Single-Family Housing: Progress Made, but Opportunities Exist to 
Improve HUD's Oversight of FHA Lenders (12-NOV-04, GAO-05-13).	 
                                                                 
Every year, the Department of Housing and Urban Development	 
(HUD), through its Federal Housing Administration (FHA), insures 
billions of dollars in home mortgage loans made by private	 
lenders. Oversight of lenders has historically been a challenge  
for HUD. In January 2003, GAO reported that, due in part to poor 
lender oversight, HUD's single-family mortgage insurance programs
remained a high-risk area. This report examines (1) how well HUD 
follows its guidance when granting lenders direct endorsement	 
authority (the ability to underwrite loans and determine their	 
eligibility for FHA mortgage insurance without HUD's prior	 
review), (2) the extent to which HUD uses a risk-based approach  
when monitoring FHA lenders, and (3) the extent to which HUD	 
holds accountable lenders that it identifies as not complying	 
with its performance requirements.				 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-13						        
    ACCNO:   A13453						        
  TITLE:     Single-Family Housing: Progress Made, but Opportunities  
Exist to Improve HUD's Oversight of FHA Lenders 		 
     DATE:   11/12/2004 
  SUBJECT:   Accountability					 
	     Eligibility determinations 			 
	     Monitoring 					 
	     Mortgage loans					 
	     Mortgage protection insurance			 
	     Noncompliance					 
	     Performance measures				 
	     Quality assurance					 
	     Risk management					 
	     Policies and procedures				 
	     HUD Computerized Homes Underwriting		 
	     Management System					 
                                                                 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-13

                 United States Government Accountability Office

                     GAO Report to Congressional Addressees

November 2004

SINGLE-FAMILY HOUSING

Progress Made, but Opportunities Exist to Improve HUD's Oversight of FHA Lenders

                                       a

GAO-05-13

[IMG]

November 2004

SINGLE-FAMILY HOUSING

Progress Made, but Opportunities Exist to Improve HUD's Oversight of FHA Lenders

  What GAO Found

HUD has not consistently followed its guidance for granting direct
endorsement authority. The guidance requires that, to receive the
authority, lenders must, within a 1-year period, submit for HUD's approval
at least 15 mortgage loans that HUD assesses "good" or "fair" using its
assessment criteria, including the last 5 consecutive loans. However, we
found that HUD deviated from this guidance when granting authority to some
of the 49 lenders that were approved between October 1, 2002, and April
30, 2004. For example, HUD granted authority to 7 lenders who did not
submit the minimum 15 loans rated "good" or "fair."

HUD has been using a risk-based approach to monitoring lenders, employing,
among other things, aggregate loan performance data to target lenders for
review. However, certain factors limit the usefulness of its monitoring
tools. First, the rating system HUD uses when performing technical
reviews-desk audits to evaluate the underwriting quality of loans insured
by FHA-does not currently reflect the different levels of risk that
detected underwriting errors pose to the insurance fund. HUD is in the
process of revising the system to improve its usefulness. Second, while
GAO found that, in fiscal year 2003 and the first half of fiscal year
2004, HUD generally reviewed those lenders that met its targeting
criteria, its reports on lender reviews do not distinguish between those
conducted on-site (at lenders' offices) and off-site ("desk" reviews).
HUD's guidance allows desk reviews, but on-site reviews are preferred
because, among other things, they allow for direct observation and the
ability to easily review more loans. HUD's reports do not identify the
number of off-site reviews, but a manual search of records showed that 70
of the 910 lender reviews conducted in fiscal year 2003 were off-site
reviews.

HUD's efforts to hold poor performing lenders accountable have not been
comprehensive. HUD has made limited use of its ability to suspend the
direct endorsement authority of noncompliant lenders, suspending 7 (of
about 2,900 lenders with direct endorsement authority) in fiscal year 2003
and the first half of fiscal year 2004. Further, HUD's Mortgagee Review
Board can take over a year to take action, during which time noncompliant
lenders may continue to make FHA-insured loans.

                 United States Government Accountability Office

Contents

  Letter

Results in Brief
Background
Homeownership Centers Have Not Consistently Followed HUD's

Guidance for Granting Direct Endorsement Authority HUD's Monitoring Is
Risk-Based, but Certain Factors Limit the Usefulness of Its Monitoring
Tools Efforts to Hold Lenders Accountable for Poor Performance Have

Not Been Comprehensive Conclusions Recommendations for Executive Action
Agency Comments and Our Evaluation

1 3 5

9

13

27 35 36 36

Appendixes

Appendix I: Objectives, Scope, and Methodology 40

Appendix II:Comments from the Department of Housing and Urban Development
43

Appendix III:GAO Contacts and Staff Acknowledgments 52 GAO Contacts 52
Staff Acknowledgments 52

Figures	Figure 1: Figure 2:

Figure 3:

Figure 4:

Figure 5:

Figure 6:

Figure 7:

FHA Mortgage Application Process 6
Geographical Jurisdictions of HUD's Four
Homeownership Centers and Lender Branches in Each
Jurisdiction 8
Preclosing Review Loan Rating Categories and Scoring
System 10
Steps to Be Taken Each Quarter to Target Lenders for
Lender Reviews 14
Extent to Which Lenders Reviewed by Homeownership
Centers Were Targeted Lenders, Fiscal Year 2003 and
First Two Quarters of Fiscal Year 2004 15
Extent to Which Lenders Targeted by Homeownership
Centers Were Reviewed, Fiscal Year 2003 and First Two
Quarters of Fiscal Year 2004 16
Percentage of Loans Receiving Technical Reviews, Fiscal
Year 2003 and First Two Quarters of Fiscal Year 2004 19

Contents

Figure 8:	Percentage of Technical Review Contractors' Work Reviewed by
Homeownership Centers in Fiscal Year 2003 21

Figure 9:	Technical Review Contractors' Fiscal Year 2003 Performance 22

Figure 10: Percentage of "Poor" Ratings Assigned during Technical Reviews
Performed in Fiscal Year 2003 and First Two Quarters of Fiscal Year 2004,
by Category 23

Figure 11: Deficiencies Most Commonly Cited during Technical Reviews
Performed in Fiscal Year 2003 and First Two Quarters of Fiscal Year 2004
25

Figure 12: Results of the First 19 Rounds of HUD's Credit Watch Program 28
Figure 13: Status of the Mortgagee Review Board's Actions on 32 Cases as
of June 2004 33

Abbreviations

CHUMS Computerized Homes Underwriting Management System
FHA Federal Housing Administration
HECM Home Equity Conversion Mortgage
HUD Department of Housing and Urban Development

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It 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.

A

United States Government Accountability Office Washington, D.C. 20548

November 12, 2004

The Honorable Paul S. Sarbanes

Ranking Minority Member

Committee on Banking, Housing, and Urban Affairs United States Senate

The Honorable Michael G. Oxley Chairman The Honorable Barney Frank Ranking
Minority Member Committee on Financial Services House of Representatives

Every year, the Department of Housing and Urban Development (HUD), through
its Federal Housing Administration (FHA), insures billions of dollars in
home mortgage loans made by private lenders.1 During fiscal year 2003
alone, FHA insured over 1.3 million mortgages valued at about $160
billion. While FHA insures lenders against nearly all losses resulting
from foreclosed loans, it relies on the lenders to underwrite the loans
and determine their eligibility for FHA mortgage insurance.2 Oversight of
FHA lenders has historically been a challenge for HUD. In April 2000, we
reported on weaknesses in HUD's lender approval, monitoring, and
enforcement efforts.3 We also reported in January 2003 that, due in part
to poor lender oversight, HUD's single-family mortgage insurance programs
remained a high-risk area for HUD.4 Furthermore, HUD's Office of Inspector
General noted in its most recent semiannual report to Congress

1FHA is a part of HUD, and the Assistant Secretary for Housing is also the
Federal Housing Commissioner.

2Underwriting refers to a risk analysis that uses information collected
during the origination process to decide whether to approve a loan.

3GAO, Single-Family Housing: Stronger Oversight of FHA Lenders Could
Reduce HUD's Insurance Risk, GAO/RCED-00-112 (Washington, D.C.: Apr. 28,
2000).

4GAO, High-Risk Series: An Update, GAO-03-119 (Washington, D.C.: Jan. 1,
2003).

that FHA's single-family mortgage insurance programs continue to be a
major management challenge for the department.5

While over 10,000 lending institutions are approved to participate in
FHA's single-family mortgage insurance programs, only about 2,900 of these
institutions have direct endorsement authority, meaning that they can
underwrite loans and determine their eligibility for FHA mortgage
insurance without HUD's prior review. Lenders with this authority
underwrite virtually all FHA-insured mortgages for single-family homes.
This report examines (1) how well HUD follows its guidance when granting
lenders direct endorsement authority, (2) the extent to which HUD uses a
risk-based approach when monitoring the lenders participating in FHA's
mortgage insurance programs, and (3) the extent to which HUD is holding
lenders that it identifies as not complying with its requirements
accountable for their performance. We conducted this review on the
initiative of the Comptroller General.

To address our objectives, we reviewed the activities of HUD's
headquarters and its four homeownership centers in Atlanta, Georgia;
Denver, Colorado; Philadelphia, Pennsylvania; and Santa Ana, California,
which administer HUD's single-family housing activities in all 50 states,
the District of Columbia, and Puerto Rico. At each homeownership center,
we reviewed the documentation maintained on lenders to which HUD had
recently granted direct endorsement authority. We also obtained and
analyzed data on the lenders that HUD had targeted for reviews and on the
loans HUD had selected for technical reviews in fiscal year 2003 and the
first two quarters of fiscal year 2004. Finally, we reviewed case files
acted on by the Mortgagee Review Board-an enforcement body chaired by
HUD's Assistant Secretary for Housing-Federal Housing Commissioner. We
assessed the reliability of the HUD data we used by discussing the data
with knowledgeable agency officials, reviewing information about the
systems, and performing electronic testing to detect obvious errors in
completeness and reasonableness. We determined that the data were
sufficiently reliable for the purposes of this report.

