Residential Care Facilities Mortgage Insurance Program: 	 
Opportunities to Improve Program and Risk Management (24-MAY-06, 
GAO-06-515).							 
                                                                 
Through its Section 232 program, the Department of Housing and	 
Urban Development's (HUD) Federal Housing Administration (FHA)	 
insures approximately $12.5 billion in mortgages for residential 
care facilities. In response to a requirement in the 2005	 
Consolidated Appropriations Conference Report and a congressional
request, GAO examined (1) HUD's management of the program,	 
including loan underwriting and monitoring; (2) the extent to	 
which HUD's oversight of insured facilities is coordinated with  
the states' oversight of quality of care; (3) the financial risks
the program poses to HUD's General Insurance/Special Risk	 
Insurance (GI/SRI) Fund; and (4) how HUD estimates the annual	 
credit subsidy cost for the program.				 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-515 					        
    ACCNO:   A54683						        
  TITLE:     Residential Care Facilities Mortgage Insurance Program:  
Opportunities to Improve Program and Risk Management		 
     DATE:   05/24/2006 
  SUBJECT:   Mortgage programs					 
	     Mortgage protection insurance			 
	     Nursing homes					 
	     Program management 				 
	     Risk management					 
	     Subsidies						 
	     Financial analysis 				 
	     Cost estimates					 
	     HUD General and Special Risk Insurance		 
	     Fund						 
                                                                 

******************************************************************
** 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-06-515

     

     * Report to Congressional Addressees
          * May 2006
     * RESIDENTIAL CARE FACILITIES MORTGAGE INSURANCE PROGRAM
          * Opportunities to Improve Program and Risk Management
     * Contents
          * Results in Brief
          * Background
          * HUD's Decentralized Management Provides Field Offices
            Flexibility, but Varying Awareness of Underwriting and Monitoring
            Practices and Concerns Over Insufficient Staff Expertise Increase
            Program's Potential Risks
               * Some Field Offices Were Not Aware of All Current Program
                 Requirements
               * Field Offices Do Not Systematically Share Information on
                 Practices
               * Officials Cited Concerns about Adequate Levels of Staff
                 Expertise
          * FHA's Coordination with States' Oversight of Quality of Care for
            Section 232 Residential Care Facilities Is Limited
               * FHA Requires Some Coordination with States' Oversight of
                 Quality of Care for Section 232 Residential Care Facilities
               * FHA's Limited Coordination with States on Oversight Issues
                 May Lead to Missed Identification of Risk Indicators
          * Program and Industry Trends Show Sources of Potential Risks to
            the GI/SRI Fund
               * The Section 232 Program Represents a Small Percentage of the
                 GI/SRI Fund
               * Program Trends Show Sources of Potential Risk
                    * Higher Claim Rates for Recent Loan Cohorts
                    * Changes in Claim Rates by Loan Purpose
                    * Changes in Claim Rates by Facility Type
                    * Increase in Loan Prepayments
                    * Concentration of Loans
               * Industry Faces Uncertainties
               * FHA Uses a Number of Tools to Mitigate Risks
          * HUD's Model for Estimating Credit Subsidy Costs Excludes Some
            Potentially Relevant Factors
               * HUD Uses a Model to Estimate Credit Subsidy Costs
               * Features of the Credit Subsidy Model May Lead to Unreliable
                 Credit Subsidy Estimates
                    * Prepayment Penalties or Restrictions
                    * Debt Service Coverage Ratio at Point of Loan
                      Origination
                    * Loan-to-Value Ratio at Point of Loan Origination
                    * Types of Facilities Insured and Changes in Interest
                      Rates
                    * Use of Proxy Data
          * Conclusions
          * Recommendations for Executive Action
          * Agency Comments and Our Evaluation
     * Objectives, Scope, and Methodology
     * Information on the Application Processing, Underwriting, and Oversight
       of Section 232 Loans
          * Application Processing and Underwriting for Section 232 Loans
          * Oversight and Monitoring of Section 232 Loans
     * HUD Officials Addressed Some Issues Raised by the Agency's Inspector
       General in 2002, but Several Key Items Remain Unresolved
     * HUD's Use of Proxy Data for Refinance Loans
          * HUD's Comparison Did Not Allow for Differences in the Age of
            Loans
          * Substantial Differences Exist between Section 207 and Section 232
            Loans
     * Comments from the Department of Housing and Urban Development
     * GAO Contact and Staff Acknowledgments
     * PDF6-Ordering Information.pdf
          * Order by Mail or Phone

Report to Congressional Addressees

May 2006

RESIDENTIAL CARE FACILITIES MORTGAGE INSURANCE PROGRAM

Opportunities to Improve Program and Risk Management

Contents

Figures

May 24, 2006Letter

Congressional Addressees

Through its Section 232 Mortgage Insurance for Residential Care Facilities
program (Section 232 program), the Department of Housing and Urban
Development's (HUD) Federal Housing Administration (FHA) insures mortgages
for nursing homes, assisted living facilities, board and care homes, and
intermediate care facilities. As of December 31, 2005, the program insured
mortgages with an unpaid principal balance of approximately $12.5
billion.1 The program insures HUD-approved private lenders against
financial losses from loan defaults; insured loans can be used to finance
the purchase, construction, or rehabilitation of a facility, enable
borrowers to refinance projects that do not need substantial
rehabilitation, or to install fire safety equipment.

For budget and accounting purposes, the Section 232 program is part of
HUD's General Insurance/Special Risk Insurance (GI/SRI) Fund; other
programs in the GI/SRI Fund provide mortgage insurance for various types
of multifamily housing projects and for hospitals. HUD is required to
annually estimate the subsidy cost, or the cost to the federal government
of guaranteeing credit to residential care facilities over the life of the
loans.2 This estimate requires FHA to forecast future cash flows
associated with the loans, which can be influenced by factors that are
associated with the potential risks facing the program's loan portfolio.

While private lenders may finance the purchase or construction of nursing
homes, public funding, including Medicare and Medicaid, has accounted for
an increasing percentage of spending on nursing home care.3 For example,
in 2000 Medicare and Medicaid financed 39 percent of the nation's

spending on nursing home care, up 28 percent from 1990.4 In 2004, Medicare
accounted for 14 percent and Medicaid accounted for 44 percent of the
nation's spending on nursing home care, and the total of all public funds,
including Medicare and Medicaid, accounted for approximately 61 percent.5
Federal and state governments share responsibility for oversight of
nursing homes that participate in the Medicare and Medicaid programs. The
U.S. Department of Health and Human Services defines standards that
nursing homes must meet to participate in the Medicare and Medicaid
programs and contracts with states to conduct annual inspections.6
Generally, states license nursing homes (and in some cases related
facilities) and oversee their operations through inspections.

The 2005 Consolidated Appropriations Conference Report mandated that we
review the design and management of two FHA mortgage insurance programs
-those for the Section 232 program and the Section 242 Hospital Mortgage
Insurance program.7 In addition, the Ranking Member of the Subcommittee on
Housing and Transportation, Senate Committee on Banking, and others
requested that we review several aspects of the Section 232 program.
Accordingly, this report provides both the results of the mandated review
and our response to the request. Specifically, we examined: (1) HUD's
management of the program, including loan underwriting and monitoring; (2)
the extent to which HUD's oversight of insured residential care facilities
is coordinated with the states' oversight of the quality of care provided
by those facilities that are subject to state licensing or inspection; (3)
the financial implications of the program to the GI/SRI Fund, including
risk posed by program and market trends; and (4) how HUD estimates the
annual credit subsidy for the program, including the factors and
assumptions used. In addition, we examined HUD's action in response to a
HUD Inspector General report that concluded that HUD's Office of Housing
did not have adequate controls to effectively manage the Section 232
program; this information is summarized in appendix III.

To address these objectives, we reviewed program manuals and documentation
of loan processing procedures and underwriting requirements and analyzed
program financial data that we tested and found reliable for our
purposes.8 In examining HUD's management of the program, we focused on how
underwriting and loan monitoring activities were carried out through
visits to five of HUD's field offices (Atlanta, Georgia; Buffalo, New
York; Chicago, Illinois; Los Angeles, California; and San Francisco,
California), where we interviewed program officials and obtained relevant
documents in each office.9 We also reviewed documentation of the model HUD
uses to estimate program subsidy costs, applicable program laws,
regulations, and policy statements. We obtained relevant program
documentation and interviewed headquarters officials in HUD's Office of
Multifamily Development and Office of Asset Management and HUD's Office of
Evaluation and Office of Inspector General. We also interviewed
representatives of residential care associations; lenders with loans
insured by the program, as well as other private lenders that offer
non-FHA-insured residential care loans; and representatives of nursing
homes and assisted living facilities. Our review did not include an
evaluation of the need for the Section 232 program. See appendix I for
more detailed information on our objectives, scope, and methodology.

We conducted our work in Washington, D.C., and the HUD field office
locations noted above between February 2005 and April 2006 in accordance
with generally accepted government auditing standards.

Results in Brief

While HUD's decentralized management of the Section 232 program allows
field offices some flexibility in their specific practices, in our visits
to five field offices we found a lack of awareness of some current program
requirements, potentially useful loan underwriting and monitoring
practices developed in individual offices that were not systematically
shared with other offices, and concerns by field office managers about
current or future levels of staff expertise. For example, four of the
field offices were not aware of a notice that disqualifies potential
Section 232 borrowers if they have had a bankruptcy in their past, and
four offices were unaware of required addendums to the programs' standard
regulatory agreement regarding certain state licensing requirements for
nursing homes. While individual offices had developed useful practices for
implementing the program's loan underwriting and monitoring requirements,
they lacked a mechanism for systematically sharing practices with other
offices. For example, two field offices included asset management
staff--persons that monitor and oversee a loan after it has been
insured-in the underwriting stages of loans to better assess their risks,
while the other field offices did not. We also found that field office
officials were concerned about adequate current or future levels of staff
expertise-a critical factor in avoiding unwarranted risk in the Section
232 program, in that health care facility loans are generally more
complicated and require more specialized knowledge and expertise compared
with loans insured under HUD's other multifamily programs. Lack of
awareness of current requirements and insufficient staff expertise can
contribute to the program insuring loans with increased risks.

FHA's coordination with states' oversight of residential care facilities'
quality of care provided to residents is limited. FHA requires field
office officials to review the most recent annual state administered
inspection report for existing state-licensed facilities as part of the
application. HUD recommends, but does not require, that officials continue
monitoring annual inspection reports for a residential care facility once
it is insured, particularly if the officials do not perform an on-site
management review (an examination of operations, occupancy, financial
management, and possible quality of care issues) of the facility. Four of
the five HUD field offices we visited do not routinely review annual
inspection reports for the insured facilities they oversee; further, HUD
field offices conduct a limited number of management reviews of Section
232 facilities. While annual inspection reports are but one of several
means of monitoring insured properties, FHA's limited use of them may lead
the agency to miss potential indicators of risk for some of its insured
loans. Because serious quality of care deficiencies can have a variety of
implications that affect cash flow streams, ranging from a related
reduction in occupancy to the more direct financial implications such as
civil money penalties and loss of licensing and reimbursements, they may
ultimately affect a facility's ability to repay the loan. Some private
lenders told us they use the annual state inspection reports in
coordination with other financial indicators to assess the financial risk
of loans to facilities subject to state inspections. HUD is in the process
of revising its residential care facility regulatory agreements-which
establish loan conditions applicable to an owner and potential operator-to
require owners or operators to (1) submit to HUD annual inspection reports
and (2) report to HUD any notices of inspection violations. However, the
proposed revisions have been awaiting approval since August 2004, and it
is not clear when the revised agreements will be approved.

