMP
modifications. For example, Treasury’s data on borrowers’ LTV ratios at
the time of modification ranged from 0 to 999, with 1 percent of non-GSE
borrowers in active permanent modifications reporting ratios over 400
percent, implying that some borrowers who received HAMP modifications
did not have a mortgage, and others had loan amounts more than 4 times
the value of their homes. Some data elements also included internal
inconsistencies. For example, a borrower’s back-end DTI (the ratio of total
monthly debt-to-gross monthly income) includes the front-end DTI (the
ratio of monthly housing debt-to-gross monthly income) and, therefore,
should always at least be equal to the front-end DTI. However, according
to Treasury’s database, 29 percent of those in trial modifications and 40
percent of those who had trial modifications canceled had back-end DTIs
that were less than their front-end DTIs. The quality of these data
improved for those who received permanent modifications, with only 3
percent of these borrowers showing back-end DTIs that were less than the
front-end DTIs.
Treasury acknowledged that its HAMP database contained some
inconsistencies, despite edit checks conducted by Fannie Mae as the
HAMP administrator. According to Treasury, the inconsistencies continue
because of servicers’ data-entry errors, data formatting mistakes such as
entering percentages as decimals rather than whole numbers, and data
mapping problems. Treasury said it was continuing to work with Fannie
Mae to refine and strengthen data quality checks and that the data has and
will continue to improve over time. For example, Treasury noted that
since September 2010, it has worked to improve the quality of borrower
and loan attributes such as back-end DTI and modification terms. Treasury
officials said that the error rate on these data elements has dropped from
16 percent and 12 percent for trial and permanent modifications,
respectively, to 2 percent and 10 percent.
Treasury’s HAMP database also was missing a significant amount of
information on borrowers’ race and ethnicity, resulting in an inability to
date to assess whether HAMP is being fairly implemented across servicers.
For example, as of September 30, 2010, race and ethnicity information was
not available for 65 percent of non-GSE borrowers in active trial
modifications. A significant portion of borrowers declined to report this

Page 30

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

information—that is, for 45 percent of non-GSE borrowers in active trial
modifications the category was marked as “not provided by borrower.”
However, for another 20 percent, some data are simply missing, with no
category marked. Some of this information may be missing because
servicers were not required to report borrowers’ race and ethnicity until
after December 1, 2009. As a result, Treasury lacks complete information
needed to be able to determine whether the first-lien modification
program has been implemented fairly across all borrowers.
In addition, Treasury acknowledged data-mapping problems with race and
ethnicity data that resulted in some data being included in the system of
record, but inadvertently excluded from the database. Combined, these
factors resulted in a large proportion of borrowers without race and
ethnicity information, as of September 30, 2010. According to Treasury
officials, Fannie Mae was making improvements to the data mapping,
which should allow Treasury to better evaluate whether HAMP is being
implemented fairly across all borrowers. Treasury officials told us they
anticipated that the more complete data would be ready to use in early
2011. On January 31, 2011, Treasury announced the availability of loanlevel HAMP data to the public for the first time. The data files were as of
November 30, 2010, and included information on borrowers’ race and
ethnicity. According to Treasury, these data indicated that roughly 31
percent of borrowers who started trial modifications after December 1,
2009, did not report race and ethnicity data. Treasury also reported
approximately 6 percent of data as not applicable or not reported by the
servicer. In addition, roughly 57 percent of those who were denied or did
not accept trial modifications did not report or were missing this
information.
Finally, Treasury’s HAMP database did not contain information on all
borrowers who were denied HAMP, as some borrowers were denied
before income information was collected for a net present value test.
Treasury currently requires servicers to report identifying information,
such as borrowers’ names and Social Security numbers, as well as the
reason for denial for all borrowers denied modification, but other data
elements—including income information, level of delinquency, LTV, and
GSE or non-GSE status—is not required to be collected by servicers if
borrowers are denied because they do not meet basic eligibility
requirements such as the property being owner-occupied. According to
data we received from Treasury, through September 30, 2010, some
information was lacking on 85 percent of borrowers who were denied
HAMP trial modifications, including monthly gross income amounts and
the number of months in delinquency. Treasury noted that these data are

Page 31

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

incomplete because they are unobtainable by the servicers and not a good
use of servicer resources to obtain. While we recognize that servicers may
be unable to collect information from borrowers who were previously
denied trial modifications, going forward it will be important for Treasury
to collect sufficient information from servicers to assess program gaps.
According to Treasury, it has requested servicers to report on borrowers
who were denied HAMP when low volumes of these data were received.

Most Borrowers
Denied or Canceled
from Trial
Modifications Appear
to Have Avoided
Foreclosure To Date,
but Weaknesses in
Treasury’s Data
Collection Limit its
Ability to Understand
the Outcomes of
These Borrowers

Because there have been more HAMP trial modification cancellations than
conversions to permanent modifications, we evaluated Treasury’s
reporting of the disposition paths, or outcomes, of borrowers who were
denied or canceled from HAMP trial modifications and obtained additional
information from six large MHA servicers to understand the extent to
which these borrowers have been able to avoid foreclosure to date. While
it appears that the majority of these borrowers had been able to avoid
foreclosure as of the time of our data collection and Treasury’s survey, if
borrowers are being evaluated for a loss mitigation option such as a
proprietary modification and the servicer has also started foreclosure
proceedings, Treasury’s data reporting template will result in a loan being
reported only as a proprietary modification or the other applicable loss
mitigation category, understating the number of borrowers who have had
foreclosure proceedings started. In addition, Treasury’s reporting of
outcomes for these borrowers does not differentiate between borrowers
who received proprietary modifications and those who were still being
evaluated for these modifications, some of whom will not ultimately
receive them. For example, for six large servicers, Treasury reported that
43 percent of borrowers who had their trial modification canceled
received proprietary modifications. 32 However, the reported 43 percent
includes both borrowers who had received proprietary modifications and
those who were being evaluated for proprietary modifications. Data we
collected from the same servicers indicate that only 18 percent of
borrowers with canceled trial modifications received permanent
proprietary modifications, while another 23 percent had pending but not
yet approved permanent modifications. Without a complete picture of the
outcomes of those borrowers who were denied or canceled from HAMP,
Treasury cannot accurately evaluate the outcomes for these borrowers

32

Treasury publicly reports these outcomes for the eight largest HAMP servicers, but we
calculated the percentages for six of these servicers based on Treasury’s report in order to
compare them with the data we received from these same servicers.

Page 32

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

and determine whether further action may be needed to assist this group
of borrowers.

Treasury’s Reporting Does
Not Fully Reflect the
Current Disposition
Actions for Borrowers
Denied or Canceled from
HAMP

According to HAMP guidelines, servicers must consider all potentially
HAMP-eligible borrowers for other loss mitigation options, such as
proprietary modifications, payment plans, and short sales, prior to a
foreclosure sale. To report the current outcomes of borrowers who
applied for but did not receive a HAMP trial modification or had a HAMP
trial modification canceled, Treasury surveys the eight largest HAMP
servicers each month and publishes these data in the monthly servicer
performance reports. However, Treasury’s requirements for reporting
these data produce results that do not fully reflect all outcomes for
borrowers who were denied or canceled from HAMP and overstate the
proportion of some outcomes. First, in order to prevent double counting of
transactions, the survey does not allow servicers to place a borrower in
more than one outcome category. Additionally, servicers must follow the
order in which Treasury lists the outcomes on the survey. However, this
does not allow for the accurate reporting of borrowers being considered
for multiple potential outcomes. For example, a servicer could be
evaluating a borrower who had been denied a HAMP modification for a
proprietary modification at the same time that the servicer started
foreclosure proceedings. But the Treasury survey would capture only the
proprietary modification, because that category is the first in the list of
possible outcomes. Because servicers are allowed to evaluate borrowers
for loss mitigation options while simultaneously starting foreclosure,
Treasury’s requirement that borrowers be included in only one category,
starting with proprietary modifications, likely overstates the proportion of
borrowers with proprietary modifications while also understating the
number of borrowers who have started foreclosure.
Furthermore, a comparison of Treasury’s data to data we received from
six large MHA servicers on the outcomes of borrowers denied a HAMP
trial modification showed that Treasury’s requirement that servicers place
borrowers according to a specific order of outcomes may result in an
understatement of the number of borrowers becoming current. For
example, according to the data we received, almost 40 percent of
borrowers who were denied a HAMP trial modification became current
without any additional assistance from the servicer as of August 31, 2010.
In comparison, Treasury reported only 24 percent of borrowers became
current after applying for but not receiving a HAMP trial modification
through these same servicers. While differences may exist between the
populations of these data, a servicer we spoke with noted one reason that

Page 33

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

the percentage of current borrowers in the Treasury survey was lower
than the percentage reported in our data was Treasury’s requirement that
servicers report outcomes in a certain order, with “borrower current”
being in last place. 33 As a result, borrowers are reviewed for all other
outcomes before being reflected in this category. Placing borrowers only
in one category according to a specific order may not reflect all of the
outcomes experienced by these borrowers and may understate outcomes
further down the list, such as starting foreclosure or becoming current.
Second, while Treasury’s survey includes an “action pending” category, all
six of the servicers we spoke with told us that Treasury had instructed
them to include borrowers who were being evaluated for an outcome in
their respective outcome categories, such as proprietary modification,
rather than the “action pending” category. Treasury recently instructed
servicers to use the action pending category only if a borrower had
recently been denied a HAMP trial modification, had a HAMP trial
modification canceled, or fallen out of another disposition path such as a
proprietary modification, and the servicer has not yet determined the next
step for the borrower. Because the proprietary modification category
includes borrowers who are still being evaluated for modifications as well
as those who have received them, the number of borrowers who actually
received a proprietary modification cannot be determined from Treasury’s
data. For example, for the outcomes of borrowers who had a canceled
HAMP trial modification, we asked six large MHA servicers to separate
borrowers who were being evaluated for permanent proprietary
modifications from those who had actually received them. For these same
six servicers, while Treasury reported that 43 percent of borrowers who
canceled from a HAMP trial modification through August 2010 were in the
process of obtaining a proprietary modification, the data we received
indicated that 18 percent of these borrowers had received permanent

33

Several differences exist between the populations of borrowers reported in our data and
Treasury’s report. First, our data included only borrowers who were denied a HAMP trial
modification, while Treasury’s also included borrowers who were offered but declined this
option. Second, Treasury did not begin requiring servicers to report on borrowers who
applied for but did not receive a HAMP trial modification until December 1, 2009, so some
servicers did not have data on these borrowers until that date. Third, one servicer reported
to us borrowers in a business division not reported in Treasury’s survey, and another
reported borrowers to Treasury for a business division not included in our report. Fourth,
our data may reflect loans that have not yet been reported to Treasury due to differences in
timing of reporting. Finally, Treasury defines “current” as being less than 60 days
delinquent, while our “current” category includes borrowers who ranged from 0 to 30 days
past due.