5U.S. Department of Housing and Urban Development, Office of Inspector
General, Semiannual Report to Congress, October 1, 2003 through March 31,
2004 (Washington, D.C.).

We performed our work from December 2003 to September 2004 in accordance
with generally accepted government auditing standards. Appendix I provides
additional details on our scope and methodology.

Results in Brief	HUD's homeownership centers are not consistently
following the department's guidance for granting direct endorsement
authority. FHAapproved lenders must demonstrate "acceptable performance"
in underwriting at least 15 mortgage loans before receiving direct
endorsement authority. HUD's four homeownership centers perform
evaluations, known as preclosing reviews, of these loans in order to
assess lenders' performance. According to HUD's guidance, acceptable
performance is defined as submitting a minimum of 15 loans that are rated
"good" or "fair" within a 1-year probationary period, with the last 5
consecutive cases rated "good" or "fair." We found, however, that the
homeownership centers have not consistently followed this guidance, based
on our analysis of the preclosing reviews performed for all 49 lenders
that entered the probationary period on or after October 1, 2002, and were
granted direct endorsement authority by April 30, 2004. For example, 7 of
the 49 lenders were granted direct endorsement authority, although they
did not submit the minimum 15 loans rated "good" or "fair."

HUD uses a risk-based approach to monitoring lenders, employing aggregate
loan performance data, complaints of irregularities or fraudulent
practices, the results of technical reviews of individual loans, and/or
other factors to target lenders for review. However, certain factors limit
the usefulness of its monitoring tools.

o 	HUD's technical reviews do not distinguish between different levels of
risk. Technical reviews are desk audits to evaluate the underwriting
quality of individual loans insured by FHA. In February 2004, HUD
implemented an algorithm that allows it to select loans for technical
reviews based on certain risk factors, such as loans made to first-time
homebuyers and adjustable rate mortgages. However, the ratings that are
assigned during technical reviews do not currently reflect the different
levels of risk that underwriting errors pose to the insurance fund.
According to our analysis of technical reviews conducted in fiscal year
2003 and the first two quarters of fiscal year 2004, 70 percent of the
loans rated on mortgage credit analysis received "poor" ratings, meaning
that the lenders made mistakes in evaluating the borrowers'
creditworthiness. Under the current rating system, there is no way to
distinguish a "poor" that represents a deficiency posing a risk to the

insurance fund from a "poor" that represents a compliance or documentation
issue (such as an undated or unsigned form). The homeownership centers are
in the process of revising the rating system to make it more risk-based.
Despite their numbers-over 130,000 in fiscal year 2003 and the first two
quarters of fiscal year 2004-technical reviews serve a limited purpose and
do not help HUD identify loans that have a high probability of default or
loans susceptible to fraud.

o 	HUD's reports on lender reviews do not distinguish between on-site and
desk reviews. One of HUD's primary tools for evaluating the quality of
lenders' mortgage-lending practices is lender reviews, which are generally
on-site evaluations of lenders' operations. Since May 2000, the
homeownership centers have been selecting lenders for lender reviews based
on their default and claim rates on FHA-insured mortgages. We found that,
in fiscal year 2003 and the first half of fiscal year 2004, HUD generally
reviewed those lenders that met its targeting criteria. However, HUD's
reports on lender reviews do not identify the number of reviews that were
performed as desk reviews (off-site reviews). Although HUD's guidance
allows staff to complete desk reviews of lenders' operations, the guidance
and homeownership center officials acknowledge that on-site reviews at the
lender's main office or branch are the preferred method of monitoring
lenders' operations. Because HUD's reports do not routinely track the
number of desk reviews, HUD officials conducted a manual search of their
records and determined that 70 of the 910 lender reviews conducted in
fiscal year 2003 were offsite reviews.

HUD's efforts to hold poorly performing lenders accountable for their
performance have not been comprehensive. HUD has recently proposed changes
to improve the effectiveness of its Credit Watch program-an enforcement
tool used to terminate the loan origination authority of lenders with
excessive default and claim rates on FHA-insured loans. Specifically, it
has proposed holding the lenders that underwrote the loans, in addition to
the lenders that originated the loans, accountable for excessive defaults
or insurance claims. Although HUD's guidance allows the homeownership
centers to suspend the direct endorsement authority of lenders that fail
to comply with FHA's underwriting requirements, the homeownership centers
have made limited use of this ability. In fiscal year 2003 and the first
two quarters of fiscal year 2004, the Philadelphia homeownership center
suspended the direct endorsement authority of seven lenders; however, the
other three homeownership centers did not take this action against any
lenders. Additionally, the Mortgagee Review

Board's (Board) process for sanctioning lenders is time consuming. The
Board, which can impose administrative sanctions against lenders, has
taken over a year to complete its actions, during which time the lender
can continue to make dozens of loans.

This report contains recommendations designed to improve HUD's processes
for approving and monitoring FHA mortgage lenders and sanctioning them for
unacceptable performance. We provided HUD with a draft of this report for
review and comment. HUD agreed with our recommendations but disagreed with
some of our findings and stated that the report does not fully recognize
the accomplishments resulting from changes it has made to lender
oversight. We did not change our findings because HUD provided no new
evidence, and we believe that the report appropriately recognizes the
progress HUD has made.

Background	Established by the National Housing Act, FHA insures lenders
against losses on mortgages for single-family homes.6 Lenders usually
require mortgage insurance when a homebuyer has a down payment of less
than 20 percent of the value of the home. FHA mortgage insurance allows a
homebuyer to make a modest down payment and obtain a mortgage for the
balance of the purchase price. FHA plays a particularly large role in
certain market segments, including low-income borrowers and first-time
homebuyers. During fiscal years 2001 to 2003, the number of single-family
mortgage loans that FHA insured annually averaged about 1.2 million. For
the 3 years combined, FHA insured about 3.7 million mortgages with a total
value of about $425 billion.

A homebuyer seeking a FHA-insured mortgage must submit a mortgage
application to a FHA-approved lender. Once the lender approves the loan,
it sends the loan documents to HUD for approval of FHA mortgage insurance.
(See fig. 1.) If the borrower defaults and the lender subsequently
forecloses on the loan, the lender can file an insurance claim with HUD
for the unpaid balance of the loan. FHA insures most of its mortgages for
single-family housing under its Mutual Mortgage Insurance Fund (Fund). To
cover lenders' losses, FHA collects insurance premiums that borrowers pay
to lenders and deposits the premiums in the Fund. The Fund has
historically been self-sufficient. An actuarial study by Deloitte & Touche

6Single-family loans insured by FHA may be used to finance the purchase of
new or existing one-to-four-family properties. 12 U.S.C. 1709(b).

meaning that they can accept mortgage applications, obtain employment
verifications and credit histories on applicants, order appraisals, and
perform other tasks that precede the loan underwriting process.
Approximately 2,900 of the FHA-approved lending institutions also have
direct endorsement authority, meaning that they can underwrite loans and
determine their eligibility for FHA mortgage insurance without HUD's prior
review.8 Underwriting refers to a risk analysis that uses information
collected during the origination process to decide whether to approve a
loan. Virtually all FHA-insured mortgages for single-family homes are
underwritten by lenders with direct endorsement authority.

Some FHA-approved lenders with direct endorsement authority, known as
sponsoring lenders, enter into agreements to underwrite and fund loans
originated by other FHA lenders who do not have direct endorsement
authority, known as loan correspondents. About 71 percent of FHA's
approved lenders are loan correspondents, meaning that they originate
FHA-insured mortgages and sell or transfer the loan paperwork to
sponsoring lenders for underwriting and approval. According to HUD's
regulations, sponsoring lenders are responsible for the loan origination
activities of their loan correspondents.

HUD's 2020 Management Reform Plan, which was announced in 1997,
consolidated the single-family mortgage housing activities of HUD's 81
field offices into four homeownership centers, each of which is
responsible for a multistate area. (See fig. 2.) The homeownership centers
are located in Atlanta, Georgia; Denver, Colorado; Philadelphia,
Pennsylvania; and Santa Ana, California; and report directly to the Deputy
Assistant Secretary for Single Family Housing in HUD's Washington, D.C.,
headquarters.

8To be eligible to receive direct endorsement authority and to underwrite
FHA-insured loans, a lender, in addition to meeting other HUD
requirements, must be one of the following: (1) a member of the Federal
Reserve System or an institution whose accounts are insured by the Federal
Deposit Insurance Corporation or the National Credit Union Administration;
(2) a financial institution whose principal activity is lending or the
investing of funds in real estate mortgages; or (3) a federal, state, or
municipal government agency.

Figure 2: Geographical Jurisdictions of HUD's Four Homeownership Centers
and Lender Branches in Each Jurisdiction

Source: GAO analysis of HUD data.