Although the Section 232 program is a small component of the GI/ SRI
Fund-representing approximately 16 percent of the total unpaid principal
balance-program and industry trends may pose financial risks to the fund.
For example:

o In recent years, HUD has insured increasing numbers of mortgages that
are refinances of existing loans, as well as loans for assisted living
facilities. Because these types of loans are relatively new to the
portfolio, there are limited data to observe long-term claim trends,
making their risk difficult to assess. However, the 5-year claim rate (the
portion of loans leading to a claim within 5 years of origination) was
significantly higher for more recent assisted living facility loans.

o The proportion of loans that terminate due to prepayment within 10 years
of origination is increasing. Prepayment occurs when a borrower pays a
loan in full before the loan reaches maturity. As more borrowers prepay
their loans, HUD loses future cash flows of premiums. Such losses could be
offset to some extent, in that prepayments may ultimately result in fewer
claims.

o Program loans are concentrated in several states and among relatively
few lenders. As of 2005, five states (California, Illinois, Massachusetts,
New York, and Ohio) accounted for 51 percent of active loan amounts, with
one state-New York-representing 24 percent. Further, while a total of 109
lenders held active loans, just 6 held over half of the active loan
portfolio. Geographic concentration makes the program vulnerable to swings
in regional economic conditions, while concentration among lenders
potentially makes the program more vulnerable if one or a few large
lenders encounters financial difficulty.

In addition, industry developments and uncertainty in the funding of the
Medicaid and Medicare programs pose potential risks. Projected shifts in
demand for residential care facilities could affect not only the
facilities currently insured by HUD but also the overall market for the
particular types of facilities that HUD insures under the program.

To estimate the subsidy cost of the Section 232 program, HUD uses a cash
flow model to project the expected cash flows for all of these loans over
their entire life. The cash flow model uses assumptions based on
historical and projected data to estimate the amount and timing of claims,
subsequent recoveries from these claims, prepayments, and premiums and
fees paid by the borrower. We found that HUD's model does not explicitly
consider certain factors such as loan prepayment penalties or lockouts
(the period of time during which prepayment is prohibited), which can
affect whether and when a loan is prepaid and may also show changes in the
risk of claim and expected collections of premiums. HUD's cash flow model
also does not fully capture the effects on existing loans when market
interest rates change, nor explicitly consider differences in loan
performance between different types of facilities. Furthermore, the model
includes some proxy data with borrower characteristics and performance
that is not comparable to Section 232 loans. The model's exclusion of
potentially relevant factors and its use of this proxy data could affect
the reliability of HUD's credit subsidy estimates.

This report contains recommendations to HUD designed to ensure that field
offices understand and implement current program requirements, including
sharing practices among field offices. We also recommend that HUD
incorporate reviews of annual inspection reports for nursing homes and
other residential care facilities into its loan monitoring process,
complete its proposed revision to the residential care facility regulatory
agreement in a timely manner, and consider including additional variables
and methods in its credit subsidy modeling. We provided a draft of this
report to HUD and received written comments from the Assistant Secretary
for Housing, which are discussed later in this report and in appendix V.
In its response, HUD generally concurred with our recommendations intended
to ensure that field offices are aware of and implement current program
requirements and policies, but disagreed with most parts of our
recommendation related to HUD's credit subsidy model. Specifically, HUD
did not agree to consider factoring additional information into its credit
subsidy model including prepayment penalties and restrictions, initial
loan-to-value and debt service coverage ratios, and the ratio of contract
rates and market rates. Because we believe that factoring such information
into the credit subsidy model could be useful, we did not modify our
recommendation.

Background

Section 232 of the National Housing Act, as amended, authorizes FHA to
insure mortgages made by private lenders to finance the construction or
renovation of nursing homes, intermediate care facilities, board and care
homes, and assisted living facilities.10  Congress  established the
Section 232 program in 1959 to provide mortgage insurance for the
construction and rehabilitation of nursing homes. The Housing and
Community Development Act of 1987 expanded the program to allow for the
insuring of refinancing or purchase of FHA-insured facilities and, in
1994, HUD issued regulations implementing legislation to expand the
program to allow for the insuring of assisted living facilities and the
refinancing of loans for facilities not previously insured by FHA. Since
1960, FHA has insured 4,372 loans through the Section 232 program in all
50 states, the District of Columbia, the U.S. Virgin Islands, and Puerto
Rico. As of the end of fiscal year 2005, there were 2,054 currently
insured loans.

FHA does not insure all residential care facilities, as there are
approximately 16,500 nursing home facilities and over 36,000 assisted
living facilities in operation.11 We did not identify any private mortgage
insurance that is currently available for loans made to nursing homes or
other similar facilities. According to HUD officials, in recent times, the
Section 232 program exists, in part, to support the market for residential
care facilities when the private market is reluctant to finance such
projects due to market conditions. The loans are advantageous to borrowers
because they are nonrecourse loans whereby the lender (in this case the
lender and the insurer, FHA) has no claim against the borrower in the
event of default and can only recover the property. The loans are also
generally long term (in some cases up to 40 years) and, according to HUD
and lender officials, offer an interest rate that is, in many cases, lower
than what private lenders offer for non-FHA insured loans made to nursing
homes and other similar facilities. Additionally, FHA insures 99 percent
of the unpaid principal balance plus accrued interest.

HUD administers the Section 232 program through its field offices, with
HUD headquarters oversight. HUD's field structure consists of 18 Hub
offices and 33 program centers. Generally, each Hub office has a number of
program centers that report to it. Program centers administer multifamily
programs within the states in which they are located or portions thereof.
Hub offices also administer multifamily programs, as well as augment the
operations of and coordinate workload between their program centers.

Under Medicaid, states set their own nursing home payment rates
(reimbursement rates), and the federal government provides funds to match
states' share of spending as determined by a federal formula. Within broad
federal guidelines, states have considerable flexibility to set
reimbursement rates for nursing homes that participate in Medicaid but are
required to ensure that payments are consistent with efficiency, economy,
and quality of care.12 Under Medicare, skilled nursing facilities receive
a federal per diem payment that reflects the resident's care needs and is
adjusted for geographic differences in costs.

HUD's Decentralized Management Provides Field Offices Flexibility, but
Varying Awareness of Underwriting and Monitoring Practices and Concerns
Over Insufficient Staff Expertise Increase Program's Potential Risks

While the decentralization of the program allows field offices some
flexibility in their specific practices, the results of our visits to five
field offices revealed differences in the extent to which field office
staff were aware of current program requirements. Further, while
individual offices had developed useful practices for implementing the
program's loan underwriting and monitoring requirements, they lack a
mechanism for systematically sharing practices with other offices. We also
found that field office officials were concerned about adequate current or
future levels of staff expertise-a critical factor in avoiding unwarranted
risk in the Section 232 program, in that health care facility loans are
generally more complicated and require specialized expertise compared with
loans insured under HUD's other multifamily programs. Lack of awareness of
current requirements and insufficient staff expertise can contribute to
insuring loans with increased risks. Both factors are related to
recommendations made in the HUD Office of Inspector General's 2002 report
that HUD has not fully addressed (see app. III for further information on
weaknesses identified by HUD's Inspector General).

Some Field Offices Were Not Aware of All Current Program Requirements

FHA has numerous underwriting requirements for loans insured under the
Section 232 program; for example, facilities must provide evidence of
market need; a (real estate) appraisal; and be in compliance with limits
on loan-to-value and debt service coverage ratios.13 FHA also requires a
variety of reviews for monitoring Section 232 loans. (Loan underwriting
and monitoring requirements, which can involve fairly complex reviews and
analyses, are described in more detail in app. II.)

According to HUD headquarters officials, the field offices that administer
the Section 232 program are required to follow all program statutes and
regulations, but the decentralization of the program allows field offices
some flexibility in their specific practices. For example, individual
field offices can designate how to staff the underwriting and monitoring
of Section 232 loans, depending on such factors as loan volume relative to
other multifamily programs, to fully utilize resources. HUD headquarters
provides guidance on program policies and requirements; when necessary,
reviews applications for certain types of loans, such as those submitted
by nonprofit entities; and provides technical assistance or additional
guidance and support if contacted by field offices. HUD headquarters staff
also conduct Quality Management Reviews, which are management reviews of
field offices administering HUD programs and services. For these reviews,
evaluators visit offices and coordinate subsequent reports. The process
also involves reporting the status of follow-up corrective actions. While
not focused on the Section 232 program, this process helps to oversee the
program by reviewing the management of the field offices that administer
it.

We found that the five field offices that we visited varied in their
understanding and awareness of policies related to the Section 232
program. For example, staff in two field offices said that their standard
regulatory agreement (that serves as the basic insurance contract and
spells out the respective obligations of FHA, the lender, and the
borrower) did not include language that would require operators of insured
facilities to submit financial statements on new loans.14 According to
officials at HUD headquarters, field offices should be using language
requiring these financial statements. HUD and most lender officials we
interviewed told us that operator financial statements provide information
on the legal entity operating the facility in cases where the borrower and
the operator of the residential care facility are different entities.
These officials also stated that, in such situations, borrower financial
statements may not disclose expenses, income, and other financial
information, and may only show the transactions between the borrower and
operator, thus making operator financial statements a necessity. Also,
HUD's Inspector General identified HUD's lack of a requirement for
operators to submit financial statements electronically to be part of an
internal control weakness for the Section 232 program.

Additionally, we found that the field offices that we visited were not
always aware of specific notices that established new requirements or
processes for the Section 232 program. For example:

o Four of the five field offices that we visited were not aware of a
notice that disqualifies potential Section 232 borrowers if they have had
a bankruptcy in their past. According to a HUD headquarters officials,
this policy is intended to protect HUD from insuring a potentially risky
loan based on a borrower's financial history.

o Officials at four of the five field offices we visited did not know
about required addendums to the regulatory agreement regarding state
licensing requirements for nursing homes. HUD developed these addendums to
place a lien on a property's operational documents, such as a Certificate
of Need and state licenses, to prevent operators from taking these
documents with them upon termination of a property's lease.15 Without
these documents, a facility may not be able to operate and, consequently,
the property's value would be greatly diminished.

According to HUD headquarters officials, HUD headquarters communicates
changes in the Section 232 program's policies and procedures to field
offices in a variety of ways besides sending formal notices. For example,
HUD headquarters also posts some notices on a "frequently-asked questions"
section of a Web site available to field offices, lenders, attorneys, and
others.16 HUD headquarters officials also conduct nationwide conference
calls with the field offices in which various HUD multifamily programs,
including the Section 232 program, are discussed. The conference calls are
conducted separately for loan development staff and asset management staff
that work, respectively, on the underwriting and monitoring of loans. HUD
headquarters officials stated that these conference calls provide a forum
to disseminate information to the field offices and for individual field
offices to discuss any issues, questions, or concerns regarding any
multifamily programs, including the Section 232 program.

HUD headquarters officials stated that they plan to address the lack of
awareness we observed by updating the "Multifamily Asset Management and
Project Servicing Handbook" to clarify current policies and requirements
for the Section 232 program. HUD is also planning to update the handbook
to address the 2002 HUD Inspector General report that identified that
HUD's current handbook was not specific to Section 232 nursing home
operations. However, HUD officials told us the updates to the handbook
would not be completed until the proposed revisions to the applicable
regulatory agreements have been approved. The proposed revisions have been
awaiting approval since August 2004, and it is not clear when the revised
agreements will be approved.

Field Offices Do Not Systematically Share Information on Practices

As discussed earlier, field offices have some flexibility in practices
that they use in administering the Section 232 program. In our visits to
five field offices, we found a variety of practices that could be useful
in the underwriting and monitoring of Section 232 loans if shared with
other field offices. However, currently, HUD does not have systematic
means by which to share this information among field offices.

Officials in two of the five field offices we visited identified specific
practices they had developed to carry out loan underwriting requirements.
For example:

o Asset management staff, whose focus is monitoring the performance of
loans that are already insured, are asked to review a variety of documents
submitted in the underwriting process, such as financial statements and
information on the occupancy of the facility.

o In one of the offices, staff members may contact relevant state
officials, just before the closing of a loan, to verify that the state has
not identified any quality of care deficiencies since the facility
submitted the application for mortgage insurance.

o Officials in one office stated that they conduct an additional review
before approving a loan application for mortgage insurance to ensure that
all required steps, such as mortgage credit analysis and valuation, have
been properly performed.