Page 34

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

proprietary modifications, and 23 percent were in the process for being
approved for one. 34 By including borrowers who received permanent
proprietary modifications alongside borrowers who were still in the
process for getting one, Treasury may not fully understand the extent to
which servicers are providing permanent assistance to borrowers being
denied or canceled from HAMP trial modifications.
While Treasury has taken steps to collect data on the outcomes of
borrowers who do not receive a HAMP trial or permanent modification—
data that could be used to assess the extent to which these borrowers are
receiving other loss mitigation programs—the way in which Treasury has
asked servicers to report these data overstates the proportion of certain
outcomes and understates others, such as starting foreclosure
proceedings. In addition, Treasury’s reporting does not differentiate
between those who have received a proprietary modification and those
who are being evaluated for one. If the information presented in the
monthly servicer performance reports does not fully reflect the outcomes
of these borrowers, Treasury and the public will not have a complete
picture of their outcomes. Further, Treasury cannot determine the extent
to which servicers provided alternative loss mitigation programs to
borrowers denied or canceled from HAMP or evaluate the need for further
action to assist this group of borrowers.

Outcomes of Borrowers
Vary by Whether
Borrowers were Denied,
Canceled, or Redefaulted,
and by Servicer

We requested data from six servicers on the outcomes of borrowers who
(1) were denied a HAMP trial modification, (2) had a canceled HAMP trial
modification, or (3) redefaulted from a HAMP permanent modification.
According to the data we received, of the about 1.9 million GSE and nonGSE borrowers who were evaluated for a HAMP modification by these
servicers as of August 31, 2010, 38 percent (713,038) had been denied a
HAMP trial modification; 27 percent (505,606) had seen their HAMP trial
modifications canceled; and 1 percent (20,561) had redefaulted from a
HAMP permanent modification. 35 We requested that the servicers report all

34

We requested that servicers provide the data as of August 31, 2010, but servicers may
report borrowers with a canceled HAMP trial modification to Treasury until early
September 2010, for August 2010 reporting. In addition, servicers may have included loans
in our data request that have not yet been reported to Treasury and, therefore, would not
be reflected in the number of borrowers that Treasury reports. Lastly, one servicer
reported borrowers to Treasury for a business division not included in our data.

35

Two servicers provided the data as of their closing date for reporting August 2010 data to
Treasury, September 6, 2010, and September 8, 2010, respectively.

Page 35

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

of the outcomes borrowers had received and they separate those who
were being evaluated for an outcome from those who had received them. 36
According to the data we received, borrowers experienced different
outcomes, depending on whether they were denied a HAMP trial
modification, received but were canceled from a trial modification, or
redefaulted from a permanent modification.
According to these servicers’ data through August 31, 2010, borrowers
who were denied HAMP trial modifications were more likely to become
current on their mortgages without any additional help from the servicer
(39 percent) than to have any other outcome (see fig. 5). 37 According to
one servicer, borrowers who were denied a HAMP trial modification were
often current when they applied for a HAMP modification and, once
denied, were likely to remain current. In addition, 9 percent of these
borrowers paid off their loans. Twenty-eight percent of borrowers who
had been denied trial modifications received or were in the process for
receiving a permanent proprietary modification or a payment plan. 38
Servicers initiated foreclosure proceedings on 17 percent at some point
after being denied, while only 3 percent of borrowers completed
foreclosure. 39 Several servicers explained that loss mitigation efforts can
often work in tandem, so a borrower could be referred for foreclosure and
evaluated for another outcome at the same time, and borrowers who were
referred for foreclosure may not necessarily complete it.

36
Because we requested that servicers report all outcomes that a borrower received, a
borrower may be reflected more than one time across these outcomes. One servicer only
provided the most recent outcome of these borrowers.
37

We requested that servicers provide the number of borrowers who were 0 days past due
on their original loan without need for further loss mitigation efforts. Two servicers
provided the number of borrowers who were 0 to 29 days delinquent, while another
servicer provided the number of borrowers who were 0 to 30 days delinquent.

38

Three servicers were unable to provide the number of borrowers who had a payment plan
pending because they only track payment plans once the payment plan has been set up or
the borrower begins making payments.

39

We requested that servicers provide the number of borrowers who were referred for
foreclosure at any time after redefaulting, having their modification canceled, or being
denied.

Page 36

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Figure 5: Outcomes of Borrowers Denied a HAMP Trial Modification, through
August 31, 2010 (Six large MHA servicers)
Percentage
40
35
30
25
20
15
10
5
0
Borrower
current

Loan
payoff

Proprietary
modification

Payment
plan

Foreclosure Foreclosure Foreclosure
alternativea
start
completion

Other
categoriesb

Pending action
Source: GAO analysis of data received from six large MHA servicers.

Note: Borrowers may be included in more than one category.
a

The percentage of borrowers who received a foreclosure alternative may include borrowers who
have a short-sale agreement signed but have not closed on the short sale.

b

Other categories include borrowers who had a bankruptcy in process and no other loss mitigation
effort was allowed at some point after being denied, canceled, or redefaulting; borrowers who had
action pending outside of a proprietary modification, payment plan, or foreclosure alternative; and
borrowers not able to be reflected in any of the other outcomes, such as borrowers who currently
have no workout plan in process.

Of those borrowers who were canceled from a HAMP trial modification,
servicers often initiated actions that could result in the borrower retaining
the home. Specifically, 41 percent of these borrowers received or were in
the process for receiving a permanent proprietary modification, and 16
percent received or were in the process for receiving a payment plan (see
fig. 6). However, servicers started foreclosure proceedings on 27 percent
of borrowers at some point after the HAMP trial modification being
canceled, but, similar to borrowers who were denied a HAMP trial
modification during this time period, a small percentage completed
foreclosure (4 percent). Compared with borrowers who were denied,
borrowers who had a HAMP trial modification canceled were less likely to
become current on their mortgages (15 percent) or to pay off their loan
(4 percent).

Page 37

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Figure 6: Outcomes of Borrowers who Had a Canceled HAMP Trial Modification,
through August 31, 2010 (Six large MHA servicers)
Percentage
50

40

30

20

10

0
Borrower
current

Loan
payoff

Proprietary
modification

Payment
plan

Foreclosure Foreclosure Foreclosure
alternativea
start
completion

Other
categoriesb

Pending action
Source: GAO analysis of data received from six large MHA servicers.

Note: Borrowers may be included in more than one category.
a

The percentage of borrowers who received a foreclosure alternative may include borrowers who
have a short-sale agreement signed but have not closed on the short sale.

b

Other categories include borrowers who had a bankruptcy in process and no other loss mitigation
effort was allowed at some point after being denied, canceled, or redefaulting; borrowers who had
action pending outside of a proprietary modification, payment plan, or foreclosure alternative; and
borrowers not able to be reflected in any of the other outcomes, such as borrowers who currently
have no workout plan in process.

There were wide ranges in the outcomes among servicers we contacted
for borrowers who were canceled from HAMP trial modifications (see
table 1). For example, of those borrowers who had a canceled HAMP trial
modification, one servicer reported that 26 percent had obtained a
proprietary modification through August 31, 2010, compared with 14
percent for another servicer. In addition, for borrowers who had a
canceled HAMP trial modification, one servicer reported foreclosure
completion rates of almost 7 percent, while another servicer reported
foreclosure completion rates of roughly 1 percent. Servicers reported a
wide range of outcomes, which depend on factors such as the composition
of loan portfolios and proprietary loss mitigation programs, including
modifications, payment plans, and short sales. These programs can differ
in design and may have, among other things, different eligibility
requirements for borrowers.

Page 38

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Table 1: Selected Outcomes of Borrowers who Had a Canceled HAMP Trial Modification by Servicer, through August 31, 2010
Action pending
Proprietary
modification

Payment
plan

Current

Foreclosure
alternativea

Foreclosure
completion

Proprietary
modification

Payment
plan

Foreclosure
alternative

Servicer 1

24%

4%

12%

1%

5%

26%

N/Ab

4%

Servicer 2

22

76

35

2

2

6

N/A

2

Servicer 3

26

1

3

1

7

24

N/A

2

Servicer 4

14

1

15

2

1

16

6%

14

Servicer 5

20

1

1

2

3

39

<1

6

Servicer 6

16

0

5

0

6

42

<1

4

18%

14%

15%

1%

4%

23%

3%

7

Average (all
servicers)

Source: GAO analysis of data received from six large MHA servicers.

Note: This table does not reflect all outcomes that borrowers may have received, such as being
referred for foreclosure or currently being reviewed for loss mitigation options. Borrowers may be
included in more than one category.
a

The percentage of borrowers who received a foreclosure alternative may include borrowers who
have a short-sale agreement signed but have not closed on the short sale.

b

N/A indicates servicer was unable to report this data.

Finally, of the borrowers who redefaulted from a HAMP permanent
modification, almost half were reflected in categories other than
proprietary modification, payment plan, becoming current, foreclosure
alternative, foreclosure, or loan payoff (see fig. 7). Twenty-eight percent of
borrowers who redefaulted from permanent modifications were referred
for foreclosure at some point after redefaulting, but, like borrowers denied
or canceled from a HAMP trial modification, the percentage of borrowers
who completed foreclosure remained low relative to other outcomes (less
than 1 percent). Unlike borrowers who were denied or canceled,
borrowers who redefaulted were less likely to receive or be in the process
for receiving a permanent proprietary modification or payment plan after
redefaulting, with 27 percent of borrowers receiving or in the process for
receiving one of the outcomes. In addition, less than 1 percent of
borrowers who redefaulted became current as of August 31, 2010. 40

40

Because borrowers who redefault on a HAMP modification would still retain the terms of
their HAMP modification, we would not expect many borrowers who redefaulted to receive
a proprietary modification. One servicer, however, reported that 95 percent of those
borrowers who redefaulted from a HAMP permanent modification had an action pending
for a proprietary modification. The servicer explained that it evaluates the majority of these
borrowers for another modification program after they redefault.