The homeownership centers are responsible for processing and approving
mortgage insurance, as well as implementing several critical aspects of
HUD's lender approval, monitoring, and enforcement activities. These
responsibilities include (1) evaluating, through a process known as
preclosing reviews, loans submitted by FHA-approved lenders seeking direct
endorsement authority and granting direct endorsement authority to
qualified lenders; (2) evaluating lenders' operations, through a process
known as lender reviews, and monitoring lenders' performance through
reviews of individual loans, known as technical reviews; and (3) taking
enforcement actions against lenders that have not complied with FHA's
requirements. HUD's headquarters also has important approval, monitoring,
and enforcement functions. For example, HUD's headquarters is responsible
for annually recertifying lenders that wish to participate in FHA's
mortgage programs. HUD's Credit Watch program, an initiative to identify
and impose sanctions against lenders with unacceptably high rates of
defaults and insurance claims on FHA-insured mortgages, is managed by
HUD's Office of Lender Activities and Program Compliance. HUD's Mortgagee
Review Board, an enforcement body chaired by HUD's Assistant Secretary for
Housing-Federal Housing Commissioner, can impose

administrative sanctions against lenders, including withdrawing the
lenders' authority to make FHA-insured loans.9

In April 2000, we reported on HUD's lender approval, monitoring, and
enforcement efforts. Among other things, we noted that HUD's guidance for
granting direct endorsement authority was not clear, which led the
homeownership centers to interpret it differently. We also reported that
HUD's monitoring did not focus on the lenders and loans that posed the
greatest insurance risks to the department. In addition, we observed that
the homeownership centers were making only limited use of their ability to
suspend lenders' direct endorsement authority and that HUD's Credit Watch
program pertained only to the lenders that originated troubled loans.

  Homeownership Centers Have Not Consistently Followed HUD's Guidance for
  Granting Direct Endorsement Authority

HUD's homeownership centers, which are responsible for granting direct
endorsement authority to lenders participating in FHA's programs, have not
consistently followed HUD's guidance for granting this authority.
According to departmental guidance, FHA-approved lenders seeking direct
endorsement authority must go through a probationary period and are
required to demonstrate "acceptable performance" in underwriting at least
15 mortgage loans. During this probationary period, the lenders send to
the homeownership centers mortgage loans that have not yet been "closed"-
that is, the borrower has not yet taken on the loan obligation. The
homeownership centers then evaluate the loans against FHA's underwriting
requirements.10 During these evaluations, known as preclosing reviews, the
homeownership centers rate the quality of the construction exhibits (for
new or rehabilitated homes), the valuation of the mortgaged property, and
the mortgage credit evaluation of the borrower as "good," "fair," or
"poor."11 (See fig. 3.) According to HUD guidance, a "good" rating
indicates no underwriting deficiencies, a "fair" rating indicates the
presence of deficiencies that would not affect the insurance eligibility
determination,

9The other members of the Board are HUD's General Counsel, Chief Financial
Officer, Assistant Secretary for Administration, Assistant Secretary for
Fair Housing and Equal Opportunity, and Director of the Enforcement
Center; and the President of the Government National Mortgage Association.

10At the Atlanta, Denver, and Philadelphia homeownership centers, staff
perform these evaluations. In contrast, a contractor performs these
evaluations for the Santa Ana homeownership center.

11During a preclosing review, the Denver homeownership center also
sometimes evaluates the quality of the loan-closing documents.

and a "poor" rating indicates underwriting errors that would significantly
increase HUD's insurance risk. HUD's guidance provides specific criteria
for the homeownership centers to use in determining these ratings.

Figure 3: Preclosing Review Loan Rating Categories and Scoring System

Good

Fair

Poor

                                  Source: GAO.

Note: "N/A" means not applicable. The three loan examples are provided for
illustration only. The overall rating is equal to the lowest rating
assigned to an individual rated category. For example, if one category is
rated "poor," the overall rating for that loan is a "poor."

In our April 2000 report, we noted that HUD's guidance for granting direct
endorsement authority was unclear because it did not define what would
constitute overall acceptable performance. As a result, we found that the
homeownership centers had implemented the existing guidance differently
and had approved lenders that demonstrated varying levels of proficiency,
including lenders that had made multiple and serious underwriting
mistakes. In response to our report, HUD issued its current guidance for
granting FHA-approved lenders direct endorsement authority in September
2002. The guidance states that, in order to qualify for direct endorsement
authority, a lender must submit a minimum of 15 mortgage loans that
receive ratings of "good" or "fair" within a year, with the last 5
consecutive loans rated "good" or "fair." These 15 mortgages may include
loans for

home purchase transactions (including 203(k) loans) and full
creditqualifying refinances.12 Only 5 of the 15 required mortgages may be
from a combination of automated underwriting, streamline refinances, and
fully underwritten loans denied by other lenders.13 In addition, a lender
cannot submit more than 30 mortgage loans for HUD's review during this
probationary period. The guidance states that, if the lender has submitted
30 loans and has not met the standards to be granted direct endorsement
authority, the lender must be notified that it cannot submit mortgage
loans for at least 6 months.

Although HUD has issued specific guidance, the homeownership centers have
not consistently followed it. To determine how well each homeownership
center followed HUD's guidance, we analyzed preclosing review ratings
given to the loans submitted by all 49 lenders that entered the
probationary period on or after October 1, 2002, (after the guidance was
implemented) and were granted direct endorsement authority by April 30,
2004.14 (Approximately 290 other lenders were in the process of seeking,
but had not yet received, direct endorsement authority as of April 30,
2004.) The 49 lenders submitted an average of 24 loans to the
homeownership centers for preclosing reviews.

12Under the 203(k) Home Rehabilitation Mortgage Insurance program, a
borrower can get one mortgage loan to finance both the acquisition and
rehabilitation of the property. A refinance transaction involves repaying
an existing real estate debt from the proceeds of a new mortgage that has
the same borrower and the same property.

13When automated underwriting is used, a computer-based tool simplifies
the processing of loan applications by analyzing, among other things, how
a borrower managed credit obligations in the past and whether the borrower
has the ability to repay the mortgage loan. It then provides a
recommendation to the lender to approve the loan or refer it for manual
underwriting. A streamline refinance is a type of refinance transaction
that requires less paperwork. For example, streamline refinances can be
made without an appraisal, and HUD does not require a credit report.

14One additional lender with direct endorsement authority applied for and
was granted the authority to underwrite Home Equity Conversion Mortgages
(HECM)-mortgages that can be used by senior homeowners to convert equity
into income-during this time period. HUD's guidance states that, after
receiving direct endorsement approval, lenders may apply for approval to
underwrite specialized loans such as HECM and 203(k) loans by submitting a
minimum of five consecutive cases of that type rated "good" or "fair."
According to our analysis, the lender submitted a minimum of five
consecutive HECM cases rated "good" or "fair."

Our analysis showed that the homeownership centers granted some of the 49
lenders direct endorsement authority in violation of HUD's criteria.
Specifically, we found the following:

o 	Seven of the lenders did not submit at least 15 mortgage loans that
were rated "good" or "fair."

o 	Two lenders were granted direct endorsement authority although the last
5 consecutive loans they submitted were not rated "good" or "fair."15

o 	One lender exceeded the allowed 1-year probationary period, and eight
lenders submitted more than the 30 loans allowed before being granted
direct endorsement authority. The number of mortgage loans submitted by
these lenders ranged from 31 to 73.

Our analysis of the loans submitted by the 49 lenders was based on
information contained in a log maintained for each lender seeking direct
endorsement authority. According to homeownership center officials, they
use this log to determine if a lender has met HUD's standards. When
commenting on the results of our analysis, HUD officials stated that some
of the information in the logs we reviewed was incorrect. For example,
they noted that some of the loans were incorrectly entered as automated
underwriting cases. Because the information in the log is what the
homeownership center officials use to determine if the standards have been
met, we did not change our findings based on the new information provided
by HUD. HUD officials also noted that they had sometimes used management
discretion when applying the standards. For example, for one case in which
we determined that the last five consecutive cases were not rated "good"
or "fair," the homeownership center staff determined that, despite the one
loan rated "poor" out of the last five, the lender had submitted a
sufficient number of loans rated "good" or "fair" to be approved.

15In three additional cases, the lender continued submitting loans after
it had satisfied HUD's requirements of having a total of 15 loans rated
"good" or "fair," with the last 5 of the 15 loans rated "good" or "fair."
One of the extra loans submitted was rated "poor," which meant that the
last five consecutive cases submitted by the lender were not rated "good"
or "fair."

  HUD's Monitoring Is Risk-Based, but Certain Factors Limit the Usefulness of
  Its Monitoring Tools

Although HUD has adopted a risk-based approach to monitoring lenders,
certain factors limit the usefulness of the tools it employs. The
homeownership centers rely on two primary monitoring tools to ensure
lenders' compliance with FHA's mortgage requirements: (1) lender reviews,
which are generally on-site evaluations of lenders' operations, performed
by HUD staff and (2) technical reviews, which are desk audits of the
underwriting quality of individual loans already insured by FHA, performed
mainly by contractors. Since May 2000, the homeownership centers have
generally been targeting for review those lenders they consider to be high
risk, but HUD's reports do not distinguish between on-site and desk
reviews. HUD has started selecting loans for technical reviews based on
characteristics associated with risk and done a better job of tracking the
performance of the contractors that perform most of HUD's technical
reviews. However, its technical review rating system does not currently
reflect the different levels of risk that underwriting errors pose to the
insurance fund.

    Although HUD Is Following Its Guidance in Targeting High-Risk Lenders, Its
    Reports Do Not Identify Desk Reviews

Most Lenders Reviewed Were Targeted Based on Risk

Since May 2000, the homeownership centers have targeted lenders for review
based on indicators of risk, and our analysis shows that they have
generally reviewed the lenders that they have identified as high-risk
lenders. Although on-site reviews are the preferred method of monitoring,
HUD's reports do not identify the number of desk reviews performed.