According to the officials in these two offices, it is necessary to take
these additional steps in order to adequately underwrite a loan under this
program. They stated that the additional steps result in the better
screening of loan risk and could result in the rejection of a risky loan
they might otherwise approve.

We found a similar variety of practices in the monitoring of Section 232
loans. In some cases, field offices we visited had taken additional steps
beyond those required by HUD. For example:

o While HUD requires a review of the annual financial statements of
insured facilities, two field offices that we visited require monthly
financial accounting reports from facilities either for the first year of
the loan or until the facility has reached stable occupancy.

o Two field offices had developed their own specialized checklists for
monitoring Section 232 loans. These checklists were specifically designed
for the oversight of residential care facility loans and included items
such as the facility's replacement reserve accounts and professional
liability insurance, among other items.

o One of the offices had established a Section 232 working group, where
underwriting and asset management staff met periodically to discuss loans
in the portfolio and issues related to the overall management of the
program in the field office. Additionally, three of the five field offices
we visited had specialized staff with expertise in overseeing residential
care facility loans. These were asset management staff whose primary or
sole responsibility was oversight of the Section 232 portfolio.

o While HUD headquarters officials stated that they do not require
management reviews of Section 232 facilities, three of the five field
offices we visited conducted management reviews on some part of their
Section 232 portfolios.

o One field office obtained the state annual inspection reports on its
Section 232 facilities on a regular basis.

According to officials in these offices, the unique characteristics
associated with residential care facilities make the additional measures
necessary.

Officials in field offices we visited that had developed these specific
practices stated that the practices result in better underwriting and
monitoring of loans and could potentially help to prevent claims. However,
HUD field offices do not have a systematic means by which to share
information with other field offices about practices they have developed.
While field office officials can raise concerns and issues through
conference calls with HUD headquarters officials, most explained that
these conference calls are not particularly designed for field offices to
share practices with other field offices. Officials in the five field
offices that we visited told us that they occasionally contact their
counterparts in other field offices regarding loan processing or asset
management questions or issues. Additionally, officials in some field
offices said that they occasionally see their counterparts at regional
lender conferences. However, aside from these forms of contact, there was
no systematic method by which to learn about other field office practices.
Consequently, officials in one field office are likely to be unaware of
additional steps or practices taken by another field office that are
intended to help officials improve underwriting or monitoring of Section
232 loans. Officials at all field offices that we visited told us that
they could benefit from the sharing of such practices regarding
underwriting and monitoring procedures established by different offices.

Officials Cited Concerns about Adequate Levels of Staff Expertise

Officials in two of the five field offices stated that a lack of expertise
on residential care facility loans, either in underwriting or loan
oversight, is a current concern in their office. They specifically noted a
lack of expertise in residential care facilities and their overall
management. Officials in all of the field offices that we visited stated
that additional training on Section 232 loans would be beneficial to
provide more knowledge and expertise, as there has been very little
Section 232-specific training. In its 2002 report, HUD's Office of
Inspector General also identified that field office project managers did
not have sufficient training on reviewing Section 232 loans and dealing
with the issues unique to Section 232 properties.

All of the private lenders we interviewed-those that offer non-FHA insured
loans to residential care facilities and face similar risks to FHA-had a
specialized group that conducted the underwriting of these loans. All of
the individuals that conducted the underwriting of these loans were part
of a health care lending unit that focused exclusively on loans made to
health care facilities. According to the lenders, they believed it was
necessary to have specialized staff underwriting such loans due to the
unique nature of lending money to a facility that was designed for a
residential health care business. Additionally, almost all of the private
lenders we interviewed had specialized staff that monitored their
residential care facility loans. According to lender staff we interviewed,
nursing home and assisted living facility loans require an understanding
of the market, trends, expenses, income, and other such unique
characteristics associated with these types of facilities.

While officials in only two of the five offices expressed concern about
the expertise of current staff, officials in all field offices we visited
stated that they are concerned about the ability to adequately staff the
Section 232 program in the next 5 years. They stated that as older staff
retire in the next 5 years or so, any expertise that such staff currently
have will take time to replace. All of the field offices that we visited
staffed the underwriting process for Section 232 loans similar to that of
other multifamily programs, based on workload and staff resources.
However, while two field offices assigned their Section 232 properties,
along with other multifamily properties, to general asset management staff
for oversight, three field offices designated specific staff to oversee
Section 232 properties. This was due to the latter field office officials'
belief, similar to that of the private lenders we interviewed, that the
properties require a certain level of knowledge and expertise associated
with residential care facilities. Expertise in Section 232 loans allows
for a better understanding of the distinct issues associated with
oversight of residential care facilities. In one of the offices that had
general asset management staff overseeing the portfolio, eight project
managers shared responsibility for monitoring Section 232 properties in
conjunction with other multifamily program properties. In contrast, in one
of the offices with staff designated specifically for the Section 232
program, one member of the asset management staff was responsible for the
entire Section 232 portfolio. Officials from the two field offices that
have experienced staff specialized in monitoring Section 232 loans stated
that they are concerned about losing their specialized staff over time and
acknowledged that they will need to find replacements in order to continue
to adequately monitor Section 232 loans. Their concern stems in part from
the fact that Section 232 facilities, unlike other multifamily properties,
require specialized knowledge and an understanding of the marketing,
trends, and revenue streams associated with residential care facilities.

According to officials in all of the field offices that we visited,
monitoring of Section 232 loans, when compared with other FHA-insured
multifamily programs, requires additional measures. Section 232 loans
contain a complex business component-the actual assisted living service or
the nursing service operating in a facility-making them different from
other multifamily programs that are solely realty loans. Consequently, for
Section 232 loans, field office officials monitor the financial health of
the business, including expenses, income, and other such items. Some field
office officials also stated that it is important to monitor the operator
to ensure that the facility is adequately managed. Additionally, some
field office officials stated that to ensure the facility is generating
enough income, they have to monitor Medicare and Medicaid reimbursement
rates, as well as occupancy rates.

According to HUD headquarters officials, as part of its overall strategic
human capital efforts, HUD is currently assessing the loss of human
capital in field offices over time. However, this effort is not focused on
the Section 232 program specifically but is intended to examine general
human capital issues and needs.

FHA's Coordination with States' Oversight of Quality of Care for Section
232 Residential Care Facilities Is Limited

FHA requires field office officials, when processing applications for
Section 232 mortgage insurance from existing state-licensed facilities to
review the most recent annual state-administered inspection report for the
facilities, but does not require the continued monitoring of annual
inspection reports for state-licensed facilities once it has insured them.
Four of the five HUD field offices we visited do not routinely collect
annual inspection reports for the insured facilities they oversee. While
such reports are but one of several means of monitoring insured
properties, FHA's limited use of them may lead the agency to overlook
potential indicators of risk for some of its insured loans.

FHA Requires Some Coordination with States' Oversight of Quality of Care
for Section 232 Residential Care Facilities

State inspections or surveys of residential care facilities may stem from
state licensing requirements or the facilities' participation in Medicare
or Medicaid. Nursing homes are state licensed, while states vary in their
licensing requirements for assisted living facilities. The Department of
Health and Human Services' Centers for Medicare & Medicaid Services
requires that nursing homes receiving Medicare and Medicaid funding be
federally certified, and all certified facilities are subject to annual
federal inspections administered by the states. State survey agencies,
under agreements between the states and the Secretary of Health and Human
Services, conduct the annual federally required inspections. To complete
the annual inspections, teams of state surveyors visit Medicare and
Medicaid participating facilities and assess compliance with federal
facility requirements, particularly whether care and services provided
meet the assessed needs of the residents. These teams also assess the
quality of care provided to residents of the facilities, looking at
indicators such as preventing avoidable pressure sores, weight loss, or
accidents. Overall, annual inspections provide a regular review of quality
of care by officials with relevant backgrounds, such as, registered
nurses, social workers, dieticians, and other specialists. For facilities
that are applying for mortgage insurance under the Section 232 program,
FHA requires a copy of the state license needed to operate the facility
and a copy of the latest state annual inspection report on the facilities'
operation.

HUD's "Multifamily Asset Management and Servicing Handbook" recommends
that, once nursing home loans are insured under the program, HUD officials
responsible for loan monitoring continue to review state annual inspection
reports if they do not undertake management reviews of the facility.
Management reviews focus on an insured facility's financial indicators and
general management practices, but, particularly if conducted on-site,
could provide some information on issues related to the quality of care at
a facility. Because of their wider scope, however, management reviews
would not likely go into the same depth on quality of care issues as
annual inspections. HUD headquarters officials told us that the handbook's
recommendation applies to all Section 232 facilities; further, HUD
headquarters officials stated that management reviews for Section 232
properties should be conducted based on need and available resources. We
found that two of five field offices we visited did not regularly conduct
any regular management reviews and did not review annual inspection
reports during loan monitoring. Of the three field offices that did
conduct management reviews on some Section 232 properties, one also
reviewed annual inspection reports during loan monitoring. Additionally,
the offices that did not review annual inspection reports had little
direct interaction with the state agencies. Private lenders overseeing
non-FHA insured residential care facilities told us that they regularly
conduct various levels of management reviews and review annual inspection
reports on a consistent basis.

FHA has emphasized the importance of ongoing coordination with state
oversight agencies in its proposed revisions to its regulatory agreements,
which require owners or operators of insured facilities to report any
state or federal violations to FHA. HUD's proposed revisions to the
regulatory agreements also include a requirement that the owner or
operator provide HUD with copies of annual inspection reports that can be
used as part of loan monitoring. However, the proposed revisions to the
regulatory agreements have yet to be approved.

FHA's Limited Coordination with States on Oversight Issues May Lead to
Missed Identification of Risk Indicators

Serious quality of care deficiencies can have a variety of implications
that affect cash flow streams, ranging from a related reduction in
occupancy to the potential for civil money penalties and loss of licensing
and reimbursements. Consequently, quality of care concerns can ultimately
affect a facility's financial condition. For many Section 232 properties,
in particular nursing homes, state oversight of quality of care helps to
determine whether a facility is licensed and eligible to receive Medicaid
and Medicare reimbursements. This is particularly important to the Section
232 program because, as noted earlier in this report, Medicaid and
Medicare reimbursements typically account for a significant portion of
nursing home income.

Federal or state annual inspection reports, to the extent that they are
available for facilities, provide regular evaluations of nursing homes and
other residential care facilities. As discussed earlier, annual
inspections provide a review of quality of care by officials with relevant
backgrounds. In a 2005 report, we found inconsistencies across states in
conducting surveys and state surveyors understating serious deficiencies
in quality of care.17 Nonetheless, annual inspection reports serve as an
important indicator of a property's risk related to problems with the
quality of care to residents.

Annual inspection reports, coupled with other information such as facility
staffing profiles, resident turnover, and data from financial statements,
could assist HUD's field offices in overseeing loan performance.
Additionally, reviewing facilities' quality of care records over time, as
well as any corrective action plans needed to come into compliance with
state and federal quality of care requirements could further the field
offices' ability to identify loan performance risks. The reports may also
prompt HUD field office officials to communicate with federal or state
nursing home regulatory agencies for further information on facilities
that appear to be high risk. These agencies may have available information
on civil money penalties and sanctions, which serve as additional
indicators of quality of care risk. Private lenders we spoke with
acknowledged that annual inspection reports provided insight into the
management of a facility and coupled with other information could help to
assess financial risk.

Program and Industry Trends Show Sources of Potential Risks to the GI/SRI
Fund

The Section 232 program represents a relatively small share of the broader
GI/ SRI Fund. However, program and industry trends show sources of
potential risks that could affect the future performance of the Section
232 portfolio and the GI/SRI Fund. FHA uses a number of tools to mitigate
risk to the program and to the fund.