Page 39

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Figure 7: Outcomes of Borrowers Who Redefaulted on a HAMP Permanent
Modification, through August 31, 2010 (Six large MHA servicers)
Percentage
50

40

30

20

10

0
Borrower
current

Loan
payoff

Proprietary
modification

Payment
plan

Foreclosure Foreclosure Foreclosure
alternativea
start
completion

Other
categoriesb

Pending action
Source: GAO analysis of data received from six large MHA servicers.

Note: Borrowers may be included in more than one category.
a

The percentage of borrowers who received a foreclosure alternative may include borrowers who
have a short-sale agreement signed but have not closed on the short sale.

b

Other categories include borrowers who had a bankruptcy in process and no other loss mitigation
effort was allowed at some point after being denied, canceled, or redefaulting; borrowers who had
action pending outside of a proprietary modification, payment plan, or foreclosure alternative; and
borrowers not able to be reflected in any of the other outcomes, such as borrowers who currently
have no workout plan in process.

Page 40

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Proprietary Modifications
Offer Additional Flexibility
and More Borrowers Have
Received Them than
HAMP Permanent
Modifications

As noted above, servicers have reported that many borrowers who were
denied, canceled, or redefaulted from HAMP have received or were being
evaluated for proprietary modifications. According to HOPE NOW,
servicers completed over 1.2 million proprietary modifications from
January 2010 through December 2010, compared with roughly 513,000
permanent HAMP modifications (see fig. 8). 41
Figure 8: Number of Proprietary and HAMP Modifications Started Each Month,
January through December 2010
Modifications (in thousands)
140

120

100

80

60

40

20

0
Jan.

Feb.

Mar.

Apr.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

Month
Permanent proprietary modifications
HAMP permanent modifications
Sources: HOPE NOW for proprietary modifications and Treasury for HAMP modifications.

In designing the HAMP program, Treasury stated that it had to balance the
needs of taxpayers, investors, and borrowers and develop a program that
would ensure consistent and equitable treatment of borrowers by multiple
servicers. In contrast, servicers told us they had greater flexibility with
respect to the types of borrowers and conditions under which they could
offer proprietary modifications. First, several servicers told us their

41

HOPE NOW is an alliance between counselors, mortgage companies, investors, and other
mortgage market participants. According to its December 2010 Industry Extrapolations and
Metrics report, HOPE NOW estimates the survey covers 88 percent of the industry market.

Page 41

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

proprietary modification programs had fewer documentation
requirements. According to HAMP guidelines, borrowers must submit all
required documentation in order to be evaluated for and offered a HAMP
modification, including a Request for Modification and Affidavit, a tax
form, documentation to support income, and a Dodd-Frank Certification
form. 42 While Treasury has taken steps to streamline documentation
requirements in the past, both Treasury and servicers acknowledge that
borrowers’ failure to submit required documentation was one of the
primary reasons for being denied or canceled from a HAMP trial
modification. However, a servicer can offer a proprietary modification
even if the borrower lacked all of the required documentation. For
example, one servicer told us that if a borrower who was required to
submit 10 documents for a proprietary modification submitted only 6, the
servicer could still offer a modification if the 6 documents provided
sufficient information.
Second, several servicers told us they were able to offer more proprietary
modifications than HAMP modifications or help borrowers whom HAMP
cannot, because their proprietary modifications had fewer eligibility
requirements, such as restrictions on occupancy type. Treasury
announced early on that the HAMP program was not designed to help all
borrowers, such as those with investment properties and second homes.
For a borrower to be eligible for a modification under HAMP, the property
must be owner occupied, and according to Treasury’s HAMP data, through
September 2010, servicers have denied roughly 63,000 HAMP applicants (7
percent) who they said failed to meet this requirement. 43 But all six
servicers who provided us with information offered proprietary
modification programs without this restriction, allowing them to reach
borrowers who were ineligible for HAMP. One servicer we spoke with
noted that it had a large portfolio of investment properties that do not
meet the eligibility requirements for a HAMP modification.
In addition, while HAMP guidelines require borrowers to have a front-end
DTI above 31 percent, all of the servicers we spoke with indicated their
proprietary modification programs also served borrowers who had front-

42

The Dodd-Frank Certification form certifies that the borrower has not been convicted of a
financial felony, such as felony larceny, money laundering, or tax evasion, within the past
10 years.

43

Because Treasury does not report whether borrowers had GSE or non-GSE loans for all
borrowers denied HAMP, this percentage includes both GSE and non-GSE HAMP denials.

Page 42

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

end DTIs below 31 percent. The servicers explained that even with low
DTIs many of these borrowers were still unable to make their mortgage
payments because they had high levels of back-end debt, such as credit
card balances and car loans. We previously reported that HAMP requires
borrowers with high total household debt levels (postmodification DTI
ratios greater than 55 percent) to agree to obtain counseling, but it does
not require documentation that they actually received this counseling. 44
We continue to believe that it is important that Treasury determine
whether borrowers are receiving this counseling and whether the
counseling requirement is having its intended effect of limiting redefaults,
as we recommended. When asked about the differences between effective
proprietary modifications and HAMP modifications, roughly 63 percent of
housing counselors who responded to this question on our Web-based
survey ranked the ability of proprietary modifications to reach borrowers
with DTIs less than 31 percent as one of the main differences. According
to Treasury’s HAMP data, through September 2010, roughly 215,000
borrowers (24 percent) who were denied HAMP were denied because they
had a front-end DTI of less than 31 percent. Almost all of the servicers we
received information from indicated that the eligibility requirements for
their proprietary modification programs allowed mortgage balances that
exceeded HAMP limits. 45 One servicer noted that the majority of its
portfolio comprised super-jumbo loans, many of which fell outside the
HAMP mortgage balance limits. Roughly 106,000 borrowers (12 percent)
who were denied HAMP trial modifications through September 2010 were
denied because of ineligible mortgages. Fifty-two percent of housing
counselors also identified higher mortgage balance limits as another key
difference between proprietary modifications and HAMP modifications.
Lastly, the servicers we received information from offered proprietary
modifications with more flexible terms than HAMP modifications and
could more easily be adapted to the circumstances of individual
borrowers. HAMP guidelines require servicers to modify the terms of a
mortgage through interest-rate reductions, term extensions, and other
steps to bring the borrower’s front-end DTI ratio down to 31 percent (see
table 2). Several of the proprietary modification programs we reviewed
had variable target housing ratios—with one going down to 24 percent—

44

GAO-09-837.

45

For a one-unit property, the unpaid principal balance limit to be eligible for the HAMP
program is $729,750; for a two-unit, $934,200; for a three-unit, $1,129,250; for a four-unit,
$1,403,400.

Page 43

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

allowing servicers to bring a borrower’s payment down to a more
affordable level for some borrowers. In addition, for a servicer to be
required to offer a borrower a HAMP modification, HAMP requires the
borrower to pass the NPV test with a front-end DTI ratio of 31 percent.
However, some borrowers may fail the test at this level but would be able
to pass with a higher DTI ratio—for example, at 38 percent. These
borrowers may not be able to receive a HAMP modification, even though a
DTI ratio of 38 percent may have been more affordable than their current
mortgage payment. Some borrowers who are denied a HAMP modification
due to a negative NPV result but have a positive NPV result with a higher
front-end DTI may be offered a proprietary modification. For example, one
servicer plans to use variable front-end DTI thresholds to bring borrowers’
DTI ratios into more affordable ranges. The servicer will calculate
borrowers with front-end DTI ratios greater than 31 percent based on 31
percent, 35 percent, and 38 percent thresholds, and borrowers with frontend DTI ratios less than 31 percent could be brought down to a DTI as low
as 24 percent if they pass the NPV test at this level. The servicer estimates
that of 3,370 borrowers who were denied a HAMP trial modification
because their front-end DTI was already below 31 percent or as a result of
a negative NPV, 2,415 would pass the NPV test using the flexible front-end
DTI ratio thresholds and could receive a proprietary modification.
Table 2: Terms of Selected Proprietary Modification Programs Compared to HAMP
Proprietary
Modification
Program 2

Proprietary
Modification
Program 3

Proprietary
Modification
Program 4

HAMP

Proprietary
Modification
Program 1

Target DTI ratio

31%

None

31-42%

31-38%

24-38%

Interest rate floor

2%

2%

2%

1%

2%

Term extension

Up to 40 years

Up to 50 years

Up to 40 years

Up to 40 years

Up to 40 years

Principal forbearance

Yes

Yes

Yes

Yes

Yes

Principal forgiveness allowed

Yes

No

No

Yes

No

Duration of reduced interest rate 5 years
a
until raised to cap

Up to 5 years

3 years

Up to 5 years

5 years

Trial period

3 months

None

3 months

3 payments

3 months

Net spendable income per
month limit

None

10%, or minimum
of $250, maximum
of $1,000

None

At least $900 and
$200/dependent

None

Source: GAO analysis of documentation received from servicers.
a

After this length of time, the reduced interest rate under HAMP and each of these proprietary
modification programs may step up, or incrementally increase, to a maximum interest rate.