Lender reviews typically involve an in-depth analysis of a sample of loans
and assessments of lenders' internal control systems for making loans. If
a lender review finds serious deficiencies with specific loans or the
lender's internal controls, HUD may take actions that reduce the
department's insurance risk, such as requiring the lender to compensate
HUD for financial losses that HUD has incurred or may incur on certain
loans. Staff assigned to each homeownership center's quality assurance
division are responsible for scheduling and performing these reviews. In
fiscal year 2003, HUD's homeownership centers conducted 910 lender
reviews, exceeding the department's goal of 900 reviews.

In April 2000, we reported that, contrary to HUD's guidance, the
homeownership centers were not always reviewing the lenders that they
considered to pose the highest risks and concluded that HUD lacked a
systematic process for identifying and prioritizing such lenders for
review. In response, HUD issued a May 2000 memo calling for the
homeownership

centers to target lenders for lender reviews based on indicators of risk.
Because early defaults and claims-loans reported as 90 days or more
delinquent and loans terminated by claim within the first 24 months of
origination-are an indicator of poor lending practices that may ultimately
result in insurance losses, HUD considers them to be the primary risk
factors in targeting lenders for review. Thus, each quarter the
homeownership centers use data from HUD's Neighborhood Watch Early Warning
System (Neighborhood Watch)-an information system that displays loan
performance data by loan types and geographic areas-to identify the
lenders that pose the highest risk to the insurance fund in terms of
defaults and claims.16 In addition, the guidance lists other factors to be
considered when targeting lenders, including the length of time since
their last review, complaints or reports of irregularities or fraudulent
activity in a lender's practices, and the results of HUD's technical
reviews of individual lenders' loans. (See fig. 4.) According to HUD, the
target reports developed each quarter to identify the lenders to be
reviewed are "fluid;" for example, changes may result if there is a large
number of complaints about a particular lender.

Figure 4: Steps to Be Taken Each Quarter to Target Lenders for Lender
Reviews

Source: GAO analysis of HUD guidance.

Note: Other factors that the homeownership centers are to consider during
the targeting process include high-risk programs such as the 203(k)
program and sudden increases in business volume.

16The loan information in Neighborhood Watch is displayed for a 2-year
origination period and is updated monthly.

We found that HUD's homeownership centers are generally following this
guidance when targeting lenders for reviews. All four homeownership
centers provided us with lists of the lenders they targeted for review in
fiscal year 2003 and the first two quarters of fiscal year 2004 and the
lenders they reviewed during the same time period. Overall, our analysis
of these lists showed that about 80 percent of the lenders reviewed by the
four homeownership centers during these six quarters were lenders on their
target lists. As shown in figure 5, the percentage of lenders reviewed by
each homeownership center that were on their target lists ranged from 68
percent in Denver to 89 percent in Atlanta.

Figure 5: Extent to Which Lenders Reviewed by Homeownership Centers Were
Targeted Lenders, Fiscal Year 2003 and First Two Quarters of Fiscal Year
2004

Lenders reviewed

1,500

1,200

900

600

300

0 Atlanta Denver Philadelphia Santa Total Ana

Homeownership center

Nontargeted lenders Targeted lenders Source: GAO analysis of data provided
by HUD's four homeownership centers.

Furthermore, about 69 percent of the lenders that were targeted by the
homeownership centers had been reviewed by the end of the six quarters. As
shown in figure 6, the percentage of lenders that were targeted and

reviewed ranged from 57 percent for Santa Ana to 82 percent for
Philadelphia. According to HUD, reviews not completed during the quarter
are carried over to the subsequent quarter and nearly all are completed.

Figure 6: Extent to Which Lenders Targeted by Homeownership Centers Were
Reviewed, Fiscal Year 2003 and First Two Quarters of Fiscal Year 2004

Atlanta Denver Philadelphia Santa Ana Total

                       Targeted lenders not yet reviewed

                           Targeted lenders reviewed

Source: GAO analysis of data provided by HUD's four homeownership centers.

HUD is seeking to improve its risk-based monitoring of lenders. According
to HUD officials, the department has hired a contractor to help it analyze
all collected FHA loan data to determine if more of it can be used to
target lenders for review. According to the statement of work, the
contractor is to design, among other things, a risk-based model using HUD
data that will identify lenders that pose a risk to the FHA insurance
fund.17 In developing this model, the contractor is to (1) analyze
risk-based models used by Fannie Mae, Freddie Mac, and private mortgage
insurers to determine how these entities evaluate the risk related to
single-family loans; (2) analyze the data available in HUD's data systems
that can be used to develop riskbased model(s); and (3) recommend
additional data not already available

17The risk-based model that is to be developed by the contractor must be
adaptable and compatible with HUD's Neighborhood Watch System. The lender
information that may be used in developing the model includes, among other
things, the percentage of originations with late up-front mortgage
insurance premiums, the percentage of signed indemnification agreements
(which require the lender to compensate HUD for financial losses that HUD
has incurred or may incur on certain loans), and high default and claim
rates.

in HUD's systems that should be used in the development of a risk-based
model.18

Reports Did Not Distinguish According to HUD's guidance on conducting
lender reviews, a desk review

Between On-Site and Desk may be acceptable for a focused review-a review
of a specific operational

Reviews	area, specific loan type, or specific issue-and necessary when
travel funds are constrained. Even so, both HUD and homeownership center
officials acknowledge that on-site reviews are the preferred method of
monitoring lenders' operations. General HUD guidance on monitoring states
that onsite monitoring reviews are essential for high-risk programs. In
addition, its guidance on conducting lender reviews lists certain factors
that should be considered, including determining if the lender's office
facilities meet HUD's requirements. For example, when conducting an
on-site review a reviewer should, among other things, observe whether the
public can properly identify the lender. Homeownership center officials
also note that on-site reviews are preferable because they can request
additional loans to review on short notice and they sometimes get leads
from employees who want to disclose problems. On-site reviews also give
HUD an opportunity to provide technical assistance to the lenders.

All four homeownership centers are performing some off-site lender reviews
(i.e., desk reviews); however, their reports do not identify the number of
desk reviews performed. At our request, homeownership center officials
conducted a manual search of their records and determined that 70 of the
910 lender reviews performed in fiscal year 2003 (about 8 percent) were
desk reviews. Although all four homeownership centers performed at least
some desk reviews, HUD's Office of Lender Activities and Program
Compliance, the headquarters office that oversees lender reviews,
described all of the reviews that the homeownership centers performed in
fiscal year 2003 as on-site reviews in its annual report.

18Both Fannie Mae and Freddie Mac are federally-chartered corporations
that purchase residential mortgages and convert them into securities for
sale to investors; by purchasing mortgages, they supply funds that lenders
may loan to potential homebuyers.

    HUD Now Selects Loans for Technical Reviews Based on Risk and Better
    Oversees Contractors, but Its Reviews Serve a Limited Purpose

Selection Is Generally Based on Loan Risk Characteristics

In response to recommendations in our April 2000 report, HUD has started
selecting loans for technical reviews based on loan risk characteristics
and improved its oversight of the contractors that perform technical
reviews. However, technical reviews serve a limited purpose because the
system used to rate lender performance on individual loans does not
identify the underwriting errors that pose the greatest risk, and the
reviews do not help HUD identify (1) loans that have a high probability of
default or claim or (2) loans susceptible to fraud.

Technical reviews are desk audits that evaluate the underwriting quality
of individual loans already insured by FHA. They are similar to preclosing
reviews in that HUD evaluates the quality of the construction exhibits
(for new or rehabilitated homes), the valuation of the mortgaged property,
the mortgage credit evaluation of the borrower, and the loan-closing
documents and assigns a "good," "fair," or "poor" rating in each
applicable category. Reviews revealing serious deficiencies may result in,
among other things, HUD's requiring the lenders to compensate the
department for financial losses or HUD's suspending the lenders' direct
endorsement authority. In total, the four homeownership centers performed
133,446 technical reviews in fiscal year 2003 and the first two quarters
of fiscal year 2004, representing 7 percent of the loans that FHA insured
during that time period (see fig. 7).

Figure 7: Percentage of Loans Receiving Technical Reviews, Fiscal Year
2003 and First Two Quarters of Fiscal Year 2004

Percentage

8

7

6

5

4

3

2

1

0 Atlanta Denver Philadelphia Santa Total Ana

Homeownership center

Source: GAO analysis of HUD data.

Prior to February 2004, HUD used the Computerized Homes Underwriting
Management System (CHUMS)-a computer system that assists and supports HUD
staff in processing mortgage insurance for single-family homes-to randomly
select a certain percentage of each lender's loans for technical reviews.
However, as we noted in our April 2000 report, HUD's guidance recommends
that the loans selected for technical reviews should be those that pose a
high risk of loss to the insurance fund. In February 2004, HUD implemented
a CHUMS algorithm that provides a risk-based statistical process for
selecting loans for review at time of approval. The algorithm selects
loans for review based on certain characteristics-such as loans made to
first-time homebuyers, loans with adjustable rate mortgages, and loans for
multiple housing units. According to HUD officials, loans that exhibit
these high-risk characteristics are, all other things being equal, more
likely to be subject to default and/or contain underwriting errors than
loans that do not.