The Section 232 Program Represents a Small Percentage of the GI/SRI Fund

The Section 232 program is a relatively small share of the total GI/SRI
Fund. HUD estimated that the program would represent only about 5.3
percent of the fund's fiscal year 2006 commitment authority.18 Similarly,
the Section 232 program represents a little less than 16 percent or a
little more than $12.5 billion of the nearly $80 billion in unpaid
principle balance in the GI/SRI Fund (see fig. 1). Despite its small size,
a significant worsening in the performance of the Section 232 program
could negatively affect the performance of the GI/SRI Fund. The extent,
though, of the impact on the overall performance of the GI/SRI Fund would
depend upon numerous factors including changes in the size and performance
of the other programs in the fund.

Figure 1: The Section 232 Program Comprises a Relatively Small Part of the
GI/SRI Fund

Note: Numbers have been rounded to closest whole number.

Program Trends Show Sources of Potential Risk

As discussed below, several trends exist within the Section 232 program
that pose potential risks to the Section 232 portfolio and, therefore, to
the GI/SRI Fund.

Higher Claim Rates for Recent Loan Cohorts

To identify potential trends in loan performance, we analyzed 5- and
10-year claim rates for Section 232 loans based on data that spanned from
fiscal year 1960 through the end of fiscal year 2005, for the entire
portfolio, as well as by type of loan purpose and type of insured
facility. The analysis of the entire portfolio showed that the 10-year
claim rates for more recent loan cohorts (loans originated between 1987
and 1991 and loans originated between 1992 and 1996) ranked among the
highest historical cohort claim rates (see fig. 2).19 The 5-year claim
rate for loans originated between 1997 and 2001 also ranked among the
highest historical cohort claim rates. A continued increase in claim rates
could have a negative effect on the performance of the GI/SRI Fund.

Figure 2: Overall 5- and 10-Year Claim Rates for the Most Recent Cohorts
of the Section 232 Program Are Among the Highest Historical Claim Rates

Changes in Claim Rates by Loan Purpose

Section 232 loans can have a loan purpose in one of two categories-new
construction/substantial rehabilitation loans or refinance/purchase loans.
New construction loans are for loans that involve the construction of a
new residential care facility. Substantial rehabilitation loans are for
loans that meet HUD criteria for substantial rehabilitation of a
residential care facility, such as two or more building components being
substantially replaced. Purchase loans are for loans in which the borrower
is acquiring an existing residential care facility, while refinance loans
are the refinancing of an existing HUD insured loan or a loan not
previously insured by HUD. As described earlier in the report, HUD began
to allow for the refinancing of FHA-insured facilities and non-FHA insured
facilities in 1987 and 1994, respectively. When analyzing Section 232 loan
data by loan purpose, we found that new construction/substantial
rehabilitation loans have a higher 5-year claim rate than
refinance/purchase loans for the most recent cohort

for which data are available (see fig. 3).20 New construction/substantial
rehabilitation loans originated between 1997 and 2001 also have the
highest historical 5-year cohort claim rate for these type of loans.
Because of the higher claim rates in recent years, continued monitoring
will be important. In contrast, the number of refinance and purchase loans
endorsed in the last 5 years is more than double those endorsed in the
previous 5 years. The future impact of the refinance and purchase loans on
the overall performance of the Section 232 program is uncertain since they
have existed for a shorter period of time and thus there is currently
limited data available to assess the relative risk of claims.

Figure 3: The 5-Year Claim Rate for New Construction Loans Has Increased
in the Most Recent Cohort for Which Claim Rate Data Are Available

Changes in Claim Rates by Facility Type

As discussed earlier in the report, HUD insures different types of
residential care facilities that include nursing homes, intermediate care
facilities, assisted living facilities, and board and care facilities.
Assisted living facilities are relatively new to the portfolio, and the
number of these loans have been increasing. Our analysis of Section 232
loan data by facility type found that board and care facilities had a
slightly higher 10-year claim rate than nursing home facilities in the
most recent cohorts; however, these loans remain a very small percentage
of the active portfolio and are being made in decreasing numbers. There
are limited data to observe claim trends on assisted living facilities,
making their risk difficult to assess, but the 5-year claim rates for
assisted living facilities have increased significantly in the most recent
cohort years for which claim rate data are available (see fig. 4). A
continued high claim rate in assisted living facilities could negatively
affect the performance of the Section 232 program and the GI/SRI Fund.
However, lenders and HUD officials told us that, although assisted living
facilities had high claim rates in the past, they believe the market has
stabilized and lessons have been learned.

Figure 4: The 5-Year Claim Rates for Assisted Living and Board and Care
Facilities Have Increased in the Most Recent Cohorts for Which Claim Rate
Data Are Available

Increase in Loan Prepayments

Another observable trend is the increase in the portion of loans in each
cohort that is prepaid. (Prepayment occurs when a borrower pays a loan in
full before the loan reaches maturity.) There have been 1,688 prepayments
in the Section 232 program from 1960 through the end of fiscal year 2005
and loans that terminate do so overwhelmingly because of prepayment.
Moreover, the proportion of loans that terminate due to prepayment within
10 years of origination is increasing. Specifically, the 10-year
prepayment rates for the three most recent cohorts for which 10-year claim
rates are available are more than double that of some earlier cohorts. As
more borrowers prepay their loans, HUD loses future cash flows from
premiums; thus, higher prepayment rates will likely make the net present
value of cash flows decrease. However, the decrease could be offset to the
extent that higher prepayment rates result in fewer claims (a prepaid loan
cannot result in a claim).

Concentration of Loans

Market concentration also poses some risks to the GI/SRI Fund. The Section
232 program is concentrated in several large markets and in loans made by
relatively few lenders. As of 2005, five states (California, Illinois,
Massachusetts, New York, and Ohio) held 51 percent of active Section 232
loan dollars and 38 percent of active loan properties (see fig. 5). New
York holds close to 24 percent of the active loan dollars in the
portfolio. This is an improvement since 1995 when we found that eight
states accounted for 70 percent of the portfolio, and New York accounted
for 32 percent of the portfolio. However, the current market concentration
could still pose risk to the portfolio if a sudden market change took
place in one or more of the states with a larger percentage of the insured
Section 232 loans. We also found significant loan concentration among a
small group of lenders. While a total of 109 lenders held active loans, 6
hold over half of the active loan portfolio. GMAC Commercial Mortgage
Corporation holds more than 17 percent of all active mortgages in the
Section 232 program, the single largest share of any lender. This
concentration among lenders potentially makes the program more vulnerable
if one or a few large lenders encounter financial difficulty.

Figure 5: Section 232 Properties Are Concentrated in Several States

Note: Active loan dollars are from F47 as of the end of calendar year
2005, and number of loans are from F47 as of the end of fiscal year 2005.
This also does not include one loan for which property state information
was not available in F47. Numbers have been rounded to closest whole
number.

Industry Faces Uncertainties

The Section 232 program may also face risks from trends in the residential
care industry at large that include uncertainty about sources of revenue
and occupancy. Nursing home revenue is generated in large part from the
Medicare and Medicaid programs, which make up 58 percent of national
nursing home spending. Private lenders we interviewed that offer
non-FHA-insured residential care facility loans explained that one of the
primary reasons their loans are shorter-term loans than those of HUD is
due to their perception of the potential, long-term uncertainty in the
funding of the Medicaid and Medicare programs, which generally account for
a large share of patient payments in nursing homes. We and others have
reported that Medicare and Medicaid spending may not be sustainable at

current levels.21 In our 2003 report on the impact of fiscal pressures on
state reimbursement rates, however, we found that even in states that
recently faced fiscal pressures, reimbursement rates remained largely
unaffected.22 At that time, we concluded that any future changes to state
reimbursement rates remain uncertain. If program cuts occur in federal
spending on Medicaid that result in shifting costs from the federal
government to state governments, states could contain costs by taking a
number of steps, including freezing or reducing reimbursement rates to
providers. An ongoing tension exists, however, between what federal and
state governments and the nursing home industry believe to be reasonable
Medicare and Medicaid reimbursement rates to operate efficient and
economic facilities that provide quality care to public beneficiaries. As
the federal and state governments face growing long-term financial
pressure on their budgets, these budgetary pressures may have some
spillover effects on Medicare and Medicaid revenue streams for the nursing
home industry.

Uncertainty also exists about the future demand for residential care
facilities and the corresponding effects on occupancy. As the number of
Americans aged 65 and older increases at a rapid pace, lenders we
interviewed projected an increased need for residential care facilities.23
Industry officials also noted a rise in alternatives to nursing home care,
such as assisted living facilities and home and community-based care
options. As patients choose alternative care options, traditional nursing
homes may face occupancy challenges. Overall, these changes to the nursing
home facilities patient base may lower occupancy and income levels for
nursing homes, including those in the Section 232 portfolio. However,
these changes may positively affect the occupancy and income levels of
other types of residential care facilities, including those in the Section
232 portfolio.

FHA Uses a Number of Tools to Mitigate Risks

As described elsewhere in this report, FHA uses a number of tools to
mitigate risks to the program and to the GI/SRI Fund. These tools include
imposing requirements prior to insuring loans to help prevent riskier
loans from entering the Section 232 portfolio. FHA also uses various
tools-such as reports on physical inspections of facilities, and financial
and other information captured in data systems-to monitor the status of
insured facilities and the performance of their loans. Additionally, FHA
officials use quality control reviews to mitigate the risk for the program
as a whole using two processes: Quality Management Reviews and Lender
Qualifications and Monitoring Division reviews (the latter reviews are
described in app. II).

HUD's Model for Estimating Credit Subsidy Costs Excludes Some Potentially
Relevant Factors

HUD's model for estimating annual credit subsidies-which incorporates
assessments of various risks that loan cohorts will face and includes
assumptions consistent with the Office of Management and Budget (OMB)
guidance-does not explicitly consider the impacts of some potentially
important factors. These factors include: variables to capture the impact
of prepayment penalties or restrictions on prepayments, the loan-to-value
ratio and debt service coverage ratios of Section 232 properties at the
time of loan origination and differences between types of residential care
facilities. Further, the model does not fully capture the effects on
existing loans to changes in market interest rates, and it uses proxy data
that are not comparable to the loans in the Section 232 program. As a
result, HUD's model for estimating the program's credit subsidy may result
in over- or underestimation of costs.

HUD Uses a Model to Estimate Credit Subsidy Costs

Federal law requires HUD to estimate a credit subsidy for its loan
guarantees. The credit subsidy cost is the estimated long-term cost to the
government of a loan guarantee calculated on a net present value basis and
excluding administrative costs. HUD estimates a credit subsidy for each
loan cohort. This estimate reflects HUD's assessment of various risks,
based in part on the performance of loans already insured. Since 2000, HUD
has annually estimated two credit subsidy rates for the Section 232
program, reflecting its two largest risk categories: loans for new

construction and substantial rehabilitation, and loans for refinance and
purchase loans.24 HUD uses an identical methodology for each estimate.

To estimate the initial subsidy cost of the Section 232 program, HUD uses
a cash flow model to project the cash flows for all identified loans over
their expected life. The cash flow model incorporates regression models
and uses assumptions based on historical and projected data to estimate
the amount and timing of claims, subsequent recoveries from these claims,
prepayments, and premiums and fees paid by the borrower. The regression
models incorporate various economic variables such as changes in GDP,
unemployment rate, and 10-year bond rates. The model also has broken out
claim and prepayment data into new construction and refinance loans since
these loans are expected to perform differently.

HUD inputs its estimated cash flows into OMB's credit subsidy calculator,
which calculates the present value of the cash flows and produces the
official credit subsidy rate. A positive credit subsidy rate means that
the present value of cash outflows is greater than inflows, and a negative
credit subsidy rate means that the present value of cash inflows is
greater than cash outflows. For the Section 232 program, cash inflows
include premiums and fees, servicing and repayment income from notes held
in inventory, rental income from properties held in inventory, and sale
income from notes and properties sold from inventory. Cash outflows
include claim payments and expenses related to properties held in
inventory.