Page 44

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

In addition, having the flexibility to bring borrowers’ front-end DTI ratios
to below 31 percent allows servicers to account for borrowers’ back-end
DTI ratios when offering proprietary modifications. Several of the
servicers we spoke with had proprietary modification programs that
considered borrowers’ overall affordability, or ability to pay, when
modifying a mortgage, and the servicers calculated affordability
differently. For example, one servicer addressed overall affordability by
using a net spendable income calculation to determine a borrower’s
monthly mortgage payment. According to the servicer, its net spendable
income calculation factors in all of the borrower’s income and deducts all
expenses, including credit cards and utility bills. This proprietary
modification program was designed to leave the borrower with
approximately 10 percent of net spendable income, with a minimum of
$250 and a maximum of $1,000. Another servicer reported using family size
to determine affordability. The servicer indicated that it calculated
borrowers’ monthly payments based on the nature of the borrowers’
hardship, their current financial situation, and their change in
circumstances, as well as a postmodification monthly net disposable
income of $600 and an additional $100 per dependent. By incorporating
family size, this proprietary modification program may be able to help
some borrowers who may otherwise not qualify for HAMP.
Because servicers had a variety of proprietary modification programs that
calculated affordability in a number of ways, and because their loan
portfolios differed, the changes in mortgage terms as a result of
proprietary modifications varied across servicers. According to data we
received from six servicers, roughly 655,000 borrowers had permanent
proprietary modifications as of August 31, 2010. These borrowers had their
interest rate reduced by an average of 2.35 to 3.87 percentage points,
depending on the servicer. In addition, the amount of term extension
varied by each servicer. Specifically, servicers extended mortgage terms
by an average of 87 to 178 months for borrowers who had permanent
proprietary modifications. Lastly, servicers forbore varying amounts of
principal, ranging from an average of $33,971 to $116,488, or 16 percent to
60 percent of the unpaid principal balance prior to modification. 46

46
One servicer was unable to report the amount of principal forbearance for roughly 99
percent of those borrowers who had proprietary modifications and, therefore, is not
included in this range.

Page 45

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

The Sustainability of Both
HAMP and Proprietary
Modifications Remain
Unclear

While the number of proprietary modifications has outpaced the number
of HAMP modifications, the sustainability of both types of modifications is
still unclear. HAMP redefault rates have been relatively low to date, but it
is likely too soon to draw conclusions about HAMP redefaults. While data
on the redefault rates of HAMP and proprietary modifications are limited,
the Office of the Comptroller of the Currency (OCC) and Office of Thrift
Supervision (OTS) reported that 11 percent of HAMP modifications and 22
percent of proprietary modifications that started in the fourth quarter of
2009 were 60 or more days delinquent after 6 months. 47 In addition, one
servicer reported the redefault rates for its proprietary modifications were
26 percent at 6 months and roughly 40 percent at 12 months after the loan
was modified, while another servicer reported redefault rates of 32
percent at 6 months and 51 percent at 12 months.
Proprietary modifications may not reduce monthly mortgage payments as
much as HAMP modifications, potentially affecting the ability of
borrowers to maintain their modified payments. According to OCC and
OTS, during the third quarter of 2010, proprietary modifications reduced
monthly mortgage payments by an average of $332 per month, while
HAMP modifications reduced them by an average of $585 per month.
According to our analysis of Treasury’s HAMP data, borrowers who had a
GSE or non-GSE HAMP permanent modification as of September 30, 2010,
had their payments reduced by an average of $632, or 33 percent of the
average payment before modification. According to the data we received
from six servicers, for GSE and non-GSE loans, borrowers with a
permanent proprietary modification as of August 31, 2010, had their
monthly mortgage payments reduced from an average of $100 to $691 per
month, or 7 to 30 percent of the average monthly payment before
modification. In response to our survey, housing counselors provided
several examples of borrowers who had received proprietary
modifications that did not substantially reduce monthly mortgage
payments and that, in some cases, increased payments.
As we have seen, the extent to which modifications reduce monthly
mortgage payments may correlate with the ability of borrowers to

47

The OCC and OTS publish a quarterly mortgage metrics report that includes data on firstlien residential mortgages serviced by national banks and thrifts, focusing on credit
performance, loss-mitigation efforts, and foreclosures. OCC and OTS collect these data
from the eight national banks and one thrift with the largest mortgage-servicing portfolios
among national banks and thrifts. The data represent 64 percent of all first-lien residential
mortgages outstanding in the country.

Page 46

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

maintain modified payments. Specifically, OCC and OTS reported that
modifications made in 2010 that reduced monthly mortgage payments by
20 percent or more resulted in a redefault rate of 12 percent 6 months after
modification compared with 28 percent for modifications that reduced
payments by 10 percent or less. However, servicers have told us their
proprietary modification programs can serve borrowers with front-end
DTIs below 31 percent—borrowers who would be ineligible for a HAMP
modification. As a result, the average percentage monthly reduction for
these borrowers may not be as high as it would be for those with a HAMP
modification, because their premodification front-end DTI ratios were
lower than those of borrowers who received a HAMP modification. Going
forward, it will be important for Treasury to monitor redefault rates and
understand how they differ across servicers and modification terms. We
will also be looking at the redefault rates of HAMP and non-HAMP
modifications, as well as the effectiveness of other foreclosure mitigation
efforts, as part of our ongoing work looking at the broader federal
response to the foreclosure crisis.

Conclusions

HAMP and the newer MHA programs were part of an unprecedented
response to a particularly difficult time for our nation’s mortgage markets.
However, 2 years after Treasury first announced that it would use $50
billion in TARP funds for various programs intended to preserve
homeownership and protect home values, foreclosure rates remain at
historically high levels. While Treasury originally estimated that 3 to 4
million people would be helped by these programs, only 550,000
borrowers had received permanent HAMP first-lien modifications as of
November 30, 2010, and the number of borrowers starting trial
modifications has been rapidly declining since October 2009. Moreover,
Treasury has experienced challenges in implementing its other TARPfunded housing initiatives. In particular, the 2MP program, which Treasury
has stated is needed to create a comprehensive solution for borrowers
struggling to make their mortgage payments, has had a slow start.
According to six large MHA servicers, they have faced difficulties—
matching errors and omissions—in using the database required for
identifying second liens eligible for modification under the program. As a
result, servicers told us that relatively few second liens had been modified
as of August 2010, a year after program guidelines were first issued.
Treasury has taken some steps to address the issues that have slowed
down implementation of the program, but more could be done to inform
potentially eligible borrowers about 2MP. Specifically, borrowers whose
second liens may be eligible for modification under 2MP may not be aware
of the program or of any errors in the matching process, as servicers are

Page 47

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

not required to inform borrowers receiving HAMP first-lien modifications
that they could also be eligible for 2MP. Consequently, missed matches of
first and second liens could go undetected, and some borrowers who were
eligible for but not helped by the program are less able to keep up the
payments on their first-lien HAMP modifications.
HAFA and PRA, two other key components of Treasury’s TARP-funded
homeownership preservation effort, have also had slow starts. In fact,
servicers we spoke with did not expect HAFA to increase the overall
number of short sales performed, primarily due to extensive program
requirements that lengthen the time frames associated with a short sale
under the program. While Treasury has recently revised its HAFA program
requirements to allow servicers to bypass the HAMP first-lien program
eligibility review for borrowers interested solely in participating in HAFA
and relaxed other HAFA program requirements, it remains unclear the
extent to which these changes will result in greater program participation.
Additionally, because of the voluntary nature of the PRA program and
concerns over the lack of program transparency, including the level of
public reporting that will be available at the servicer level, it remains
unclear how many borrowers will receive principal reductions under PRA.
Treasury has stated that it will report on PRA activity when data are
available, and we continue to believe that it will be important that this
reporting includes the extent to which servicers determined that principal
reduction was beneficial to investors but did not offer it, as we
recommended in June 2010. If HAFA and PRA do not result in increased
program participation, Treasury’s efforts to combat the negative effects
associated with avoidable foreclosures will be compromised, potentially
limiting the ability of Treasury efforts to preserve homeownership and
protect home values.
Further, Treasury could do more to apply lessons learned from its
experience in implementing early HAMP programs to its more recent
initiatives. We reported in June 2010 that the implementation of other
TARP-funded homeownership preservation programs could benefit from
lessons learned in the initial stages of HAMP implementation. Specifically,
we noted that it would be important for Treasury to expeditiously develop
and implement new programs while also developing sufficient program
planning and implementation capacity, meaningful performance measures
and remedies, and appropriate risk assessments in accordance with
standards for effective program management. Already, 2MP, HAFA, and
PRA have undergone several revisions, and servicers cited changing
guidelines and short implementation periods as significant challenges in
fully implementing the programs. In July 2009, we recommended that

Page 48

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Treasury place a high priority on fully staffing vacancies in the
Homeownership Preservation Office (HPO) and evaluating staffing levels
and competencies. As of January 2011, Treasury has filled key positions in
HPO, but has not conducted a formal assessment of its staffing levels
despite the implementation of the newer programs. We continue to believe
that it is essential that Treasury ensure that it has enough staff with the
appropriate skills to govern TARP-funded housing programs effectively.
While Treasury has conducted reviews of the readiness of servicers
participating in 2MP, HAFA, and PRA to successfully implement the
programs, a large majority of servicers did not provide all documentation
required to demonstrate that the key tasks needed to support these
programs were in place. It is imperative that Treasury take swift action to
ensure that servicers have the ability to implement these programs since,
as we have seen with the slow progress of the HAMP first-lien
modification program, the success of these TARP-funded initiatives will be
largely driven by the capacity and willingness of servicers to implement
these programs in an expeditious and effective manner. In addition,
Treasury has not developed program-specific performance measures
against which to measure these programs’ success and has not specified
the remedies it will take if servicers are not meeting performance
standards. Without specific program measures and remedies, Treasury will
not be able to effectively assess the outcomes of these programs and hold
servicers accountable for performance goals. We continue to believe that
it is important for Treasury to develop such performance measures and
clear goals for them, as we have recommended.
Treasury requires servicers to submit data on borrowers who have been
evaluated for HAMP, and these data provide important information and
insights on the characteristics of borrowers who are in trial and permanent
HAMP modifications, who have been canceled from trial modifications, and
who have redefaulted from permanent modifications. However, Treasury’s
HAMP database also contains inaccurate or missing information on certain
key variables, including LTV ratios and borrowers’ race and ethnicity.
Treasury has stated that it is working to improve the quality of its data, and
it will be important that the agency do so expeditiously. Complete and
accurate information is important for Treasury to fully understand the
characteristics of borrowers who HAMP has been unable to help or
determine program compliance. Moreover, this information is important to
identify what additional steps or adjustments that could be made to existing
TARP-funded programs to better achieve the mandated goals of preserving
homeownership and protecting property values.