Homeownership center staff also have the ability to adjust, in CHUMS, the
percentage of a lender's loans selected for technical reviews to more

closely monitor certain lenders. For example, HUD's guidance states that
the homeownership centers should perform technical reviews of 100 percent
of at least the first 30 FHA-insured loans made by newly approved direct
endorsement lenders. However, our analysis of loans made by the 49 lenders
that entered the probationary period on or after October 1, 2002, and were
granted direct endorsement authority by April 30, 2004, shows that the
homeownership centers have not followed this guidance. As of June 2004,
only 16 of these lenders had used their direct endorsement authority to
make loans. Contrary to HUD guidance, the homeownership centers had
reviewed only about 7 percent of these lenders' early loans. According to
homeownership center officials, they do not always select the first 30
loans to review because some of the lender's early loans may have been
made by a new branch office of which they are unaware.19 Also, CHUMS does
not automatically maintain the 100 percent designation used to flag a new
lender's early loans for review. As part of an effort to control the
volume of technical reviews, CHUMS revises some of the review percentages
each quarter. As a result, the 100 percent designation for newly
authorized lenders is sometimes changed to less than 100 percent, causing
the homeownership centers to miss some of these lenders' loans.

Tracking of Technical Review The large majority of HUD's technical reviews
are performed by firms

Contractors' Performance Has under contract with the homeownership
centers, and HUD has done a

Improved in Recent Years	better job of tracking these contractors'
performance in recent years.20 In our April 2000 report, we noted that the
technical review contracts in place at the time contained specific
performance standards expressed as the maximum acceptable percentage of
reviews that could contain significant errors or omissions. However, we
found that three of the four homeownership centers were not tracking the
contractors' work against these standards. As a result, these
homeownership centers lacked the information necessary to evaluate the
quality of the contractors' work or to determine whether actions should be
taken against the contractors for poor performance.

19Direct endorsement authority is granted to a lender's home office and
applies to all of the lender's branch offices.

20Virtually all of the Atlanta, Philadelphia, and Santa Ana homeownership
centers' technical reviews are performed by contractors. In contrast,
Denver homeownership center staff performed 44 percent of the
homeownership center's reviews in fiscal year 2003. At the time of our
study, the Santa Ana homeownership center had two firms under contract,
while the Atlanta, Denver, and Philadelphia homeownership centers each
used a single contractor.

The four homeownership centers are currently evaluating the quality of
their contractors' work. Each homeownership center's technical review
contract states that the contractor must deliver 90 percent of the
completed reviews without any errors. An error is defined as any instance
in which HUD has to change a "poor" rating to a "good" or "fair" rating or
when it has to change a "fair" or "good" rating to a "poor" rating. To
determine if the contractor is meeting this standard, the contracts
require HUD to randomly select and evaluate a minimum of 5 percent of the
contractor's completed reviews. If the contractor's error rate exceeds 10
percent for the review period, HUD's payment will be reduced by 1 percent
for each error percentage point above 10 percent. As shown in figure 8,
the homeownership centers reviewed at least 5 percent of their
contractors' work in fiscal year 2003.

Figure 8: Percentage of Technical Review Contractors' Work Reviewed by
Homeownership Centers in Fiscal Year 2003

Percentage

14

12

10

8

6

5% guidance

benchmark 4

2

0 Atlanta Denver Philadelphia Santa Ana Homeownership center

Source: GAO analysis of data provided by HUD's four homeownership centers.

The homeownership centers have used the results of their quality assurance
reviews to track their contractors' performance. For each month in fiscal
year 2003, the four homeownership centers calculated their

contractors' error rate (a total of 60 calculated error rates because
Santa Ana uses two contractors). As a result, they were able to track
whether their contractors exceeded the allowed error rate of 10 percent.
As shown in figure 9, three of the four centers identified that their
contractors had exceeded the allowed error rate.

Figure 9: Technical Review Contractors' Fiscal Year 2003 Performance

Months

12

10

8

6

4

2

0 Atlanta Denver Philadelphia Santa Ana Santa Ana

                         (contractor 1) (contractor 2)

Homeownership center

Months contractor stayed under or equaled 10% error rate

Months contractor exceeded 10% error rate

Source: GAO analysis of data provided by HUD's four homeownership centers.

Despite better tracking, the homeownership centers were not always able to
hold their contractors accountable for unacceptable performance. Only the
Santa Ana center could provide us with support that it assessed
contractors a penalty when appropriate in fiscal year 2003. According to
an Atlanta center official, they did not assess a disincentive to the
contractor in fiscal year 2003 when its error rate exceeded 10 percent
because they did not complete their quality assurance reviews within 30
days. The official also stated that they had solved the timing problem as
of October 2003 and that all reviews are currently completed within the
30-day requirement in

order to properly assess disincentives. Similarly, a Denver center
official stated that the center did not assess any disincentives because
system problems made it difficult to calculate the error rates correctly.
The center has since corrected the problem, according to the same
official. The Philadelphia center's contractor did not exceed the allowed
error rate in fiscal year 2003.

Technical Reviews Serve Limited Homeownership center staff have questioned
the usefulness of technical

Purpose 	reviews because the rating system does not identify the
underwriting errors that pose the greatest risk to the insurance fund.
According to homeownership center officials, the current rating system
results in too many "poor" ratings being assigned. To determine the
percentage of "poor" ratings assigned, we requested data from all four
homeownership centers for fiscal year 2003 and the first two quarters of
fiscal year 2004. As shown in figure 10, the percentage of "poor" ratings
was over 50 percent for at least one category at three of the four
homeownership centers. At each homeownership center, the highest
percentage of "poor" ratings was in mortgage credit. Overall, 70 percent
of the loans that were rated in the mortgage credit category received
"poor" ratings.

Source: GAO analysis of Underwriting Reports System data provided by HUD's
four homeownership centers.

Note: Not all loans are evaluated in all four categories. For example,
only loans for new or rehabilitated homes receive an architecture and
engineering rating.

Although HUD guidance states that a "poor" rating indicates underwriting
errors that significantly increased HUD's insurance risk, homeownership
center officials said that the current system does not distinguish between
a "poor" rating that represents a compliance or documentation issue and a
"poor" rating that represents a risk to the insurance fund. Our analysis
of the most common deficiencies cited during technical reviews performed
in fiscal year 2003 and the first two quarters of fiscal year 2004 shows
that the majority of them are compliance issues. As shown in figure 11,
the most common deficiencies cited often involve problems with paperwork.
According to homeownership center officials, only 3 of the 10 deficiencies
that we identified as the most common represent a risk to the FHA
insurance fund.

Compliance issue Risk issue

Source: GAO analysis of Underwriting Reports System data provided by HUD's
four homeownership centers.

Note: These were the most commonly cited deficiencies that must result in
a "poor" rating. For some deficiencies, the reviewer has the discretion to
decide, based on the severity of the deficiency, whether the rating should
be "fair" or "poor."

aThe Credit Alert Interactive Voice Response System is a federal
government database of delinquent federal debtors maintained to prescreen
direct loan applicants for creditworthiness. HUD maintains a list of
parties who have been issued a limited denial of participation-an action
that excludes a party from further participation in a HUD program area.
The General Services Administration maintains a list of parties excluded
from receiving (1) federal contracts or certain subcontracts and (2)
certain types of federal financial and nonfinancial assistance and
benefits.

bThe "For Your Protection: Get a Home Inspection" form stresses the
importance of obtaining a home inspection prior to purchase.

cThe purpose of the homebuyer summary form is to provide timely
information to the buyer for repairs to be completed or property
conditions that have to be satisfied prior to FHA insurance endorsement.

dThe HUD-1 form, also called the Settlement Statement, records the money
flows that take place when the ownership of a home is transferred from a
seller to a buyer.

HUD's financial statement auditors have also questioned the usefulness of
technical reviews. In the audit of FHA's financial statements for fiscal
years 2002 and 2003, the independent auditors noted that technical
reviews, as currently designed, assisted the homeownership centers in
reporting documentation and processing errors back to the lenders but did
not help them identify loans that have a high probability of default or
claim as a result of poor lender underwriting practices.21 When the
auditors analyzed data on the 2,000 lenders with the highest volume of
originations, they found no strong correlation between the percentage of
technical review "poor" ratings received by a lender and the lender's
early default and claims rates. As a result, the auditors recommended that
HUD consider redesigning the technical review process as an early warning
control that better predicts loan performance so that it could be used not
only as a lender monitoring tool but also as an effective tool to assist
FHA in identifying lenders that originate loans that have a high
probability of going to default or claim.

HUD officials acknowledged that technical reviews were not designed to
help HUD identify loans that have a high probability of default or claim
or loans susceptible to fraud; instead, they were designed to evaluate the
quality of the lender's underwriting. The department has taken several
steps to make technical reviews more meaningful, according to HUD
officials. First, HUD has converted the Underwriting Reports System-the
system used to track the results of technical reviews-to a web-based
system, which will allow it to perform more analysis of the technical
review ratings. Second, HUD plans to revise the deficiency codes used to
assign technical review ratings. Currently, there are over 250 codes for
mortgage credit and valuation issues. The homeownership centers have
proposed reducing that number substantially and replacing the "poor"
rating with a "risk issues" rating-reserved for those deficiencies that
affect the

21To complete the FHA audit, HUD's Inspector General contracted with the
independent certified public accounting firm of KPMG LLP. U.S. Department
of Housing and Urban Development, Office of Inspector General, Audit of
the Federal Housing Administration's Financial Statements for Fiscal Years
2003 and 2002, 2004-FO-0001 (Washington, D.C.: Nov. 25, 2003).

eligibility of the loan-and the "fair" rating with a "compliance issues"
rating, given for those errors that do not affect eligibility.