Since HUD began estimating the initial subsidy cost of the Section 232
program, it has estimated that the present value of cash inflows would
exceed the outflows. As a result, the initial credit subsidy rates for the
Section 232 program were negative. However, estimates from more recent
years showed that the negative subsidy rates on new construction and
substantial rehabilitation loans have generally been shrinking, meaning
that the projected difference between the program's cash inflows and cash
outflows was decreasing. In HUD's most recent estimate (for the fiscal
year 2007 cohort), the estimated cash inflows exceed the estimated cash
outflows by a considerably greater margin than in any previous year's
estimate. This may reflect increased premiums for Section 232 loans; the
President's proposed budget for fiscal year 2007 specifies increases in
mortgage insurance premiums for almost all FHA programs, including
increasing the rate for Section 232 refinance and new construction loans
to 80 basis points from 57 basis points. Figure 6 shows changes in the
initial estimated credit subsidy rate over time for both loan categories.

Figure 6: Initial Credit Subsidy Estimates for Section 232 Program New
Construction and Substantial Rehabilitation Loans and for Section 232
Program Refinance and Purchase Loans Have Not Indicated a Need for
Subsidies

Note: Initial credit subsidy estimates were not available for 1997 for new
construction and substantial rehabilitation loans and for 1996-1999 for
refinance and purchase loans.

Features of the Credit Subsidy Model May Lead to Unreliable Credit Subsidy
Estimates

HUD's model for estimating credit subsidy rates incorporates numerous
variables, but the model's exclusion of potentially relevant factors and
its use of proxy data from another FHA loan program may negatively affect
the quality of the estimates. Including additional information in the
model could enhance the predictive value of the model.

Prepayment Penalties or Restrictions

According to some economic studies, prepayment penalties, or penalties
associated with the payment of a loan before its maturity date, can
significantly affect borrowers' prepayment patterns.25 This is also
important for claims, since if a loan is prepaid it can no longer go to
claim. HUD's model does not explicitly consider the potential impact of
prepayment penalties or restrictions, even though they can influence the
timing of prepayments and claims and collections of premiums. According to
FHA officials, FHA does not place prepayment penalties on FHA-insured
nursing home loans. However, according to the Section 232 program's
regulations, a lender can impose a prepayment penalty charge and place a
prepayment restriction on the mortgage's term, amount, and conditions.26
We reviewed a sample of Section 232 loans and found that prepayment
penalties and restrictions were consistently applied to these loans.27

According to FHA officials and mortgage bankers, prepayment restrictions
on Section 232 loans typically range from 2 to 10 years of prepayment
restrictions and 2 to 8 years of prepayment penalties. While FHA does not
specifically maintain data on insured residential care facility financing
terms, prepayment restrictions are specified on the mortgage note, which
is available to FHA. Incorporation of such data into the Section 232
program's credit subsidy rate model could refine HUD's credit subsidy
estimate by enhancing the model's ability to account for estimated changes
in cash flows as a result of prepayment restrictions.

According to HUD officials responsible for HUD's cash flow model,
prepayment penalties and restrictions are not incorporated into the model
because HUD does not collect such data. HUD officials added that even
though the cash flow model does not explicitly account for prepayment
penalties and restrictions, its use of historic data implicitly captures
trends that may occur as a result of prepayment penalties and
restrictions. The model's projections are influenced by the average level
of prepayment protection in the historical data but not by the trend. If
prepayment penalties and other restrictions have changed over time in the
past, or change in the future, then not incorporating this information
could lead to less reliable estimates.

Debt Service Coverage Ratio at Point of Loan Origination

Initial debt service coverage ratios are another important factor that may
affect cash flows, as loans with lower initial debt service coverage
ratios may be more likely to default and result in a claim payment. HUD's
cash flow model does not consider the initial debt service coverage ratio
of Section 232 loans at the point of loan origination. By initial debt
service coverage ratio, we are referring to the projected debt service
coverage ratio that is considered during loan underwriting. According to
the HUD official responsible for HUD's cash flow model, the initial debt
service coverage ratio of a residential care facility is not included as a
part of the cash flow model because it (1) is not a cash flow, (2) does
not vary, and (3) has no predictive value. We agree that a debt service
coverage ratio is not a cash flow. However, initial debt service coverage
ratios potentially affect relevant cash flows, as do other factors that
are included in HUD's model but are also not cash flows to HUD, such as
prepayments. For example, the model considers estimated prepayments
because they potentially affect future cash inflows from fees and future
cash outflows from claim payments.

Our analysis of available projected debt service coverage ratios, which
include the amount of new debt being insured, shows that these ratios
varied from 1.1 to 3.6.28 All other factors being equal, loans with debt
service coverage ratios of 3.6 are generally considered to have less risk
than a loan with only a 1.1 debt service coverage ratio.

Economic theory suggests that the debt service coverage ratio is an
important factor in commercial mortgage defaults. However, empirical
studies show mixed results regarding the significance of the impact of
debt service coverage ratios upon commercial mortgage defaults. Some
studies indicate that debt service coverage ratios are meaningful factors
in modeling default risk and are helpful in predicting commercial mortgage

terminations.29 Other studies find initial debt service coverage ratios to
be statistically insignificant in modeling commercial mortgage defaults.30
These mixed results may be the consequence of relatively small sample
sizes and model specification issues.

Loan-to-Value Ratio at Point of Loan Origination

Initial loan-to-value ratios are another important factor that may affect
cash flows, as loans with higher initial loan-to-value ratios may be more
likely to default and result in a claim payment. By initial loan-to-value
ratio, we are referring to the projected loan-to-value ratio that is
considered during loan underwriting. HUD's cash flow model also does not
consider the initial loan-to-value ratio of Section 232 loans at the point
of loan origination.

According to the HUD official responsible for HUD's cash flow model, the
initial loan-to-value ratio of a Section 232 property is not included as a
part of the cash flow model because it does not vary and has no predictive
value. However, our analysis of available projected loan-to-value ratios,
which include the amount of new debt being insured, shows that these
ratios varied from 66 percent to 95 percent.31 All other factors being
equal, loans with loan-to-value ratios of 66 percent are generally
considered to have less risk than a loan with only a 95 percent
loan-to-value ratio. While economic theory suggests that the loan-to-value
ratio is an important factor in commercial mortgage defaults, empirical
studies show mixed results regarding its significance. Some studies
indicate that loan-to-value ratios are meaningful factors in modeling
default risk and are helpful in predicting

commercial mortgage terminations.32 Other studies find initial
loan-to-value ratios to be statistically insignificant in modeling
commercial mortgage defaults.33 These mixed results may be the consequence
of relatively small sample sizes and model specification issues.

Types of Facilities Insured and Changes in Interest Rates

The model's ability to reliably forecast claim rates may be enhanced by
incorporating a variable indicating facility type into the regression
analysis. HUD's cash flow model does not explicitly consider differences
in loan performance between types of facilities, such as nursing homes,
assisted living facilities, and board and care facilities. However, when
looking at the most recent cohorts for which 5-year claim rates are
available, our analysis found the 5-year claim rates for assisted living
facilities to be significantly higher than the 5-year claim rates for
nursing homes (6.7 percent 5-year claim rate for nursing homes versus 13.6
percent for assisted living facilities).

In addition, we found that HUD's cash flow model generally incorporates
the interest rate on the individual loans (the contract rate) and the
prevailing market interest rate (captured by the 10-year bond rate) as
separate variables. Economic theory suggests, when modeling mortgage
terminations, that considering these two variables jointly as a single
variable in the form of a ratio is the best way to capture the effects on
existing loans when market interest rates change.34 For example, if market
rates fall below the contract rate on existing Section 232 loans, then it
may become more attractive for borrowers to prepay. However, if market
rates fall but remain above the contract rates, then it may not become
more attractive for borrowers to prepay. Using a ratio captures the
distinction between these two examples because it considers the relative
cost to the borrower of the mortgage given the contract rate, as compared
to the mortgage with the market interest rate. By generally considering
the contract rate and market interest rate separately, HUD potentially
loses the ability to capture this distinction and predict large responses
when market rates fall and small responses when market rates rise.35

Use of Proxy Data

HUD's use of Section 207 loans as a proxy for Section 232 refinance loans
could lead to less reliable credit subsidy estimates for the Section 232
program. HUD uses certain Section 207 loans-refinance loans for existing
multifamily housing properties-as proxy data for the claim regression for
Section 232 refinance loans. The Section 207 loans are not residential
health care facility loans. According to HUD officials, HUD uses the
Section 207 loans because there are insufficient data on Section 232
refinance loans. A HUD official told us that Section 207 loans were
selected as proxy data because they are refinance loans and because they
have similar performance to the Section 232 refinance loans, as indicated
by the cumulative claim rates they calculated.

Consideration of the basis for using proxy data is important. When using
the experience of another agency or a private lender as a proxy, the
Federal Accounting Standards Advisory Board (FASAB) suggests that an
agency explain why this experience is applicable to the agency's credit
program and examine possible biases for which an adjustment is needed,
such as different borrower characteristics.36 HUD could reasonably be
expected to follow the FASAB guidance when using data from a different
program at HUD. HUD told us that they did not compare borrower
characteristics for Section 207 loans and Section 232 loans. A HUD
official told us that HUD agreed that they would not expect borrowers of
Section 207 loans to have similar characteristics to borrowers of Section
232 loans.

HUD analyzed the comparability of Section 207 and Section 232 refinance
loans using cumulative claim rate analysis, but we question the
methodology the agency used to make this comparison. Additionally, we
compared the refinance loans for each of the programs by calculating
conditional claim and prepayment rates as well as 5-year cumulative claim
and prepayment rates, and we found significant differences between the
programs (see app. IV for a further description of HUD's methodology and
our comparison of the two programs).

We question HUD's use of Section 207 loans as a proxy for Section 232
loans, given the differences we observed. We cannot fully estimate the
overall impact on the credit subsidy estimate, and the effects of the
claim and prepayment rates could partially offset each other. The higher
prepayment rates for Section 207 loans could lead to HUD underestimating
future revenues for Section 232 loans (HUD would project that many of
these loans would terminate, although they would actually remain active
and pay premium revenue to HUD.) The lower claim rates on Section 207
loans could result in HUD estimating that fewer of its Section 232 loans
would result in a claim and thus lead it to underestimate future costs.

In the future, more data will be available on the actual performance of
Section 232 refinance loans that can be used in estimating credit subsidy
needs. To avoid using questionable proxy data in the interim, one possible
approach, among others, would be to use a simpler estimation method, such
as using average claim and prepayment rates over time as is done in
estimating credit subsidy rates for the Section 242 Hospital Mortgage
Insurance program.

Conclusions

The Section 232 program is the only source of mortgage insurance for
residential care facilities. Accordingly, it is important to ensure good
program and risk management practices. While some field offices we visited
had adopted practices to better manage risks of their Section 232 loans,
varying awareness of program requirements and insufficient levels of staff
expertise contribute to increased financial risk in the Section 232
program loan portfolio and thus the GI/SRI Fund. HUD has numerous
underwriting and monitoring guidelines and policies to manage the risks of
Section 232 loans. However, to the extent that field office staff do not
accurately implement current underwriting and monitoring guidelines and
policies, they potentially allow loans with unwarranted risks to enter the
portfolio and may miss opportunities to identify problems with
already-insured loans early enough to help prevent claims. Revising the
"Multifamily Asset Management and Project Servicing Handbook" to include
monitoring requirements specific to the Section 232 program, as the Office
of Inspector General noted in its 2002 report, would help in this regard.
So too would the sharing of additional practices, such as involving asset
management staff in the underwriting process, undertaken by some field
offices to better manage risks in their program loans. Moreover,
adequately training staff to develop expertise on residential care loans
and industry could help assure proper underwriting and oversight of
Section 232 loans, which tend to be more complex than those in other HUD
multifamily programs. Field office officials' concerns about their
existing levels of staff expertise heighten the need for appropriate
guidance and additional training specific to the Section 232 program,
while the potential loss of specialized staff within the next 5 years
underscores the need for HUD, in the context of its strategic human
capital efforts, to assure adequate program expertise in the future.