Page 49

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Finally, while Treasury has begun publicly reporting the outcomes for
borrowers who have been denied or canceled from HAMP trial
modifications, its reporting practices make it difficult to determine the
extent to which these borrowers are helped by non-HAMP (proprietary)
loan modifications. For example, data we collected from six large MHA
servicers showed that only 18 percent of borrowers canceled from a
HAMP trial modification had received a proprietary modification and an
additional 23 percent had a proprietary modification pending. However,
Treasury reported that 43 percent of these borrowers were in the process
of receiving a proprietary modification with those same six servicers.
Furthermore, Treasury’s system for reporting outcomes requires servicers
to place borrowers in only one category even when borrowers are being
evaluated for several possible outcomes, with proprietary modifications
reported first. As a result, the proportion of borrowers with proprietary
modifications is likely overstated relative to other possible outcomes, such
as foreclosure starts. Without accurate reporting of borrower outcomes,
Treasury cannot know the actual extent to which borrowers who are
denied, canceled, or redefaulted from HAMP are helped by other programs
or evaluate the need for further action to assist this group of homeowners.

Recommendations for
Executive Action

As part of its efforts to continue improving the transparency and
accountability of MHA, we recommend that the Secretary of the Treasury
take actions to

•

require servicers to advise borrowers to notify their second-lien servicers
once a first lien has been modified under HAMP to reduce the risk that
borrowers with modified first liens are not captured in the LPS matching
database and, therefore, are not offered second-lien modifications;

•

ensure that servicers demonstrate they have the operational capacity and
infrastructure in place to successfully implement the requirements of the
2MP, HAFA, and PRA programs; and

•

consider methods for better capturing outcomes for borrowers who are
denied, canceled, or redefaulted from HAMP, including more accurately
reflecting what actions are completed or pending and allowing for the
reporting of multiple concurrent outcomes, in order to determine whether
borrowers are receiving effective assistance outside of HAMP and whether
additional actions may be needed to assist them.

Page 50

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Agency Comments
and Our Evaluation

We provided a draft of this report to Treasury for its review and comment,
and we received written comments from the Acting Assistant Secretary for
Financial Stability that are reprinted in appendix III. We also received
technical comments from Treasury that we incorporated into the report as
appropriate. In its written comments, Treasury stated that it appreciated
our efforts in assessing the housing programs initiated under its TARP
program and acknowledged the draft report’s description of the
operational capacity and infrastructure challenges faced by servicers in
implementing Treasury’s housing programs. In addition, Treasury noted
that our research in proprietary modifications made by servicers outside
of MHA was useful. However, Treasury stated that it believed that the draft
report raised certain criticisms regarding the design and implementation
of MHA that were unwarranted.
First, Treasury stated that the draft report criticized Treasury for the
number of changes made to its housing programs following their
implementation, and its alleged failure to incorporate the lessons it
learned from the first-lien HAMP program into the roll out and design of
other MHA programs, such as HAFA. Treasury stated that the report
should acknowledge the circumstances under which the programs were
first implemented. In response, we added some additional language
recognizing that HAMP and the newer MHA programs were part of an
unprecedented response to a particularly difficult time for our nation’s
mortgage markets. However, servicers we spoke with noted that ongoing
changes to guidelines have presented challenges such as needing to
update internal servicing systems and retrain staff which, in some cases,
delayed program implementation. In addition, as noted in the draft report,
Treasury has repeated some of the practices that were the focus of
previous recommendations we had made for the first-lien program in its
implementation of its newer MHA programs. For example, in our July 2009
report, we found that Treasury had not developed a means of
systematically assessing servicers’ capacity to meet program requirements
during program admission, and we recommended further action in this
area to increase the likelihood of success of the program. In our review of
the newer MHA programs, we also found that Treasury had not fully
ensured that servicers had the capacity to successfully implement these
programs. We continue to believe that such action is needed to better
ensure the likelihood of success of these newer MHA programs.
Second, Treasury raised concerns about the draft report’s comparison of
HAMP modifications to proprietary modifications. Treasury noted that it
did not believe that it was constructive to assess HAMP’s performance
based on the goals of proprietary programs that are not government

Page 51

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

supported. We have added some additional language to the report to
provide additional context to the report’s discussion of proprietary
modifications. The purpose of this report was not to assess the
performance of HAMP modifications based on the goals of proprietary
modifications. Instead, the draft report provided a description of
proprietary modifications and some of the ways that they differ from
HAMP modifications. It does not suggest that the objective of HAMP
modifications and proprietary modifications are or should be the same,
particularly given Treasury’s responsibility to safeguard taxpayer dollars
under HAMP. As noted by Treasury in its comment letter, there is little
available information about these proprietary modifications, and the more
that is known about their terms and outcomes, the easier is will be for
policymakers and regulators to craft appropriate changes to MHA and
other housing programs aimed at preventing avoidable foreclosures.
Third, Treasury noted that the draft report criticized the completeness and
quality of the data collected by Treasury related to HAMP modifications,
and that it disagreed with the conclusion that missing or inaccurate
information limits Treasury’s ability to identify program gaps. Treasury
noted that it relies on data provided by the borrowers to the servicers and
it has improved significantly over the past 6 months, especially as the
program moved to verified income. Treasury stated that the data on
permanent modifications is robust, allowing Treasury to determine gaps in
programs and how to make improvements. In the draft report, we
acknowledged that Treasury is working with Fannie Mae to improve the
data and, particularly with borrower race and ethnicity information, the
data has improved over time. However, it is equally important that
Treasury obtain complete and accurate information on those who are
denied or canceled from a HAMP trial modification. Without such
information, Treasury cannot determine if servicers are implementing the
program fairly or whether additional actions may be necessary to address
the needs of borrowers who are denied or canceled from HAMP trial
modification. Going forward, it will be important for Treasury to continue
to improve the quality of its HAMP data as this information is important to
identify what additional steps or adjustments could be made to existing
TARP-funded housing programs to better achieve the mandated goals of
preserving homeownership and protecting property values.

Page 52

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

We are sending copies of this report to interested congressional
committees and members of the Congressional Oversight Panel, Financial
Stability Oversight Board, Special Inspector General for TARP, Treasury,
the federal banking regulators, and others. This report is also available at
no charge on the GAO Web site at http://www.gao.gov.
If you or your staffs have any questions about this report, please contact
me at (202) 512-8678 or sciremj@gao.gov. Contact points for our Offices of
Congressional Relations and Public Affairs may be found on the last page
of this report. GAO staff who made major contributions to this report are
listed in appendix IV.

Mathew J. Scirè
Director
Financial Markets and
Community Investment

Page 53

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

List of Congressional Committees
The Honorable Daniel K. Inouye
Chairman
The Honorable Thad Cochran
Vice Chairman
Committee on Appropriations
United States Senate
The Honorable Tim Johnson
Chairman
The Honorable Richard C. Shelby
Ranking Member
Committee on Banking, Housing, and Urban Affairs
United States Senate
The Honorable Kent Conrad
Chairman
The Honorable Jeff Sessions
Ranking Member
Committee on the Budget
United States Senate
The Honorable Max Baucus
Chairman
The Honorable Orrin G. Hatch
Ranking Member
Committee on Finance
United States Senate
The Honorable Harold Rogers
Chairman
The Honorable Norman D. Dicks
Ranking Member
Committee on Appropriations
House of Representatives
The Honorable Paul Ryan
Chairman
The Honorable Chris Van Hollen
Ranking Member
Committee on the Budget
House of Representatives

Page 54

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

The Honorable Spencer Bachus
Chairman
The Honorable Barney Frank
Ranking Member
Committee on Financial Services
House of Representatives
The Honorable Dave Camp
Chairman
The Honorable Sander Levin
Ranking Member
Committee on Ways and Means
House of Representatives

Page 55

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix I: Scope and Methodology

Appendix I: Scope and Methodology

To examine the status of the Department of Treasury’s (Treasury) secondlien modification, principal reduction, and foreclosure alternatives
programs and the design and implementation challenges Treasury and
servicers have faced with these programs to date, we spoke with and
obtained information from six large Making Homes Affordable (MHA)
servicers, including the four largest servicers participating in the SecondLien Modification Program (2MP) at the start of our review. These six
servicers were: American Home Mortgage Servicing, Inc.; Bank of
America; CitiMortgage; JP Morgan Chase Bank; OneWest Bank; and Wells
Fargo Bank. We determined these as six large MHA servicers based on the
amount of Troubled Asset Relief Program (TARP) funds they were
allocated for loan modification programs. These six servicers collectively
represented 72 percent of the TARP funds allocated to participating
servicers. For each of these six servicers, we reviewed their 2MP, Home
Affordable Foreclosure Alternatives (HAFA), and the Principal Reduction
Alternatives (PRA) guidance, policies, procedures, process flows, training
materials, and risk assessments, as applicable; and interviewed
management staff. We also reviewed 2MP, HAFA, and PRA documentation
issued by Treasury, including the supplemental directives related to 2MP,
HAFA, and PRA, readiness assessments of servicers, and reporting process
flows. We also spoke with officials at Treasury to understand the
challenges faced in implementing these programs and the steps taken by
Treasury to assess the capacity needs and risks of these programs, as well
as steps taken to measure the programs’ success. We spoke with trade
associations representing investors, mortgage insurers, servicers, and an
organization representing homeowners and community advocates. Finally,
we reviewed the Standards for Internal Control in the Federal
Government to determine the key elements needed to ensure program
stability and adequate program management.
To examine the characteristics of homeowners who the Home Affordable
Modification Program (HAMP) has been able to help, we obtained and
analyzed Treasury’s HAMP data in its system of record, Investor
Reporting/2 (IR/2), through September 30, 2010. We reviewed Treasury
guidelines on reporting requirements for HAMP, including the information
servicers are required to report for borrowers who were denied trial
modifications, and spoke with Treasury and Fannie Mae officials to
understand potential inconsistencies and gaps in the data. We determined
that the data was sufficiently reliable for our purposes. We also used the
data to perform an econometric analysis of factors that contribute to
borrowers’ likelihood of seeing their trial modifications canceled (see
appendix II for more details). We received and incorporated feedback on
our model from Treasury and others. To obtain housing counselors’ views