  Efforts to Hold Lenders Accountable for Poor Performance Have Not Been
  Comprehensive

To hold lenders accountable for poor performance, HUD may (1) terminate
their loan origination authority through its Credit Watch program, (2)
suspend their direct endorsement authority, or (3) take enforcement action
through its Mortgagee Review Board. HUD has terminated the loan
origination authority of 262 lender branch offices and has recently
proposed changes to make its Credit Watch program more effective as a
means of sanctioning lenders responsible for high rates of defaults and
insurance claims. However, the homeownership centers have rarely used
their ability to suspend direct endorsement authority. Further, HUD's
Mortgagee Review Board sometimes takes a year or more to take action
against lenders for program violations, during which time the lender can
continue to make dozens of loans.

    HUD Has Proposed Credit Watch Changes to Improve Program's Effectiveness

HUD has recently proposed changes to improve the effectiveness of its
Credit Watch program. In May 1999, HUD announced that it would begin to
use its Credit Watch program to sanction lenders with excessively high
loan default and claim rates. Initially, HUD planned on a quarterly basis
to (1) terminate the loan origination authority of any lender branch
office whose default and claim rates on mortgages insured by FHA during
the preceding 24 months exceeded both the national average and 300 percent
of the average rate for the HUD field office serving the lender's
geographic location (field office rate) and (2) place on Credit Watch
status the branch offices whose default and claim rates exceeded 200
percent of the average field office rate. While on Credit Watch status,
the branch could continue to originate FHA-insured loans, but HUD would
review the insured loans that the branch originated during a 6-month
period from the date the Credit Watch status became effective for
excessive default rates. In October 2001, HUD announced that it was
eliminating the placement of lenders on Credit Watch status because
advanced warning of excessive default and claim rates was no longer
necessary since the department had provided lenders with access to their
performance via Neighborhood Watch. Also, in September 2002, HUD stated
that it would be gradually reducing the 300 percent termination threshold
set for the HUD field office rate to the 200 percent allowed in HUD's
regulations.

As of July 2004, HUD had conducted 19 rounds under its Credit Watch
initiative, with the last round being based on analysis of the 24-month
period ending December 31, 2003.22 This program has resulted in the
department's termination of 262 branch offices' loan origination
authority.23 As shown in figure 12, the number of branch offices that were
terminated as a result of each round has varied.

Figure 12: Results of the First 19 Rounds of HUD's Credit Watch Program

Number of branch offices terminated Type of lender terminated 50 45

40

35

Loan 30 correspondents

25

20

15

10 Round

                                                           Lenders authorized 
5                                                            to underwrite 
                                                            FHA-insured loans 
0                                                       
      1  2  3  4  5  6 7 8 9 10 11 12 13 14 15 16 17 18 19 

Source: GAO analysis of HUD data.

Currently, the regulations governing HUD's Credit Watch program only allow
the department to hold the lenders that originated the troubled loans
accountable for excessive defaults or insurance claims. The regulations do
not address HUD's authority to also hold accountable those lenders that
have underwritten the loans. When originating mortgage loans, lenders

22Every quarter, HUD conducts a round of Credit Watch by reviewing the
rate of defaults and claims on loans insured by FHA within the preceding
24-month period.

23A terminated lender branch may request to have its authority to
originate FHA loans reinstated no earlier than 6 months after the
effective date of the termination.

perform such functions as accepting mortgage applications and obtaining
employment verifications and credit reports on the borrowers. When
underwriting mortgage loans, lenders use this information to determine
whether borrowers are able to make their mortgage payments and whether the
loans should be approved. As shown in figure 12, 44 percent of the lenders
terminated during the first 19 rounds of Credit Watch were loan
correspondents-lenders that sell or transfer loans that they originate to
other FHA lenders, known as sponsoring lenders, for underwriting and
approval.

In response to a recommendation in our April 2000 report, HUD has
published a proposed rule making several amendments to the Credit Watch
program, including holding underwriting lenders accountable for excessive
default and claim rates.24 The proposed rule, issued in April 2003,
includes several changes designed to strengthen HUD's capacity to
safeguard the FHA mortgage insurance fund. Specifically, it proposes,
among other things, applying the default and claim threshold to both
originating and underwriting lenders and prohibiting a lender from opening
a new branch office in the same lending area as an existing branch that
has received a notice of proposed termination. The proposed rule was open
for comments through June 2, 2003, and HUD submitted an interim rule to
the Office of Management and Budget in July 2004.

    Homeownership Centers Rarely Used Their Ability to Suspend Lenders' Direct
    Endorsement Authority

HUD's homeownership centers have made limited use of their ability to
suspend the direct endorsement authority of lenders that fail to comply
with FHA's program requirements. Lenders whose direct endorsement
authority is suspended but who wish to continue underwriting mortgages
insured by FHA must submit each proposed mortgage case file to a
homeownership center, which evaluates the lenders' underwriting decisions
before deciding whether to insure the loans. The lenders must follow this
procedure until HUD's evaluations of the case files indicate that the
lenders have demonstrated satisfactory performance in underwriting loans.

HUD's guidance allows the homeownership centers to suspend direct
endorsement authority but does not prescribe the circumstances under

24In our April 2000 report, we recommended that HUD revise the Credit
Watch program's regulations to cover lenders that underwrite FHA-insured
loans with excessive default and claim rates, as well as those lenders
that originate such loans.

which the homeownership centers must do so. HUD's handbook on the direct
endorsement program provides general guidelines. For example, the guidance
states that the homeownership centers should consider suspending lenders
that exhibit "patterns" of noncompliance, but it does not define what
would constitute a pattern. After our April 2000 report recommended that
HUD clarify and implement guidelines for identifying lenders whose direct
endorsement authority should be suspended, the department issued
supplemental guidance in a November 2000 memo. The memo describes certain
conditions under which the homeownership centers may suspend direct
endorsement authority.25

Among the four homeownership centers, we found that the Philadelphia
homeownership center was the only one that suspended the direct
endorsement authority of any lenders during fiscal year 2003 and the first
two quarters of fiscal year 2004. Specifically:

o 	The Philadelphia homeownership center took this action against seven
lenders during this time frame, citing underwriting violations identified
during technical reviews and high default rates.26

o 	Although the Denver homeownership center did not suspend any lender's
direct endorsement authority during the same time period, it had warned
seven lenders that it might do so if they did not address underwriting
deficiencies revealed in technical reviews.

o 	Rather than suspending a lender's direct endorsement authority, Atlanta
homeownership center officials told us they will work with problem lenders
to develop performance improvement plans. Such plans generally involve
raising the percentage of a lender's loans that are selected for technical
reviews, meeting with the lender to discuss its performance, and requiring
the lender's staff to take training.

25These conditions include when a lender has maintained a "poor"
percentage in excess of 20 percent for more than 90 days after being
placed on 100 percent review status or when a lender has a claim and
default rate that exceeds both the national rate and 250 percent of the
field office rate (the rate for the HUD field office serving the lender's
geographic location).

26The Philadelphia homeownership center first identifies lenders that (1)
have a default rate that exceeds 150 percent of the field office rate and
(2) have received technical review "poor" ratings over 20 percent (when
more than 20 loans have been reviewed) and begins to review 100 percent of
their loans. For lenders that do not improve after two quarters, the
homeownership center suspends their direct endorsement authority.

o 	Similarly, Santa Ana homeownership center officials said that they tend
to increase the percentage of a lender's loans that are selected for
technical reviews instead of suspending direct endorsement authority.

HUD officials provided several reasons why they do not make more use of
their ability to suspend direct endorsement authority. Officials at all
four homeownership centers told us that suspending lenders would create an
additional workload for them. Atlanta and Santa Ana officials also noted
that suspending a lender's direct endorsement authority would threaten the
lender's business. In addition, Denver officials observed that large
lenders, when faced with suspension at one branch, would just send all
their FHA loans to another branch. Finally, Santa Ana officials stated
that, as long as the percentage of "poor" ratings assigned during
technical reviews was so high, they did not want to rely on them as
grounds for suspending direct endorsement authority. Headquarters
officials noted that they want the homeownership centers to conduct lender
reviews after problems are identified during technical reviews rather than
immediately suspend lenders' direct endorsement authority.

Although HUD's homeownership centers suspended the direct endorsement
authority of relatively few lenders in fiscal year 2003 and the first two
quarters of fiscal year 2004, our analysis of HUD's technical review
ratings for the same time period showed frequent noncompliance by lenders
with FHA's requirements. About 7,800 lender branch offices received
technical review ratings for mortgage credit analysis for the FHAinsured
mortgages they underwrote.27 Of these branch offices, we identified 1,203
that had at least 10 loans reviewed and received a "poor" rating for
mortgage credit analysis-meaning that the lenders made mistakes in
evaluating the borrowers' creditworthiness-on 75 percent or more of their
reviewed loans. Even though the current rating system does not identify
the underwriting errors that pose the greatest risks, this level of
noncompliance indicates that more lenders may be candidates for
enforcement action.

27Lenders could have been counted more than once if they underwrote
FHA-insured mortgages in more than one homeownership center jurisdiction.

    Mortgagee Review Board's Process for Sanctioning Lenders Is Time-Consuming

HUD's Mortgagee Review Board (Board) can impose administrative sanctions
against FHA lenders that commit program violations, such as withdrawal of
a lender's FHA approval. HUD does not have guidelines for the time it
should take for the Board to take enforcement actions against lenders. We
found the process can take over a year from the time the lender is
notified of its violations to Board action.