Although HUD recommends that field offices obtain and review annual
inspection reports for licensed facilities insured by the program, four of
five offices we visited did not do so. By not routinely using, in
combination with other performance indicators, the results of annual
inspection reports on insured facilities subject to such inspections, HUD
may be missing important indicators of problems that could result in
claims that might otherwise have been prevented. Reviewing inspection
reports is also a means of obtaining relevant information about insured
facilities that have not been the subject of FHA management reviews. HUD's
long-proposed revisions to its residential care facility regulatory
agreement recognize the potential usefulness of information on
state-administered inspections by requiring that owners or operators
report inspection violations and supply HUD with copies of annual
inspection reports. The proposed revisions would also address a number of
the internal control weaknesses identified by the HUD Inspector General's
2002 report, but it remains unclear when the proposed revisions will be
approved, leaving the program exposed to identified weaknesses in the
interim.

While the Section 232 program represents a relatively small portion of the
GI/SRI Fund, it faces risks that could affect the performance of the loan
portfolio and the fund. HUD uses a number of tools to mitigate risks, and
it will be important to continue monitoring program trends and industry
developments. Recent increases in the numbers of assisted living facility
loans and refinance loans are a source of uncertainty, in that there is as
yet little data with which to assess their long-term performance.
Similarly, industry trends and the availability of future Medicaid and
Medicare funds are sources of uncertainty, and heighten the need for HUD
to have sufficient staff expertise with which to monitor future
developments that could affect the program and ultimately the GI/SRI Fund.

HUD's model for estimating the program's credit subsidy incorporates
assessments of various risks that loan cohorts face, but it does not
explicitly consider certain factors that could result in over- or
underestimation of costs. These factors include prepayment penalties,
lockout provisions, facility type, loan-to-value ratio, the debt service
coverage ratio of loans at commitment, and the ratio of contract rates to
markets rates, which some economic studies suggest are potentially useful
in modeling risks. Including such factors could enhance the credit subsidy
estimates and provide HUD and the Congress with better cost data with
which to assess the program. Additionally, HUD's use of Section 207
refinance loans, which we do not find to be a good proxy for Section 232
refinance loans, could specifically contribute to over- or underestimation
of the credit subsidy for the refinance loans in the program.

Recommendations for Executive Action

To ensure that field offices are aware of and implement current
requirements and policies for the Section 232 Mortgage Insurance for
Residential Care Facilities program,  and reduce risk to the GI/SRI Fund,
we recommend that the Secretary of Housing and Urban Development direct
the FHA Commissioner to take the following actions:

o Revise the "Multifamily Asset Management and Project Servicing Handbook"
in a timely manner to include monitoring requirements specific to Section
232 properties;

o Establish a process for systematically sharing loan underwriting and
monitoring practices among field offices involved with the Section 232
program;

o Assure, as part of the department's strategic human capital management
efforts, sufficient levels of staff with appropriate training and
expertise for Section 232 loans;

o Incorporate a review of annual inspection reports for insured Section
232 facilities that are subject to federal or state inspections, even in
the absence of a revised regulatory agreement; and

o Complete and implement the revised regulatory agreements in a timely
manner.

To potentially improve HUD's estimates of the program's annual credit
subsidy, we recommend that the Secretary of Housing and Urban Development
explore the value of explicitly factoring additional information into its
credit subsidy model, such as prepayment penalties and restrictions, debt
service coverage and loan-to-value ratios of facilities as they enter the
program, facility type, and the ratio of contract rates to market rates.
We also recommend that the Secretary of Housing and Urban Development
specifically explore other means of modeling the performance of Section
232 refinance loans.

Agency Comments and Our Evaluation

We provided a draft of this report to HUD for their review and comment. In
written comments from HUD's Assistant Secretary for Housing-Federal
Housing Commissioner, HUD generally concurred with our recommendations
intended to ensure that field offices are aware of and implement current
program requirements and policies. However, the agency disagreed with most
parts of our recommendation related to HUD's credit subsidy model. The
Assistant Secretary's letter appears in appendix V.

HUD stated that it has initiated a full review of the Section 232 program
and that GAO's recommendations related to ensuring that field offices are
aware of and implement current requirements are being incorporated into
plans for revising the program. More specifically, HUD stated that it:

o will draft and implement changes to the program handbook;

o will initiate staff training and assure that staff is adequately trained
in underwriting and servicing policies; and

o plans to prepare a report addressing state and federal inspections,
among other things, to enhance FHA participation in and oversight of
insured health care mortgages.

HUD also provided a timeline by which to complete and implement the
revised regulatory agreements.

Concerning our recommendation that HUD explore the value of explicitly
factoring in additional information into its credit subsidy model, HUD
stated that it agreed to take into account differences among types of
residential care facilities in its modeling, when it has sufficient
historical data and if the data indicate that loan performance varies
sufficiently by type of facility. However, HUD disagreed with considering
other factors we suggested, as follows:

o Initial loan-to-value and debt service coverage ratios. HUD stated that
(1) studies we cited in our draft report found these ratios to be
statistically insignificant in predicting commercial mortgage defaults and
(2) that data are unavailable for this analysis. We agree, as our draft
report stated, that economic studies have shown mixed results regarding
the significance of the impact of loan-to-value and debt service coverage
ratios on commercial mortgage defaults, with some studies finding them to
be significant predictors and others finding them to be insignificant
predictors. We further stated that these mixed results may be the result
of small sample sizes and model specification issues. Nevertheless, we
continue to believe that HUD should explore the value of factoring initial
loan-to-value ratio and debt service coverage ratio into its credit
subsidy model, and we did not change our recommendation. Regarding the
second point, HUD has the data for analyzing loan-to-value and debt
service coverage ratios in individual loan files and could include these
data in its credit subsidy modeling by creating an electronic record of
this information either for its entire portfolio or for a sample of the
portfolio. Consequently, we did not change the recommendation.

o Factors potentially affecting prepayments. HUD disagreed with our
suggestion that its credit subsidy model does not fully capture the
effects of prepayment penalties, stating that its use of historical data
captures the effect of prepayment penalties on project owners' behavior.
However, as we stated in the draft report, HUD's use of historic data
would not fully capture trends related to changes in prepayments. HUD also
stated that it has tested using the difference between mortgage interest
rates and the 10-year Treasury bond rates in its modeling of prepayments.
However, our recommendation was to consider a ratio of these two interest
rates, not the difference. As we noted in our report, economic theory
suggests that the use of a ratio is the best way to capture the effects on
existing loans when market interest rates change. Consequently, we did not
change the recommendation.

o Use of Section 207 loans as proxy data for refinance loans. HUD stated
that it did not believe that the differences between Section 207 and
Section 232 loans that our report noted justify concerns that residential
care refinance loans are being improperly modeled and noted a lack of
available data. We agree that sufficient relevant data on Section 232
refinance loan performance do not yet exist, but we continue to question
the use of Section 207 loan data as a proxy. While we did not change the
recommendation, we added language to our report suggesting that, until
enough Section 232 refinance loan data are available, one possible
approach, among others, would be to use a simpler estimation method, such
as using average claim and prepayment rates over time as is done in
estimating credit subsidy rates for the Section 242 Hospital Mortgage
Insurance program.

We are sending copies of this report to the Secretary of the Department of
Housing and Urban Development (HUD). 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 about this report or need
additional information, please contact me at 202-512-8678 or [email protected]
. Contact points for our offices of Congressional Relations or Public
Affairs may be found on the last page of this report. GAO staff who made
major contributions to this report are listed in appendix V.

David G. Wood, Director Financial Markets and Community Investment

List of Congressional Addressees

The Honorable Christopher Bond Chairman The Honorable Patty Murray Ranking
Member Subcommittee on Transportation,   Treasury, the Judiciary, Housing
and Urban   Development, and Related Agencies Committee on Appropriations
United States Senate

The Honorable Jack Reed Ranking Minority Member Subcommittee on Housing
and Transportation Committee on Banking, Housing, and Urban Affairs United
States Senate

The Honorable Joe Knollenberg Chairman The Honorable John W. Olver Ranking
Member Subcommittee on Transportation, Treasury, and   Housing and Urban
Development, The Judiciary,   District of Columbia, and Independent
Agencies Committee on Appropriations House of Representatives

The Honorable Lincoln Chafee United States Senate

The Honorable Patrick Kennedy House of Representatives

The Honorable James Langevin House of Representatives

Appendix I  Objectives, Scope, and Methodology 

Our objectives were to examine (1) the Department of Housing and Urban
Development's (HUD) overall management of the program, including loan
underwriting and monitoring; (2) the extent to which HUD's oversight of
insured health care facilities is coordinated with the states' oversight
of the quality of care provided by facilities; and (3) the financial
implications of the program to the General Insurance/Special Risk
Insurance (GI/SRI) Fund, including risk posed by program and market
trends; and (4) how HUD estimates the annual credit subsidy for the
program, including the factors and assumptions used. In addition, we
examined HUD's action in response to a HUD Inspector General report that
concluded that HUD's Office of Housing did not have adequate controls to
effectively manage the Section 232 program; this information is summarized
in appendix III.

To examine HUD's overall management of the Section 232 program, we
obtained and reviewed program manuals, guidance, and documentation,
including the "MAP Guide," HUD's Section 232 "Mortgage Insurance for
Residential Care Facilities Handbook," and HUD's "Multifamily Asset
Management and Project Servicing Handbook," for loan processing
procedures, underwriting policies and requirements, and oversight policies
and requirements. We also interviewed HUD officials at HUD headquarters
who are responsible for providing guidance and policies on loan
underwriting and oversight and three private lenders that offered
FHA-insured Section 232 loans. In addition, we conducted site visits to
five HUD field offices (Atlanta, Georgia; Buffalo, New York; Chicago,
Illinois; Los Angeles, California; and San Francisco, California) and
conducted interviews with HUD officials, including the Hub or acting Hub
director, appraisers, mortgage credit analysts, and project managers that
are responsible for Section 232 loan applications, underwriting, and
oversight, as well as other Federal Housing Administration (FHA) programs.
We gathered relevant program documentation from each site visit. We also
interviewed an official from one of HUD's Multifamily Property Disposition
Centers during our site visit to Atlanta. To capture a variety of Section
232 loan activity, we selected five HUD field offices on the basis of (1)
the volume of Section 232 loans the field office had processed during
fiscal year 2004 up to September 2005; (2) the dollar amount of Section
232 loans processed in the field office during fiscal year 2004 up to
September 2005; (3) the timeliness of processing Section 232 loans during
the last 2 years; (4) historical claim-rate data for the field office-that
is, the rate at which Section 232 loans processed by the field office have
gone to claim; (5) HUD's suggestions for field office site visits; and (6)
geographical dispersion.

To better understand how private lenders that do not participate in the
Section 232 program manage risks, we interviewed five private lenders that
offered non-FHA insured loans to residential health care facilities. We
also interviewed representatives of three residential care facilities with
FHA-insured Section 232 loans to better understand the borrowers
perspective of the Section 232 program.

To examine the extent to which HUD coordinated with states' oversight of
quality of care provided by facilities, we reviewed FHA requirements for
conducting management reviews and reviewing annual inspection reports. We
also interviewed officials in FHA's Office of Multifamily Development and
Office of Asset Management and field office officials about policies for
coordination between FHA and state residential care oversight and rate
setting agencies, as well as policies for review of annual inspection
reports. In addition, we interviewed private lenders of FHA-insured and
non-FHA insured residential care facilities to better understand common
industry practices for coordination between lenders and state residential
care oversight and rate setting agencies.