Page 56

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix I: Scope and Methodology

of borrowers’ experiences with HAMP, we conducted a Web-based survey
of housing counselors who received funding from NeighborWorks
America, a national nonprofit organization created by Congress to provide
foreclosure prevention and other community revitalization assistance to
the more than 230 community-based organizations in its network. We
received complete responses from 396 counselors. This report does not
contain all the results from the survey. The survey and a more complete
tabulation of the results will be part of an upcoming report.
Finally, to examine the outcomes for borrowers who were denied or fell
out of HAMP trial or permanent modifications, we reviewed HAMP
program documentation on policies for evaluating these borrowers for
other loss mitigation programs. We reviewed the outcomes of borrowers
who applied for but did not receive a HAMP trial modification or who had
a canceled HAMP trial modification as reported by Treasury in the
monthly MHA servicer performance reports. We obtained documentation
from Treasury and interviewed servicers to understand how Treasury
collects data on the outcomes of these borrowers. In addition, we obtained
data from the six large MHA servicers noted earlier in this appendix.
Specifically, we obtained and analyzed data on the outcomes of all
borrowers who were denied a HAMP trial modification, had a canceled
HAMP trial modification, or redefaulted on a HAMP permanent
modification; the number of proprietary modifications completed; and the
characteristics of the terms of these proprietary modifications. The
servicers provided the data between when they began participating in the
HAMP program and August 31, 2010, or the date in which they submitted
their August 2010 reporting to Treasury (e.g., September 6, 2010). 1
According to the data we received, the number of trial modifications
offered by these six servicers represented 72 percent of the total number
of trial modifications offered by all servicers as reported by Treasury
through September 2, 2010. We determined that these data were reliable
for the purposes of our report. To understand why servicers may offer
more proprietary modifications than HAMP modifications, we reviewed
data on the number of completed proprietary modifications published by
HOPE NOW, an alliance between counselors, mortgage companies,
investors, and other mortgage market participants. In addition, we

1
Treasury did not require servicers to report data on borrowers who were denied a HAMP
trial modification until December 1, 2009. As a result, three of the six servicers provided
the outcomes of borrowers who were denied a HAMP trial modification from December 1,
2009, through August 31, 2010, or the date they submitted their August 2010 reporting to
Treasury.

Page 57

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix I: Scope and Methodology

reviewed documentation on the terms and eligibility requirements of the
proprietary modification programs offered by the six servicers
participating in our review. In addition, we interviewed these servicers
about the features of their proprietary modification programs. Also,
through our Web-based survey of housing counselors, we received
responses on the differences between effective proprietary modifications
and HAMP modifications, as well as examples of effective and ineffective
proprietary modifications. Finally, to understand the sustainability of
HAMP and proprietary modifications, we reviewed data published by the
Office of the Comptroller of the Currency and the Office of Thrift
Supervision on the redefault rates and monthly payment reduction of
HAMP modifications, as well as data we collected from servicers on the
redefault rates, terms, and monthly payment reductions of their GSE and
non-GSE proprietary modifications.
We conducted this performance audit from July 2010 through March 2011
in accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain
sufficient, appropriate evidence to provide a reasonable basis for our
findings and conclusions based on our audit objectives. We believe that
the evidence obtained provides a reasonable basis for our findings and
conclusions based on the audit objectives.

Page 58

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations
Data and Model
Specification

To describe the characteristics of the borrowers and mortgages that have
been canceled from the trial modification, we used an econometric analysis
rather than present descriptive statistics since it allowed us to control for
the impacts of potential confounding factors, including differences across
servicers as well as default-risk differences among the borrowers that are
not observable (unobserved borrower heterogeneity). 1 Servicers who
participate in the Home Affordable Modification Program (HAMP) are
required to provide periodic loan level data to Fannie Mae in its capacity as
the administrator for the HAMP program. Specifically, servicers are required
to report data at the start of the trial modification period and during the
modification trial period, for loan set up of the approved modification, and
monthly after the modification is set up on Fannie Mae’s system. The data
used in our econometric analysis are for HAMP loans as of September 30,
2010. 2 The data have one record per loan, with information on the loan
status—whether the loan is denied for trial modification, has entered the
trial plan, and its outcome (i.e., converted to the permanent modification, or
still active in the trial plan, or has fallen out). We excluded loans that
entered the trial plan from July 2010 through September 2010 (which is the
end of the current data set) because not enough time had likely elapsed for
loans in this pool to have defaulted or been canceled.
Through September 30, 2010, several servicers have signed up for HAMP.
Seventeen of them, we categorized as “large” based on the Treasury reported
data on “estimated eligible loans 60 days or more delinquent,” have over 90
percent of the loans—the remainder of the servicers were grouped into the
“other” category.3 For the universe of 1,361,832 loans that had entered the
trial period plan as of September 30, 2010, the average cancellation rate was
50 percent. The sample we used for the regression analysis, based on data
availability, consists of 727,095 loans (53 percent of the universe), with an
average cancellation rate of 50 percent. The sample data exclude servicers
whose share of loans or fallout rates for the sample differed a lot from that of
the universe; there are 13 “large” and “other” servicers.

1

The econometric methodology and findings were reviewed by Professor Anthony
Pennington-Cross, Marquette University, Marquette, Wis., and Professor Hashem
Dezhbakhsh, Emory University, Atlanta, Ga.

2

The data used for the analysis were restricted to one-unit single family, principal
residence, owner-occupied housings located in the 50 states of the United States and the
District of Columbia.

3

Treasury, Making Home Affordable Program: Servicer Performance Report Through
September 2010 (Washington, D.C., October 2010).

Page 59

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

Following the literature, we used a reduced-form probability model to help
determine the effects of the characteristics of the borrower and mortgage
on HAMP trial loan cancellations. Accordingly, based on economic
reasoning, data availability, the HAMP guidelines, and previous studies on
loan performance, we use probabilistic models to estimate the likelihood
that a loan modified under the trial period plan does not convert to a
permanent modification. The dependent variable for the cancellations is
binary, which equals 1 if a loan that entered the trial-period plan did not
convert to permanent modification and 0 otherwise. The explanatory
variables that we include in the model are conditioned by the available
data (see table 3 for the list of the variables).
Table 3: Summary Statistics of Variables Used in Regressio
If trial modification canceled
Months since loan originationa
If credit score >=620
If back-end DTI>=55
Back-end DTI flagb
If mark-to-market LTV (MLTV) <= 80%
If MLTV between 80 & 100%
If MLTV between 100 & 120%
If MLTV between 120 & 140%
If MLTV >140%
MLTV flagc
If borrower’s payments are current
If 30 days delinquent
If 60 days delinquent
If 90 days or more delinquent
If trial length <= 4 months
If original loan had private mortgage insurance
If fixed rate at origination vs nonfixed
If no decrease in principal and interest (P&I) payment
If no decrease in principal and interest (P&I) payment
If P&I decreased between 0 & 10%
If P&I decreased between 10 & 20%
If P&I decreased more than 20%
If principal reduction between 1 & 50%
If principal forbearance between 1 & 50%
If private loan vs. GSE loan
If portfolio loan vs GSE loan
If stated income vs. verified incomed

Mean
0.1246
0.1205
0.2263
0.2352
0.1439
0.2741
0.0144
0.1627
0.0988
0.1056
0.6298
0.8567
0.2107
0.6196
0.0004
0.1021
0.1280
0.7694
0.0155
0.2131
0.3353
0.1160
0.8851

Median
0
29
0
1
0
0
0
0
0
0
0
0
0
0
1
1
0
1
0
0
0
1
0
0
0
0

Standard
Deviation
0.5000
20
0.4586
0.4776
0.3302
0.3255
0.4184
0.4241
0.3510
0.4461
0.1190
0.3691
0.2984
0.3073
0.4829
0.3504
0.4078
0.4855
0.0210
0.3029
0.3341
0.4212
0.1235
0.4095
0.4721
0.3202
1

Minimum
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Maximum
1
396
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

Source: GAO Analysis of Treasury data.

Page 60

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

Note: The total number of observations used for regression analysis is 727,095 loans.
a

The months since origination is the difference between origination date and January 1, 2009. A
condition for loan-modification eligibility is that the loan was originated between January 1, 1959 and
January 1, 2009.

b

The back-end debt-to-income (DTI) ratio after modification was replaced with the back-end DTI
before modification if the former was missing; this affected about 12% of the data (back-end DTI flag).
The regression estimates were similar when the flag was excluded.
c

The MLTV ratio was replaced with the loan-to-value (LTV) at origination if the former was missing—
this affected about 1% of the data (MLTV flag). The regression estimates were similar when the flag
was excluded.

d

For servicers who allowed borrowers to submit applications with stated incomes prior to
June 1, 2010.

We estimated cancellation rates using binomial logistic probability (logit)
models, an approach commonly used in economics to examine choices
and evaluate various events or outcomes. 4 The models included fixed
effects for the servicers, which allowed us to account for both the
observed and unobserved characteristics of the servicers. 5 We also
included state-level fixed effects to control for factors that vary across the
states but are the same within a state, such as the type of foreclosure laws
and other state policies on mortgages.

Econometric Results

The basic regression results from using the probability model, described
above and the data in table 3, are presented in table 4. Most of the
variables were statistically significant at the 5 percent level or better, and
typically consistent with our expectations as to the direction of their
impacts. We discuss below the key findings, based on statistically (and

4

We note that this approach is appropriate since we are primarily interested in the
probability that a trial loan falls out and not the hazard rate of the fallout (i.e., the
probability that a trial loan falls out at a certain time if it has not already fallen out up to
that time). Furthermore, the available data do not permit us to analyze the latter situation.
5

Since the servicers have more than one loan in HAMP, the fixed effects allow us to adjust
the standard errors for dependence among those loans, and also control for all stable
characteristics of the servicers, whether observed or unobserved. The fixed effects
technique uses the variation within the servicers to estimate the coefficients.