The majority of the cases referred to the Board are the result of lending
violations revealed in lender reviews performed by HUD's homeownership
centers. Once the Board reviews and accepts a referral, it sends the
lender a notice of violation that provides the lender 30 days to respond
in writing to the Board. After reviewing the lender's response, the Board
decides what actions to take. The Board may impose a number of sanctions
against FHAapproved lenders, ranging from a letter of reprimand-a letter
informing the lender of the existence or occurrence of a violation of
HUD's requirement and directing the lender to bring and maintain its
activities into conformity with all HUD requirements-to withdrawal of a
lender's FHA approval. During the period of withdrawal (generally 3 to 5
years), HUD will not insure any mortgage originated by the withdrawn
lender. Except for a letter of reprimand, the lender has a right to a
hearing before the Board's sanction becomes final.

The majority of the Board's actions result in settlement agreements, which
require lenders to indemnify improperly originated loans; pay fines;
and/or take actions to prevent future lending violations. In June 2004, we
reviewed the Board's records for the 32 cases involving single-family
housing lenders that the Board had acted on in the previous 12 months.28
We found that in 22 of the 32 cases, the Board had either reached a
settlement agreement with the lender (20 cases) or was still attempting to
reach a settlement agreement (2 cases).29 In 8 of the cases, the Board had
withdrawn the lenders' FHA approval. In the remaining 2 cases, the Board
had either assessed a civil money penalty (1 case) or was still pursuing a
civil money penalty (1 case).

28The 32 cases were acted on in Board meetings held in June 2003, August
2003, October 2003, December 2003, February 2004, and April 2004.

29In one case that resulted in a settlement agreement, the Board also
issued the lender a letter of reprimand.

Our analysis of the 32 cases further showed that the Board's effort to
review the cases and impose sanctions against lenders or to enter into
settlement agreements with them is frequently a time-consuming process. As
figure 13 shows, it took an average of 6.7 months from the notice of
violation to withdraw lenders' FHA approval and an average of 11.1 months
to reach settlement agreements. For the one case where the Board assessed
a civil money penalty, it took 27.6 months from the notice of violation to
complete the action. The process is even lengthier when considering the
time elapsed between the referral and the final Board action. It took an
average of 10.3 months from the referral to withdraw lenders' FHA approval
and an average of 17.7 months to reach settlement agreements. It took 34.8
months from the referral to assess one lender a civil money penalty.

Source: GAO analysis of data provided by the Mortgagee Review Board.

The length of time required by the Board to complete its actions allowed
nine of the lenders to continue making FHA-insured loans for over a year
without being held accountable for their violations. For example, in April
2003, the Board sent a notice of violation to one of these lenders because
the lender had committed several violations, including allowing non-FHA
approved entities to originate loans, failing to properly verify borrower
information, and charging excessive and/or unallowable fees. By the time

the Board withdrew the lender's FHA approval in April 2004, the lender had
made 584 additional FHA-insured mortgage loans.

The length of time required for the Board to withdraw lenders' FHA
approval has improved somewhat since our April 2000 report. We found then
that, for the six cases completed during our time frame for that report,
it took an average of 8.5 months from the notice of violation to withdraw
lenders' FHA approval (as opposed to 6.7 months for the completed cases we
reviewed in June 2004).30 The length of time required for settlement
agreements is about the same. We reported in April 2000 that it took an
average of 11.2 months to reach the five completed settlement agreements,
which is similar to the 11.1 months for the completed cases we reviewed in
June 2004. At that time, Board officials told us that they had taken some
steps to speed up the process. For example, the Board's secretary told us
that in December 1998 the Board had adopted a policy of meeting every 2
months to consider case referrals. This official told us that prior to
adopting this policy, the Board did not have an established meeting
schedule and met only whenever a sufficient number of cases had
accumulated for review. Also, to speed up the settlement agreement
process, the Board planned in future violation letters to ask the lenders
whether they would be willing to settle their cases and, if so, under what
terms and conditions.

HUD officials recognize that the Mortgagee Review Board process can be
time-consuming and have taken some steps to speed it up. They noted that
progress on certain cases can be slowed when HUD's Office of Inspector
General requests that the Board place a hold on the processing of cases
against lenders until the Inspector General has completed its
investigations. During fiscal years 2003 and 2004, the Inspector General
asked for a hold on 14 cases. In accordance with its plans, the Board has
started asking lenders prior to Board meetings whether they would be
willing to settle their cases. If a lender's settlement offer is
acceptable to the Board, a settlement agreement can be prepared and signed
immediately. If a lender's offer is not acceptable, the Board can then
make its own proposal for settling the case. Also, the Board has hired new
staff and plans to implement an internal quality control process by the
end of calendar year 2004.

30We reviewed 24 cases involving single-family housing lenders that the
Board acted on from October 1998 through April 1999. As of November 1999,
the Board had completed action on 11 of the cases, while action was still
pending on 13 cases.

Conclusions	FHA insures tens of billions of dollars in mortgages for
single-family homes each year. While FHA's Mutual Mortgage Insurance Fund
is financially healthy, poor lending practices could adversely affect the
Fund's financial position. Because lenders underwrite virtually all
FHA-insured mortgages without HUD's prior review, it is essential that HUD
allow only qualified lenders to participate in its single-family programs,
adequately monitor lenders and loans to assess the risks they pose, and
hold lenders accountable for the quality of the loans they make.

Contrary to HUD's guidance, the homeownership centers have granted direct
endorsement authority to some lenders that did not demonstrate performance
that is acceptable under HUD's criteria. Ensuring that the lenders that
are being granted direct endorsement authority demonstrate acceptable
performance would better assure HUD that these lenders are qualified to
underwrite loans without HUD's prior review. By not following HUD's
guidance, the homeownership centers may be exposing the department to
unreasonable insurance risks.

The homeownership centers generally use a risk-based approach to
monitoring, which helps them to focus on the lenders and loans posing a
high insurance risk to the department. However, HUD's oversight of lenders
could be improved. First, while the homeownership centers are targeting
high-risk lenders for examination, they do not always conduct these
reviews on-site at the lenders' offices, even though HUD's guidance states
that on-site reviews are the preferred method of monitoring. Tracking the
lender review process to distinguish between desk and on-site reviews
would help HUD ensure that the majority of high-risk lender reviews
continue to be conducted on-site. Second, contrary to HUD's guidance, the
homeownership centers have not consistently targeted for technical reviews
loans made by newly approved lenders. This lack of consistency is due in
part to weaknesses in the information system (CHUMS) HUD uses to select
loans for technical reviews. Third, the ratings that are assigned during
technical reviews do not currently reflect the different levels of risk
that underwriting errors pose to the insurance fund. The homeownership
centers have proposed revisions to the rating system, which, if
implemented, may better ensure the usefulness of technical reviews in
identifying the lenders that pose the greatest insurance risk to the
department. Because of these conditions, HUD's lender reviews and
technical reviews are not as effective as they could be in mitigating
financial losses to the department.

HUD has not taken sufficient steps to hold lenders accountable for poor
performance and program violations. HUD's reviews show that numerous
lenders did not comply with FHA's underwriting requirements, yet HUD's
homeownership centers have suspended the direct endorsement authority of
relatively few lenders. Even though the technical review rating system as
currently implemented does not properly reflect the risk that different
underwriting errors pose to the insurance fund, the extent of lender
noncompliance revealed by the reviews indicates that more lenders may be
candidates for enforcement action. By failing to suspend poorly performing
lenders, HUD leaves itself vulnerable to lending practices that increase
the department's insurance risk.

  Recommendations for Executive Action

To improve HUD's oversight of FHA mortgage lenders, we recommend that the
Secretary of HUD direct the Assistant Secretary for Housing-Federal
Housing Commissioner to take the following five actions:

o 	Ensure that the homeownership centers are following the guidance for
granting direct endorsement authority.

o  Track lender reviews to distinguish between desk and on-site reviews.

o 	Enhance FHA's information system to ensure that the first 30 loans made
by new direct endorsement lenders are reviewed as required.

o 	Expeditiously complete efforts to revise the technical review rating
system so that the ratings better reflect the risks that different
underwriting errors pose to the FHA insurance fund.

o 	Develop and implement guidance specifying the conditions under which a
homeownership center must suspend a lender's direct endorsement authority.

Agency Comments and 	We provided a draft of this report to HUD for its
review and comment. In a letter from the Assistant Secretary for
Housing-Federal Housing

Our Evaluation	Commissioner (see app. II), HUD agreed with our five
recommendations. Specifically, HUD:

o 	agreed that it would update its guidance for granting direct
endorsement authority and ensure that the homeownership centers
consistently apply the requirements for granting direct endorsement
authority;

o 	stated that its lender tracking system had been modified to distinguish
between desk and on-site reviews;

o 	noted that it was aware of the limitations of its information system to
accurately identify and count the first 30 loans made by new direct
endorsement lenders, and was pursuing system enhancements to ensure that
loans made by new direct endorsement lenders are reviewed as required;

o 	stated that it was in the final phase of implementing a new technical
review rating system, to be completed by January 2005; and

o 	stated that it was currently developing consistent standards for
returning lenders to preclosing status (i.e., suspending lenders' direct
endorsement authority) and that implementation would occur by January
2005.

HUD also highlighted its program accomplishments, commenting that, while
the report acknowledges that FHA has implemented policy and procedural
changes since our last review, it does not fully recognize the substantial
achievements and accomplishments that resulted from these changes. The
department described improvements in lender monitoring, lender approval
and recertification, and lender training, among other things. While we
agree that HUD has made improvements, a number of the accomplishments
cited were outside the scope of our review. Our draft report recognized a
number of specific improvements that were relevant to our objectives. For
example, we noted that HUD had revised its guidance for granting direct
endorsement authority and started targeting lenders and loans for review
based on risk.