To examine the financial risks that the program poses to the GI/SRI Fund,
we interviewed and obtained documentation from HUD's Office of Evaluation
and analyzed HUD data on program portfolio characteristics, including
number of loans by cohort, current insurance in force, geographic and
lender concentration of loans, and claims. We also analyzed HUD data used
for their refinance credit subsidy regression model. Specifically:

o To obtain the number of active and terminated loans and claim rate
history, we analyzed data from extracts of HUD's F47 database, a
multifamily database. We obtained extracts from HUD in May 2005, September
2005, and February 2006. Unless otherwise indicated, all analyses from the
F47 data in the report utilized the May 2005 extract with subsequent
updates from the other extracts and was current as of the end of fiscal
year 2005. To assess the reliability of the F47 database extract, we
reviewed relevant documentation, interviewed agency officials who worked
with the database, and conducted manual data testing, including comparison
to published data. Because of the small number of loans endorsed in
individual fiscal years, we conducted analyses of cohorts that were
created by combining data from 5 to 6 fiscal years. For claim rate
analyses, we analyzed 5- and 10-year claim rates for the data based on the
date of loan termination.

o Our analyses found 13 loans for which facility type information was not
able to be determined from the extract. FHA administrators were able to
determine the facility type for all but one of these loans using the
Development Application Processing (DAP) system. This one terminated loan
was excluded from facility type endorsement and claim rate analysis and,
therefore, had little impact on this report. We also determined final
endorsement date information to be missing from 799 records. Our analyses
only used initial endorsement date information for which data was
available for every record; therefore, there was no impact on this report.
We also determined there were nine loans for which the facility type
information was incorrect based on the endorsement date.1 FHA
administrators checked in the DAP system and confirmed the correct
facility type for these loans; therefore, there was no impact on the
report. We determined the data to be sufficiently reliable for analysis of
number of active and terminated loans, as well as claim rates.

o To determine the proportion of the Section 232 Mortgage Insurance
program's commitment authority to the larger GI/SRI Fund's commitment
authority, we reviewed HUD's fiscal year 2006 budget.

o To determine the proportion of the Section 232 Mortgage Insurance
program's unpaid principal balance to the larger GI/SRI Fund unpaid
principal balance, we obtained the GI/SRI Fund's unpaid principal balance
as of December 31, 2005 from HUD's Office of Evaluation. We also analyzed
data from HUD's Multifamily Data Web site, which is extracted from HUD's
F47 database, to determine the unpaid principal balance of Nursing Home
Mortgage Insurance program loans as of December 31, 2005.2

o To determine the geographic concentration of loan properties in the
program, we analyzed data current as of the end of fiscal year 2005 from
our extract of HUD's F47 database. Our analysis determined property state
data was missing for 270 project numbers. FHA administrators informed us
that loans endorsed more than 20 years ago, before electronic records were
maintained, may have missing data that is unavailable. Our analyses of
geographic concentration of loan properties utilized only one record with
missing property state data; therefore, there was little impact on our
findings. We determined the data to be sufficiently reliable for analysis
of geographic loan concentration.

o To determine the geographic concentration of loan dollars in the program
we analyzed data current as of December 31, 2005, from HUD's Multifamily
Housing Data Web site.

o To determine prepayment history in the program, we analyzed data from
our F47 extract, current as of the end of fiscal year 2005. We also
analyzed 5-and 10-year prepayment rates for the data based on the date of
loan termination.

o To determine the appropriateness of using Section 207 refinance loans as
proxy data in the Section 232 refinance loan credit subsidy estimate
regression model, we analyzed data from several extracts from HUD's Office
of Evaluation. The extracts contained the loan data used by HUD to
calculate cumulative claim rates for Section 232 and 207 refinance loans
for loans endorsed from fiscal year 1992 through fiscal year 2005. The
extracts did not include termination codes for all terminated loans. We
determined termination code data for these loans from HUD data current as
of December 31, 2005, from HUD's Multifamily Housing Data Web site. We
also combined the extracts to include all loans in one larger extract. In
addition, we performed manual data reliability assessments of these
extracts and determined that three loans should not have been included in
the extracts because they had section of the act codes that were not
within the parameters of our analysis as defined by the notes included in
HUD's extracts. These loans were not included in our analysis and,
therefore, had no impact on our findings. We determined the data to be
sufficiently reliable for analysis of the comparability of Section 207
refinance loans to Section 232 refinance loans.

We conducted a literature review and interviewed numerous officials of
lenders, residential care associations, and HUD to obtain information on
risks due to health care market trends. We also searched for Inspectors
General and agency reports through HUD Web sites. Finally, we conducted a
search on our internal Web site to identify previous work on the Section
232 program.

To determine how HUD estimates the annual credit subsidy rate for the
program, we reviewed documentation of HUD's credit subsidy estimation
procedures, reviewed the cash flow model for the program, and we
interviewed program officials from HUD's Office of Evaluation and program
auditors from the Office of Management and Budget (OMB). We also compared
the assumptions used in HUD's cash flow model with relevant OMB guidance
and reviewed economic literature on modeling defaults to identify factors
that are important for estimation. Additionally, we analyzed data provided
by HUD field offices on initial loan-to-value ratios and debt service
coverage ratios (at the time of loan application). We obtained the credit
subsidy rates from the Federal Credit Supplement of the United States
Budget.

To review the actions HUD has taken in response to the HUD Inspector
General's 2002 report on the Section 232 program, we interviewed officials
in HUD's Office of Inspector General. In addition, we reviewed the HUD
Inspector General's 2002 report, as well as HUD's Management Plan Status
Reports for Implementation of Recommendation 1A of audit 2002-KC-0002. We
also interviewed HUD headquarters officials, as well as field office
officials during our site visits.

Our review did not include an evaluation of underwriting criteria or the
need for the program. We conducted our work in Atlanta, Georgia; Buffalo,
New York; Chicago, Illinois; Los Angeles, California; San Francisco,
California, and Washington, D.C., between February 2005 and April 2006, in
accordance with generally accepted government auditing standards.

Appendix II  Information on the Application Processing, Underwriting, and Oversight of
Section 232 Loans

Application Processing and Underwriting for Section 232 Loans

The Department of Housing and Urban Development (HUD) currently processes
a majority of the Section 232 loans using the Multifamily Accelerated
Processing (MAP) program and processes some loans under Traditional
Application Processing (TAP). Under MAP, the lender conducts the
underwriting of the loan and submits a package directly to the Hub or
program center for mortgage insurance. The Hub or program center reviews
the lender's underwriting and makes a decision whether or not to provide
mortgage insurance for the loan. New construction and substantial
rehabilitation loans require a preapplication meeting where HUD reviews
required documentation up front. Under TAP, HUD, not the lender, is
primarily responsible for the underwriting of the loan and determines
whether or not to accept the loan.

FHA has numerous underwriting requirements for loans made under the
Section 232 program. Some examples include:

o Requiring documentation of a state-issued Certificate of Need (CON) for
skilled nursing facilities and intermediate care facilities, and in states
without a certificate of needs procedure, an alternative study of market
needs and feasibility.

o Requiring an appraisal of the facility (prepared by the lender under the
MAP program) and a market study with comparable properties.

o Reviewing current or prospective operators of the residential care
facility and ensuring that they meet certain standards. For example, FHA
has a requirement that operators of an assisted living facility have a
proven track record of at least 3 years in developing, marketing, and
operating either an assisted living facility or a board and care home.1

o For new construction facilities specifically, FHA requires a business
plan along with an estimate of occupancy rates and prospective
reimbursement rates with the percentage of population for patients whose
costs are reimbursed through Medicare and Medicaid.

o For existing facilities applying for a refinance loan, FHA requires the
submission of vacancy and turnover rates and current provider agreements
for Medicare and Medicaid, 3 years of balance sheet and operating
statements, as well as the latest inspection report on the project's
operation.2

o Requiring limits on loan-to-value and debt service coverage ratios,
ratios identified by field office officials we interviewed as two of the
more important financial ratios in the underwriting process. For example,
for Section 232 loans, the loan-to-value ratio cannot exceed 90 percent
for new construction loans, and 85 percent loan-to-value for refinance
loans.3

For loans processed under MAP, HUD field office officials are required to
use MAP Guide checklists to ensure that lenders follow FHA's underwriting
requirements. These checklists contained guidelines for reviewing lender
submissions and overall parameters that an application must meet. For
example, field office officials use an appraisal review checklist in the
MAP Guide to ensure that the submitted market study complies with MAP
requirements. For applications processed under TAP, field office officials
stated that they use similar checklists to the ones included in the MAP
Guide as the MAP Guide incorporates many of the Section 232 underwriting
requirements.

For MAP loans, HUD headquarters has a Lender Qualifications and Monitoring
Division (LQMD) that conducts reviews of loans. LQMD is responsible for
evaluating lender qualifications and lender performance. It reviews and
ultimately approves lenders requesting MAP lender approval for loan
underwriting. The division reviews a sample of lenders when a loan has
defaulted or there is a need for additional lender oversight. While LQMD
reviews are not specific to the Section 232 program, they help to monitor
lenders participating in the program and ultimately help to reduce the
number of risky loans that enter the portfolio.

Oversight and Monitoring of Section 232 Loans

FHA requires field office staff to conduct a number of reviews for
oversight of Section 232 loans. For example, staff address noncompliance
items that are identified by HUD's Financial Assessment Sub-System (FASS)
for each facility. Noncompliance items can include items such as
unauthorized distribution of project funds or unauthorized loans from
project funds. Using information from the annual financial statement,
FASS's computer model statistically calculates financial ratios, or
indicators, for each facility, and applies acceptable ranges of
performance, weights, and thresholds for each indicator. FASS then
generates a score for each facility based on these indicators, and this
financial score represents a single aggregate financial measure of the
facility. However, a HUD draft contractor study found that FASS did not
adequately account for the unique nature of nursing homes in the Section
232 portfolio and, therefore, was a poor predictor of a nursing home going
to claim. Field office officials we interviewed also review physical
inspections conducted by HUD's Real Estate Assessment Center (REAC), which
is responsible for conducting physical assessments of all HUD-insured
properties. Officials also ensure that the professional liability
requirement for facilities is met and conduct file reviews to identify any
activities that warrant additional oversight.4 Additionally, officials in
each field office we visited stated that they are required to monitor
projects in HUD's Real Estate Management System, the official source of
data on HUD's multifamily housing portfolio that maintains data on
properties and to conduct risk assessments on their properties at least
once a year to identify those facilities that are designated as troubled
or potentially troubled based on their physical inspection, financial
condition, and other factors.

Field offices also varied in the utilization of HUD's Online Property
Integrated Information Suite (OPIIS), a centralized resource for HUD
multifamily data and property analysis. According to officials at HUD
headquarters, field office officials can use OPIIS to conduct a variety of
portfolio analysis and view risk assessments on their properties to better
assist them in overseeing their portfolios. For example, OPIIS contains an
Integrated Risk Assessment score that combines financial, physical, loan
payment status history, and other data into a score that can be used to
identify at-risk properties and prioritize workloads. However, four of the
five field offices that we visited did not frequently use OPIIS. Some of
these offices used the system to develop risk rankings for their
properties or in trying to obtain data about a property, but none of them
regularly used the system for the monitoring of Section 232 loans. The one
field office that utilized OPIIS more frequently did so because the system
partly incorporates a loan risk and rankings system that the field office
had previously developed for its own use. Officials in this field office
stated that an issue with OPIIS is that it is not designed to capture
important, specific financial information that is unique to some Section
232 loans, such as expenses on food or medication.

Appendix III  HUD Officials Addressed Some Issues Raised by the Agency's Inspector
General in 2002, but Several Key Items Remain Unresolved

In a 2002 report, the Department of Housing and Urban Development's (HUD)
Inspector General found that HUD's Office of Housing did not have adequate
controls to effectively manage the Section 232 program.1 Because of these
weaknesses, the Inspector General found that HUD lacked assurance of the
effective operation of Section 232 properties. The Inspector General noted
that the Office of Housing had already taken steps to develop an action
plan to address the weaknesses identified by a task force, but that time
frames had not yet been established. The Inspector General recommended
that the Office of Housing establish specific time frames for implementing
the corrective actions for the 10 weaknesses identified by the task force
and that it monitor the actions to ensure timely and effective completion.