Page 61

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

economically6) significant changes in the likelihood of cancellation, using
the estimated marginal effects of the explanatory variables.7
Table 4: Probabilistic Estimates of HAMP Trial Loan Modification Cancellation Rates
Variable
Intercept
Months since origination
If credit score >= 620
If back-end DTI >= 55%
Back-end DTI flag
If MLTV between 80 & 100% vs. MLTV <= 80%
If MLTV between 100 & 120% vs. MLTV <= 80%
If MLTV between 120 and 140% vs.MLTV <= 80%
If MLTV >140 vs. MLTV <= 80%
MLTV flag
If 30 days delinquent vs. current
If 60 days delinquent vs. current
If 90 days or more delinquent vs. current
If trial length <= 4 months
If original loan had PMI
If original loan had fixed rate
If payment decrease between 10 & 20% vs. <= 10%
If payment decrease > 20% vs. <= 10%
If principal reduction between 1 & 50%
If principal forbearance between 1% & 50%
If private loan vs. GSE loan
If portfolio loan vs. GSE loan
If stated income vs verified income
Servicer 1 vs. other servicers
Servicer 2 vs. other servicers
Servicer 3 vs. other servicers
Servicer 4 vs. other servicers
Servicer 5 vs. other servicers
Servicer 6 vs. other servicers
Servicer 7 vs. other servicers
Servicer 8 vs. other servicers

Marg. effect
-1.2516
0.0006
0.0066
-0.0229
0.5749
0.0071
-0.0261
-0.0673
-0.0765
-0.1129
0.0043
0.0626
0.0921
0.5803
0.0100
0.0046
-0.0539
-0.0543
-0.0564
-0.0298
0.0406
-0.0546
0.5189
-0.1772
0.1979
0.2263
0.0350
0.4284
0.3010
0.4173
0.1130

Estimate
-8.0280
0.00375
0.0426
-0.1468
3.6875
0.0458
-0.1675
-0.4318
-0.4906
-0.7242
0.0274
0.4017
0.5909
3.7218
0.0640
0.0294
-0.3460
-0.3481
-0.3617
-0.1912
0.2607
-0.3501
3.3285
-1.1363
1.2691
1.4516
0.2242
2.7479
1.9305
2.6766
0.7245

Standard error
0.1383
0.000171
0.00717
0.00711
0.0176
0.0114
0.0118
0.0131
0.0126
0.0319
0.0129
0.0120
0.00914
0.0166
0.00804
0.00715
0.0126
0.0104
0.0416
0.00766
0.00850
0.0115
0.0619
0.0746
0.0206
0.0121
0.0152
0.0653
0.0196
0.0654
0.0188

Pr >ChiSq
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
0.0339
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001

Odds ratio
0.000
1.004
1.044
0.863
39.946
1.047
0.846
0.649
0.612
0.485
1.028
1.494
1.806
41.340
1.066
1.030
0.708
0.706
0.696
0.826
1.298
0.705
27.896
0.321
3.558
4.270
1.251
15.609
6.893
14.535
2.064

6

An economically significant result implies the impact is large and meaningful. We use a
threshold of 5 percent marginal effect.
7
The odds ratios of the estimates are also reported. An odds ratio close to one means the
variable is neither more nor less likely to influence fallouts. Deviations from one indicate
the direction and strength of the effects. An odds ratio greater than one means the factor is
more likely to impact fallouts; similarly, an odds ratio less than one means the factor is less
likely to impact fallouts.

Page 62

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

Variable
Servicer 9 vs. other servicers
Servicer 10 vs. other servicers
Servicer 11 vs. other servicers
Servicer 12 vs. other servicers
Servicer 13 vs. other servicers

Marg. effect
0.3404
0.0835
0.4570
0.0900
0.2669

Estimate
2.1834
0.5358
2.9311
0.5770
1.7121

Standard error
0.0202
0.0194
0.0712
0.0873
0.0136

Pr >ChiSq
<.0001
<.0001
<.0001
<.0001
<.0001

Odds ratio
8.876
1.709
18.748
1.781
5.540

Source: GAO Analysis of Treasury data.

Note: Likelihood Ratio Test: Chi-square = 326594, df = 81, Prob Chi-Square < 0.0001. Area under
Receiver Operating Characteristic (ROC) curve, c-statistics = 0.854. Pseudo R-square = 0.3618,
Max-rescaled R-square = 0.4825. The estimates for the states are not reported.

•

Stated income. Loans that entered the trial plan using stated income
documentation were 52 percent more likely to be canceled, compared to
verified income. 8 This effect was consistent with expectations since these
borrowers are likely not to able to provide verified documentation when
requested.

•

Trial length. Trial periods that last for 4 months or less were about 58
percent more likely to be canceled compared to those that stay in the trial
plan for a longer term. A longer stay in the trial plan implies the borrower’s
payments are current and, therefore, less likely to be canceled. This result
is generally consistent with the hazard models of mortgage performance
that indicate that loans that are likely to default do so much earlier than
later.

•

Delinquency status. Borrowers who were 60 days or 90 days or more past
due on their mortgages before the trial-period plan, compared to
borrowers who were current, were 6 and 9 percent more likely to have
their loans canceled, respectively; thus, the longer the delinquency status
the more likely the cancellation. This effect is consistent with
expectations.

•

Payment Reduction. The reduction in payment generally results from
interest rate reduction and extension of the loan term. Loans that receive
reductions in payments (of principal and interest) of more than 10 percent,
compared to reductions that are less than 10 percent or less (which
include no reductions and increases in payments), were 5 percent less
likely to be canceled. This result is expected since the payment reductions
increase the affordability of the mortgage, a key objective of HAMP.

8

These are for loans where, prior to June 1, 2010, a servicer allowed borrowers to provide
stated income documentation to start the trial plan.

Page 63

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

•

MLTV ratios. Loans with an LTV between 120 and 140 percent were 7
percent less likely to be canceled, while loans with an LTV of more than
140 percent were 8 percent less likely to be canceled, compared to those
with MLTV 80 percent or less. This effect is contrary to expectation. The
reason for this outcome is while borrowers with high MLTV were more
likely to have their trial modifications canceled due to not making their
payments, they were disproportionately less likely to have their trial
modifications canceled because of insufficient documentation, compared
to those with MLTV at or below 80 percent.

•

Principal reductions. Loans that received principal reductions in the form
of principal forgiveness of between 1 and 50 percent of their total loan
balance were 6 percent less likely to be canceled compared with those
that did not receive principal forgiveness. We note that only about 2
percent of the loans have received principal forgiveness. 9

•

Servicer effects. We estimated the changes in the likelihood of cancellation
for the servicers using the marginal effects in table 4. To examine the
extent to which there is variation in the likelihood of cancellation across
servicers, we defined three distinct borrower profiles and calculated the
likelihood of cancellation for each of these borrower profiles for each
servicer. The “typical” borrower profile used mean values for the borrower
population; the “current” borrower profile used mean values for all
characteristics except that the borrower was assumed to be current (less
than 30 days delinquent); and the “delinquent” profile used mean values
for all characteristics except that the borrower was assumed to be
delinquent by 90 days or more. Because delinquency status predicts higher
likelihood of cancellations for borrowers who are seriously delinquent (90
days or more delinquent) compared to being current (less than 30 days
delinquent), the likelihood of cancellation increases with increased
delinquency for each servicer.
The results presented in fig. 9 show significant variation across the
servicers for cancellations of trial modifications. In particular, for the large
servicers, the likelihood of cancellation increased for about one-half of
them (ranging from 1 to 24 percent) but decreased for the other half
(ranging from -2 to -39 percent) for the “typical” borrower. Although the
major reasons for the cancellations vary across the servicers, they were

9

Loans that received principal forbearance were 3 percent less likely to be canceled, and
about 20 percent of the loans received this form of principal reduction.

Page 64

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

primarily due to incomplete documentation, trial plan default, and
ineligible mortgage.
Figure 9: Estimated Change in Likelihood of Cancellation of HAMP Trial Loan
Modification by Servicer, Delinquency Status Before Modification
Percentage
30
20
10
0
-10
-20
-30
-40
-50
1
Other
Servicer

4

10

12

8

Avg.

2

3

13

6

9

7

5

11

Current borrowers
Typical borrowers
90 days delinquent borrowers
Source: GAO.

Note: “OTH” means the group of non-large servicers; the numbered servicers refer to “large”
servicers, and “AVG” is the average values for all the servicers in the sample. The “typical” borrower
profile is based on the mean values for the borrower population; the “current” borrower profile used
mean values for all characteristics except that the borrower was assumed to be less than 30 days
delinquent; and the “90DQ” profile used mean values for all characteristics except that the borrower
was assumed to be delinquent by 90 days or more, and the property is assumed to be in California.
Assuming other states will change the probabilities but not the pattern of the likelihoods across the
servicers. The changes in likelihoods of cancellation of trial modifications for each large servicer are
compared to the “OTH” servicers.

•

State-level effects. For the state-level effects we estimated the change in
the likelihood of trial cancellations across the states using the marginal
effects in table 4, similar to the analysis of the servicer effects. The results
presented in fig.10 show that the changes in the likelihoods of
cancellations are higher in most of the states, including high mortgage
foreclosure states such as Arizona, California, Florida, Michigan, and
Nevada—which have over 40 percent of the trial loans.

Page 65

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

Figure 10: Estimated Change in Likelihood of Cancellation of HAMP Trial Loan Modification by State
Percentage
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
0

ME OH NH MA WI DE CT CO MN RI UT WY HI IL PA IN NE WA NJ KY ID MD MS NY VT OR SC TN MT AL VA MI LA WV ND CA MO NC AZ NM SD NV IA OK KS GA AR TX FL DC AK

-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
Source: GAO.

Note: The changes in likelihoods of cancellation of trial modifications for each state are compared to
the state of Wyoming, the base state used to obtain the regression estimates.

Several checks were conducted to ensure that our results are reliable—
including the sample used, model specification, and estimation
techniques. 10 In all cases, our key results were generally unchanged.
Specifically, we excluded the servicer effects and state-level effects, and
included the start time of the trial to account for the housing and
economic conditions at the time of the modification. 11 We also estimated
robust standard errors to ensure that the tests of significance are reliable.
Furthermore, although we could not use the fixed-effects technique to
control for unobserved heterogeneity across the borrowers because we do
not have repeat observations on the borrowers, we attempted to
incorporate unobserved heterogeneity among the borrowers using mixed
multinomial logit estimation. This is intended to help capture the

10

The key results were generally unchanged when we used all the trial loans, including the
newer trial modifications from July 2010 to September 2010.