HUD also disagreed with some of our findings. First, HUD stated that our
discussion of the percentages of lenders targeted and reviewed (figs. 5
and

6) did not acknowledge that target reports developed each quarter to
identify the lenders to be reviewed are fluid and decisions regarding
which lenders to review are subject to outside influences. For example,
staff may be directed to complete other reviews based on information from
sources other than Neighborhood Watch. HUD also noted that the reviews not
completed during the quarter are carried over to the subsequent quarter

and nearly all are completed. We included figures 5 and 6 in our draft
report as empirical evidence supporting our statement that the
homeownership centers generally were reviewing the lenders on their
targeting list. Further, our draft report-in the text preceding figure
4-acknowledged that the centers used other information to develop their
targeting lists, specifically citing examples such as the existence of
complaints about lenders. Nevertheless, we added to this text based on
HUD's comments.

Second, regarding the Mortgagee Review Board, HUD noted that the Board
carries out a highly regulated administrative process that may lead to
serious sanctions and penalties for lenders; therefore, the department is
obligated to afford maximum due process to these lenders. HUD also stated
that our conclusion that the length of time it takes for the Board to act
on a case allows a lender to continue with inappropriate practices is
misleading because (1) HUD staff communicate the findings of lender
reviews to lenders during mandatory exit conferences and (2) the Board has
the ability to suspend a lender or move a case quickly through the process
when significant problems are found. We do not agree that the statement is
misleading. Our report stated that the length of time required by the
Board to complete its actions allowed nine lenders to continue making
FHA-insured loans for over a year without being held accountable for their
violations. While HUD may notify a lender of violations found during a
lender review, such notification does not guarantee that the lender will
choose to correct or improve its practices. Finally, although the Board
has the ability to suspend lenders, it did not choose to suspend the nine
lenders we highlighted in our report.

HUD agreed with our statement that HUD's preferred method for monitoring
is to conduct on-site reviews of lenders. It also noted that the
difference between a desk review and an on-site review is minimal, but
went on to say that its tracking system had been modified to distinguish
between desk and on-site reviews. Finally, HUD stated that it anticipates
publication of the Credit Watch regulation discussed in the report by the
second quarter of fiscal year 2005.

We are sending copies of this report to the Secretary of Housing and Urban
Development and to other interested congressional committees. We also will
make copies available to others upon request. In addition, the report will
be available at no charge on the GAO Web site at http://www.gao.gov.

If you or your staff have any questions regarding this report, please
contact me at (202) 512-8678 or [email protected] or Paul Schmidt at (312)
220-7681 or [email protected]. Staff contacts and other key contributors
are listed in appendix III.

David G. Wood Director, Financial Markets and Community Investment

Appendix I

                       Objectives, Scope, and Methodology

Our objectives were to examine (1) how well the Department of Housing and
Urban Development (HUD) follows its guidance when granting lenders direct
endorsement authority, (2) the extent to which HUD uses a riskbased
approach when monitoring the lenders participating in the Federal Housing
Administration's (FHA) mortgage insurance programs, and (3) the extent to
which HUD is holding lenders that it identifies as not complying with its
requirements accountable for their performance.

To determine how well HUD follows its guidance when granting lenders
direct endorsement authority, we reviewed HUD's regulations, procedures,
and other guidance relating to its process for approving lenders and
granting lenders direct endorsement authority. Lenders with direct
endorsement authority can underwrite and close FHA-insured mortgage loans
without prior FHA review or approval. We interviewed officials from HUD's
Office of Lender Activities and Program Compliance, Office of Single
Family Program Development, and its four homeownership centers. We
requested from each homeownership center the number of lenders that
entered the probationary period on or after October 1, 2002, and were
granted direct endorsement authority by April 30, 2004. We chose October
1, 2002, as the start date because HUD revised its guidance for granting
lenders direct endorsement authority in September 2002. We used April 30,
2004, as our ending date because we visited the four homeownership centers
in May and June 2004. For each of the 49 lenders that were approved during
this time period, we reviewed documentation maintained by the centers and
entered the ratings that the lender received on the mortgages it submitted
to the center into a data collection instrument.1 All of the data
collected was independently verified. We then analyzed the data to
determine whether the centers followed FHA's procedures for granting
lenders direct endorsement authority.

To determine the extent to which HUD is using a risk-based approach when
monitoring lenders, we reviewed HUD's guidance and procedures for
conducting lender reviews (i.e., reviews of lenders' operations by HUD
staff) and technical reviews (i.e., reviews of individual loans performed
after approval of mortgage insurance to assess the quality of lenders'
underwriting practices). We reviewed HUD's use and oversight of
contractors that perform technical reviews. We interviewed officials at

1One additional lender, which already had direct endorsement authority,
was granted the authority to underwrite Home Equity Conversion Mortgages
(HECM)-mortgages that can be used by senior homeowners to convert equity
into income.

Appendix I
Objectives, Scope, and Methodology

each of the centers on a variety of issues dealing with lender reviews and
technical reviews. The issues discussed included the (1) centers' criteria
for targeting loans and lenders for review, (2) number of off-site lender
reviews, (3) the number and type of "poor" ratings assigned during
technical reviews, and (4) procedures for monitoring the work of technical
review contractors. We also interviewed Fannie Mae and Freddie Mac to
discuss their efforts to monitor the lenders that participate in their
programs.

In addition to these steps, we obtained and analyzed lists of the lenders
that the homeownership centers targeted for lender reviews in fiscal year
2003 and the first two quarters of fiscal 2004 and the lenders they
reviewed during the same time period. (We analyzed six quarters of data to
ensure that the most recent data was considered.) We compared the lenders
reviewed with those that were targeted to assess the extent to which HUD
was performing lender reviews on lenders that it considered to pose a high
risk to the insurance fund. We also analyzed the data to determine the
percentage of targeted lenders that were reviewed.

To determine the extent to which the homeownership centers were reviewing
the first 30 loans made by new direct endorsement lenders, we analyzed
data from HUD's Single Family Data Warehouse on the loans endorsed by the
49 direct endorsement lenders that entered the probationary period on or
after October 1, 2002, and were granted direct endorsement authority by
April 30, 2004. Only 16 of the 49 lenders had endorsed loans as of June
19, 2004.

To determine the percentage of "poor" ratings each lender received during
technical reviews, we analyzed data from HUD's Underwriting Reports System
for fiscal year 2003 and the first two quarters of fiscal 2004. When we
requested the data from each homeownership center, the Philadelphia and
Santa Ana centers provided just the technical reviews where the review
date was within our six-quarter time frame. In contrast, the Atlanta and
Denver centers provided databases that included reviews outside our scope
(including reviews where the review date was blank). To be consistent with
the data that the other two centers provided, we limited our analysis of
the data provided by Atlanta and Denver to just those technical reviews
where the review date was within our six-quarter time frame. We used this
data to determine the percentage of "poor" ratings assigned during
technical reviews and the most common deficiencies cited during technical
reviews.

Appendix I
Objectives, Scope, and Methodology

To determine the extent to which HUD is holding lenders that it identifies
as not complying with its requirements accountable for their performance,
we reviewed HUD's regulations and policy guidance to determine the
enforcement options available to HUD. We interviewed officials from HUD's
Office of Lender Activities and Program Compliance, Enforcement Center,
and Mortgagee Review Board. At each of the four homeownership centers, we
discussed with cognizant officials each center's efforts to take
enforcement actions against lenders that have violated program
requirements. We determined the number and types of lenders sanctioned by
HUD under its Credit Watch program as of the end of July 2004. In June
2004, we reviewed the Board's files for the 32 cases involving
single-family mortgage lenders that the Board had acted on in the previous
12 months and determined the nature and status of the Board's actions. In
addition, we analyzed data to determine the length of time it took the
Board to take action against these lenders.

To assess the reliability of the various HUD data we used, we discussed
the data with knowledgeable agency officials, reviewed information about
the systems, and performed electronic testing to detect obvious errors in
completeness and reasonableness. We determined that these data were
sufficiently reliable for the purposes of this report.

We performed our work from December 2003 to September 2004 in accordance
with generally accepted government auditing standards.

Appendix II

Comments from the Department of Housing and Urban Development

Appendix II
Comments from the Department of Housing
and Urban Development

Appendix II
Comments from the Department of Housing
and Urban Development

Appendix II
Comments from the Department of Housing
and Urban Development

Appendix II
Comments from the Department of Housing
and Urban Development

Appendix II
Comments from the Department of Housing
and Urban Development

Appendix II
Comments from the Department of Housing
and Urban Development

                               Now on pp. 15-16.

Now on p. 17.

Appendix II
Comments from the Department of Housing
and Urban Development

                               Now on pp. 33-34.

Appendix II
Comments from the Department of Housing
and Urban Development

Appendix III

                     GAO Contacts and Staff Acknowledgments

GAO Contacts	David Wood (202) 512-8678 Paul Schmidt (312) 220-7681

Staff 	In addition to those individuals named above, Eric Diamant, Mark
Egger, Harold Fulk, Nadine Garrick, Curtis Groves, John McGrail, Marc
Molino,

Acknowledgments	Josephine Perez, David Pittman, Terry Richardson, Paige
Smith, and Raymond Wessmiller made key contributions to this report.

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 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, GAO postsGAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To Testimony have GAO e-mail you a list of newly posted products
every afternoon, go to

www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone:	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

  To Report Fraud, Contact:
  Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Congressional	Gloria Jarmon, Managing Director, [email protected] (202)
512-4400 U.S. Government Accountability Office, 441 G Street NW, Room 7125

Relations Washington, D.C. 20548

Public Affairs	Susan Becker, Acting Manager, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548
*** End of document. ***