11

We used a variable for the month of the year when the trial started to capture the
contemporaneous economic conditions when the loan entered the trial plan. In order to
obtain valid estimates, since both the start time for the trial and the stated income
variable—which are both time-related—overlap, we excluded the stated-income variable
when the start-time variable was included.

Page 66

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix II: Description of GAO’s
Econometric Analysis of HAMP Trial Loans
Modifications Cancellations

differential in risk preferences and idiosyncratic differences among the
borrowers that are not captured by the explanatory variables in the
models we estimated. 12 However, we could not estimate the mass point
locations with any precision. Finally, we noted that since loans enter and
exit HAMP over time, these results may not necessary pertain in the
future.

12

See Anthony Pennington-Cross, “The Duration of Foreclosures in the Subprime Mortgage
Market: A Competing Risk Model with Mixing,” Journal of Real Estate Finance and
Economics, vol. 40:109–129, 2010 for a discussion of this approach.

Page 67

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

Appendix III: Comments from the
Department of the Treasury
Note: GAO comments
supplementing those in
the report text appear
at the end of this
appendix.

Page 68

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

Page 69

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

See comment 1.

See comment 2.

See comment 3.

Page 70

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

See comment 4.

See comment 5.

See comment 6.

See comment 7.

Page 71

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

See comment 8.

See comment 9.

Page 72

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

The following are GAO’s comments to the Department of the Treasury’s
letter dated February 23, 2011.

GAO Comments

1. We acknowledge that Treasury’s HAMP program is part of an
unprecedented response to a particularly difficult time in our nation’s
mortgage markets. As noted in the report, we also acknowledge that
Treasury took steps to consult with servicers on the design and
implementation of 2MP, HAFA, and PRA. However, Treasury has
repeated some of the practices that were the focus of previous
recommendations we had made for the first-lien program in its
implementation of its newer MHA programs. For example, in July 2009,
we recommended that Treasury develop a means of systematically
assessing servicers’ capacity to meet program requirements during
program admission. While Treasury has begun assessing servicers’
capacity to implement 2MP, HAFA, and PRA, it has not ensured that
servicers have sufficiently demonstrated they have the capacity to
successfully and expeditiously implement these programs. In addition,
we recommended in June 2010 that Treasury finalize and implement
benchmarks for performance measures under the first-lien
modification program, as well as develop such measures for the newly
announced programs. Treasury has not developed these benchmarks,
either for the first-lien or the subsequent programs, making it difficult
to hold servicers accountable for performance and assess the extent to
which they have been successful. The pages referenced in this
comment are now pages 20 to 24.
2. Treasury indicated that the draft report criticized it for a lack of a
database that includes second liens matched to HAMP-modified first
liens. The draft report does not criticize Treasury for the lack of a
database. Rather, the report notes that Treasury worked with LPS to
develop a database and has taken steps to improve the quality of the
data. We also note that servicers reported difficulties with the
matching of first and second liens, including concerns about the
accuracy and completeness of the data, which contributed to the slow
initial implementation of 2MP. As a result of these challenges,
servicers had been modified relatively few second liens a year after
program guidelines were issued. The pages referenced in this comment
are now pages 13 and 24.
3. Treasury noted that the draft report suggested that extensive program
requirements and unclear guidance were obstacles to the program’s
success. The section of the draft report that discussed concerns about

Page 73

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

extensive program requirements was associated with the
implementation challenges with the HAFA program only. The draft
report noted that Treasury itself has acknowledged these obstacles
and has since revised many of the HAFA program requirements that
were identified as contributing to the slow implementation of the
program. With regard to Treasury’s comments about program
guidance, we clarified the language in the report to focus on the
programs’ changing guidance. Servicers told us that ongoing program
revisions presented challenges such as needing to retrain staff and in
some cases delayed program implementation. The pages referenced in
this comment are now pages 15 and 48.
4. Treasury noted that the draft report faulted Treasury for failing to set
numerical goals, especially with regard to the new programs. Treasury
stated the programs were launched under challenging and
unprecedented circumstances, making it extremely difficult to predict
how many homeowners will respond to servicer solicitations, provide
requisition documentation, or accept the modification when offered.
As we, the Congressional Oversight Panel, and SIGTARP have
previously noted, establishing key performance metrics and reporting
on individual servicers’ performance with respect to those metrics are
critical to transparency and accountability. As such, we continue to
believe that it is important that Treasury implement our June 2010
recommendation that it develop measures and benchmarks for its
newer MHA program. Without pre-established performance measures
and goals, Treasury will not be able to effectively assess the outcomes
of its MHA programs or hold servicers accountable for their
performance. The pages referenced in this comment are now pages 12,
23, and 49.
5. Treasury stated that the draft report left the impression that Treasury
chose not to collect data on proprietary modifications. In fact, the
report notes that Treasury does collect data on the post-HAMP
disposition paths. We believe the information that Treasury collects
through its eight largest HAMP servicers provides important and useful
information to policymakers on the disposition of borrowers denied or
canceled from HAMP trial and permanent modifications. However, as
noted in the draft report, we believe the way in which the information
is collected makes it difficult to understand the outcomes of these
borrowers. Without accurate reporting of borrower outcomes,
Treasury cannot know the actual extent to which borrowers who are
denied, canceled, or redefaulted from HAMP are helped by other
programs or evaluate the need for further action to assist this group of

Page 74

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

homeowners. The pages referenced in this comment are now pages 41
to 45.
6. Treasury commented that the draft report suggested a criticism of
HAMP modifications because it notes that proprietary modifications
are more flexible and easier to secure than a HAMP modification.
Treasury notes that it manages a national, publicly financed program
and must balance the interest of taxpayers, investors, and borrowers in
designing and implementing the program. We agree. Our observation is
not intended to suggest that HAMP adopt the flexibility of proprietary
modifications. We are simply describing what is known about
proprietary modifications. Moreover, the report notes that the longterm sustainability of both types of modifications is unclear,
particularly for proprietary modifications because these modifications
may not reduce the monthly payments of borrowers as much as HAMP
modifications have. The pages referenced in this comment are now
pages 41 to 45.
7. Treasury stated that it believed that the overall conclusion reached in
the draft report that the gaps in data limit Treasury’s ability to identify
program gaps is inaccurate and misleading. Treasury noted that some
of the examples of missing or inaccurate data are outliers. It also noted
that the data on permanent modifications is robust and that data
provided by servicers has improved significantly over the past 6
months. We have added text in the report to acknowledge that
Treasury’s data, particularly on the race and ethnicity of borrowers,
has improved over time, and that the reporting in the most recent
public file represents an improvement over the data we received as of
September 30, 2010. We also note that Treasury has worked with
Fannie Mae to make improvements to the data. While we acknowledge
that progress has been made in the quality and accuracy of the data
reported by servicers, we believe that it is critical that Treasury
continue to work toward improving the data so that it and
policymakers can understand the characteristics of both borrowers
who have been helped by HAMP as well as those who could not be
helped by HAMP. This information will be essential to identifying what
additional steps or adjustments could be made to existing TARPfunded programs or other government programs to prevent avoidable
foreclosures to better achieve the goals of preserving homeownership
and protecting property values. The pages referenced in this comment
are now pages 25 to 40.

Page 75

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix III: Comments from the Department
of the Treasury

8. Treasury stated that it instructed servicers to report borrowers in a
single disposition path to avoid double counting of borrowers and,
thereby, provide a clear view of the current path the population is
following through the progression of potential loss mitigation
outcomes. However, the current method of data collection can distort
the current disposition status of borrowers because borrowers are
often “dual-tracked” (e.g., being evaluated for a proprietary
modification while also starting the foreclosure process). Reflecting
the full range of possible outcomes these borrowers face would
improve Treasury’s understanding of the extent to which borrowers
are helped by other programs and assist any evaluation of the need for
further action to assist this group of homeowners.
9. Treasury stated that it disagreed with the draft report’s conclusion that
its programs had not been fully implemented. We revised the language
in the report to more clearly state that the implementation of
Treasury’s MHA programs had gotten off to a slow start and reiterated
that actions needed to be taken by Treasury to better ensure the
success of its programs. The page referenced in this comment is now
page 47.

Page 76

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

Appendix IV: GAO Contacts and Staff
Acknowledgments

Appendix IV: GAO Contacts and Staff
Acknowledgments
GAO Contacts

Mathew J. Scirè (202) 512-8678 or sciremj@gao.gov

Staff
Acknowledgments

In addition to the contacts named above, Lynda Downing, Harry Medina,
John Karikari (Lead Assistant Directors); Tania Calhoun; Emily Chalmers;
William Chatlos; Grace Cho; Rachel DeMarcus; Marc Molino; Mary Osorno;
Jared Sippel; Winnie Tsen; Jim Vitarello; and Heneng Yu made important
contributions to this report.

(250549)

Page 77

GAO-11-288 Treasury’s Foreclosure Mitigation Programs

GAO’s Mission

The Government Accountability Office, the audit, evaluation, and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO’s
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of
GAO Reports and
Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO’s Web site (www.gao.gov). Each weekday afternoon, GAO
posts on its Web site newly released reports, testimony, and
correspondence. To have GAO e-mail you a list of newly posted products,
go to www.gao.gov and select “E-mail Updates.”

Order by Phone

The price of each GAO publication reflects GAO’s actual cost of
production and distribution and depends on the number of pages in the
publication and whether the publication is printed in color or black and
white. Pricing and ordering information is posted on GAO’s Web site,
http://www.gao.gov/ordering.htm.
Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
TDD (202) 512-2537.
Orders may be paid for using American Express, Discover Card,
MasterCard, Visa, check, or money order. Call for additional information.

To Report Fraud,
Waste, and Abuse in
Federal Programs

Contact:

Congressional
Relations

Ralph Dawn, Managing Director, dawnr@gao.gov, (202) 512-4400
U.S. Government Accountability Office, 441 G Street NW, Room 7125
Washington, DC 20548

Public Affairs

Chuck Young, Managing Director, youngc1@gao.gov, (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, DC 20548

Web site: www.gao.gov/fraudnet/fraudnet.htm
E-mail: fraudnet@gao.gov
Automated answering system: (800) 424-5454 or (202) 512-7470

Please Print on Recycled Paper