HUD officials developed a plan to correct the 10 control weaknesses
identified by the Office of Housing, which included the current status of
each action and specific target dates to complete the corrective actions.
According to the Inspector General, HUD has taken action to address 2 of
the 10 control weakness findings identified by the Office of Housing Task
force and for which the Inspector General recommended that timelines for
corrective actions be established.

The eight unresolved control weaknesses identified by the Office of
Housing task force are all contingent upon approval of the proposed
revisions to the regulatory agreements. However, the proposed revisions
have been awaiting approval since August 2, 2004. According to HUD
officials, the delay is a result of numerous administrative issues, which
include changes in FHA management and extended public comment periods.

The addressed control weaknesses and respective corrective actions
involved loan underwriting. The Inspector General agreed with the Office
of Housing task force, which found that HUD's underwriting process for
Section 232 properties needed to be strengthened and that HUD needed to
complete market studies and background checks of applicants as part of the
process. The Inspector General also agreed with the Office of Housing task
force's finding of potential problems associated with the nonrecourse
nature of HUD Section 232 loans. In particular, it found that HUD needed
to strengthen the regulatory agreements and underwriting process for
Section 232 loans if these mortgages were to remain nonrecourse and to
avoid potential increase in the portfolio claim rate. HUD addressed these
findings by adding requirements for operators, reviews of operators'
financial statements, and professional liability insurance. Furthermore,
applications for projects that are considered marginal are rejected.

The eight remaining control weaknesses for which HUD has not fully
completed its corrective actions are as follows:

HUD lacks a handbook detailing monitoring requirements for nursing homes
and assisted living facilities. The Inspector General found that HUD did
not have a handbook specific to the Section 232 program monitoring
requirements ensuring that all facilities follow the applicable regulatory
agreements and state and federal requirements. In our site visits to five
field offices, we found inconsistencies in the extent to which oversight
procedures were followed, such as requiring operators to submit financial
statements. HUD plans to include Section 232 project monitoring
requirements in the "Multifamily Asset Management and Project Servicing
Handbook" once the proposed revisions to the applicable regulatory
agreements have been approved. In addition, HUD headquarters officials
told us that they plan to issue updated guidance on loan oversight for
Section 232 properties while awaiting approval of the proposed revisions
to the regulatory agreements.

HUD's regulatory agreement does not include specific requirements for
Section 232 properties. The Inspector General found that the regulatory
agreement for owners lacked requirements for Section 232 properties, such
as compliance with Medicare and Medicaid guidelines. The Inspector General
also found inconsistencies between the requirements for facilities
operated by the owners and those operated by a separate entity. The
Inspector General recognized that these omissions created an inability for
HUD to control the activities of operators and ultimately created risk to
the General Insurance/Special Risk Insurance Fund. HUD's proposed
revisions to the regulatory agreements have provisions that address these
concerns; however, they are still awaiting approval.

The Financial Assessment Subsystem (FASS) does not allow the owner and
operator to submit annual financial statements electronically, denying HUD
the ability to use the financial check and compliance feature in the
system. The Inspector General found that the Real Estate Assessment
Center's (REAC) FASS did not include all Section 232 properties.
Furthermore, operators were not required to submit annual financial
statements electronically through the system. HUD headquarters officials
agreed that, while operators are unable to submit annual financial
statements electronically, FASS has allowed electronic submissions from
owners since the system's inception. However, the Inspector General found
that because operator financial statements are not required to be
submitted electronically, HUD is unable to utilize the financial and
compliance checks performed within the system to identify and follow up on
deficiencies. HUD plans to modify FASS to allow electronic submission of
operator financial statements; however, implementation has been delayed by
funding problems and approval of the proposed revision to the operator
regulatory agreement.

The Office of Housing needs to improve monitoring and legal tools to
provide early indication of possible default. The Office of Housing task
force identified a need for improved monitoring and legal tools to provide
early indication of potential default. To better understand issues related
to monitoring loans, HUD's Office of Evaluation completed several studies
on Section 232 program performance.2 As of April 2006, all of these
studies remain in draft form. Also, to aid in monitoring, HUD has proposed
revisions to the applicable regulatory agreements to require that owners
and operators submit annual inspection reports and inform HUD of state or
federal violations. These reports can be an early indicator of quality of
care concerns and possible claim. However, the proposed revisions to the
regulatory agreements have not been made final.

The Office of Housing staff needs additional training on servicing nursing
homes and assisted living facilities. The Inspector General identified
that project managers did not have sufficient training on reviewing
Section 232 properties and dealing with the issues unique to Section 232
properties. HUD's management plan states that, as of September 2004, REAC
has conducted financial statement analysis for HUD hubs for the last 2
fiscal years. HUD has also proposed training specific to Section 232
program financial analysis upon approval of revisions to the applicable
regulatory agreements and subject to the availability of funds. However,
HUD headquarters officials stated that there were very limited funds
available for training.

Certain conditions lead to loss of Certificate of Need (CON) or license.
The HUD Inspector General identified that, in some states, the CON and
operating licenses may not transfer with the property. Consequently, an
operator may hold these operational documents and take them with them upon
termination of the lease. Without these documents, a facility is not
viable as a residential care facility and its value is significantly
diminished. This presents a large risk to HUD should the loan go to claim
or should HUD have to acquire the property. HUD's proposed revisions to
the applicable regulatory agreements address this concern by categorizing
these operational documents as part of the mortgaged properties.

Receivables need to be included in the relevant legal documents to
strengthen HUD's control over assets of the property in case of regulatory
agreement violations.3  The Inspector General established that the Section
232 security agreement language was too broad to ensure that all property
assets are covered by the mortgage. To address this concern, HUD proposed
revisions to the applicable regulatory agreements to include receivables
in the personalty pledged as security for the mortgage. Additionally, HUD
proposed added language in the owner regulatory agreement requiring the
owner to execute a security agreement and financial statement upon all
items of equipment and receivables.

Field offices do not have consistent procedures for using different
addendums for mortgages, regulatory agreements, and security agreements.
The Office of Housing's task force found inconsistencies in the field
offices' use of legal agreements between HUD and owners and operators,
such as differing addendums to mortgages, regulatory agreements, and
security agreements. We also found similar discrepancies during our five
site visits. For example, only one office used addendums to HUD's legal
agreements to prevent operators from keeping these operational documents
once the lease terminates. HUD has proposed revisions to the regulatory
agreements, and once they are approved and implemented all offices will
use the same legal documentation. In the interim, HUD headquarters
officials told us they plan to provide field offices with updated guidance
on Section 232 loan oversight.

Appendix IV  HUD's Use of Proxy Data for Refinance Loans

As discussed earlier in this report, we question the Department of Housing
and Urban Development's (HUD) use of Section 207 loans as a proxy for
Section 232 loans in the claim regression that is part of HUD's credit
subsidy estimates. This appendix provides greater detail on our analysis.

HUD's Comparison Did Not Allow for Differences in the Age of Loans

Cumulative claim rates are generally compared for a set period of time and
for loans from the same years of origination. However, HUD calculated the
cumulative claim rates without making these adjustments, which confounds
claim differences between programs with differences due solely to timing.
HUD calculated the cumulative claim rates for each program by taking the
total number of loans that went to claim during a 14- year time period and
dividing this by the total number of loans in that same time period. In
this case, HUD was comparing a program that has been expanding over time,
the Section 232 program, with a program that has had less loan volume in
recent years, the Section 207 program. From 1992 to 1998, HUD insured
1,434 Section 207 loans. From 1999 to 2005, HUD insured 870 Section 207
loans. As a result, HUD has been comparing the claim rate of loans that
have had very little time in which to default (Section 232 loans had an
average age of 4 years) with the claim rate of loans that have had
substantial time in which to default (Section 207 loans had an average age
of 7.5 years). A comparison between two programs' claim rates should allow
for differences in the age of the loans. HUD officials also told us that
they have not analyzed the comparability of these two loan types in terms
of their prepayment rates.

Substantial Differences Exist between Section 207 and Section 232 Loans

To examine the comparability of the Section 207 and Section 232 loans, we
compared the conditional claim and prepayment rates of the two types of
loans. An analysis of conditional claim and prepayment rates compares
claim and prepayment probabilities for loans of the same age, so that
comparisons based on loans of widely varying ages are avoided.

We found that the Section 207 loans generally had lower and, in some
cases, significantly lower conditional claim rates than the Section 232
loans. The differences were greater in the later years when loans more
often go to claim. (see fig. 7). For example, the conditional claim rate
for Section 207 loans in fiscal year 8 was .14 percent as compared with a
conditional claim rate of 3.88 percent for Section 232 loans in fiscal
year 8.

Figure 7: Conditional Claim Rates Are Different for Section 232 and
Section 207 Refinance Loans

Note: Through year 8, there are at least 200 loans in each category of
loan for each conditional claim rate year. Beyond year 8, the loan numbers
are small (particularly for Section 232 loans), and conclusions are less
reliable.

We found that Section 207 loans had generally higher, and sometimes
significantly higher, conditional prepayment rates compared to Section 232
loans. The differences were greater in the later years when loans more
often are prepaid. (see fig. 8). For example, the conditional prepayment
rate for Section 207 loans in fiscal year 8 was 21.72 percent as compared
to a conditional prepayment 11.25 percent for Section 232 loans in fiscal
year 8 (making the conditional prepayment rate for the Section 207 loans
93 percent higher than the conditional prepayment rate for Section 232
loans).

Figure 8: Conditional Prepayment Rates are Different for Section 232 and
Section 207 Refinance Loans

Note: Through year 8, there are at least 200 loans in each category of
loan for each conditional prepayment rate year. Beyond year 8, the loan
numbers are small (particularly for Section 232 loans), and conclusions
are less reliable.

Additionally, we examined and compared cumulative 5-year claim and
prepayment rates. Section 207 loans had a 5-year cumulative claim rate of
3 percent, while for the Section 232 loans it was approximately 6.7
percent. The 5-year cumulative prepayment rate for Section 207 loans was
about 27 percent, while for Section 232 loans it was about 11 percent.

Appendix V  Comments from the Department of Housing and Urban Development

Appendix VI  GAO Contact and Staff Acknowledgments

David G. Wood (202) 512-8678

In addition to the individual named above, Paul Schmidt, Assistant
Director; Austin Kelly; Tarek Mahmassani; John McGrail; Andy Pauline; Carl
Ramirez; Richard Vagnoni; Wendy Wierzbicki; and Amber Yancey-Carroll made
key contributions to this report.

(250238)

www.gao.gov/cgi-bin/getrpt? GAO-06-515 .

To view the full product, including the scope

and methodology, click on the link above.

For more information, contact David G. Wood at (202) 512-6878 or
[email protected].

Highlights of GAO-06-515 , a report to congressional addressees

May 2006

RESIDENTIAL CARE FACILITIES MORTGAGE INSURANCE PROGRAM

Opportunities to Improve Program and Risk Management

Through its Section 232 program, the Department of Housing and Urban
Development's (HUD) Federal Housing Administration (FHA) insures
approximately $12.5 billion in mortgages for residential care facilities.
In response to a requirement in the 2005 Consolidated Appropriations
Conference Report and a congressional request, GAO examined (1) HUD's
management of the program, including loan underwriting and monitoring; (2)
the extent to which HUD's oversight of insured facilities is coordinated
with the states' oversight of quality of care; (3) the financial risks the
program poses to HUD's General Insurance/Special Risk Insurance (GI/SRI)
Fund; and (4) how HUD estimates the annual credit subsidy cost for the
program.

What GAO Recommends

GAO recommends, among other things, that the HUD Secretary establish a
process for sharing practices among field offices, assure appropriate
levels of staff with appropriate expertise, and incorporate reviews of
federal or state inspection reports into loan monitoring. GAO also
recommends that HUD explore factoring additional information into its
credit subsidy model. In written comments, HUD agreed with all of GAO's
recommendations except exploring the value of adding certain factors to
its credit subsidy model.

While HUD's decentralized program management allows its 51 field offices
flexibility in their specific practices, GAO found differences in the
extent to which staff in the five field offices it visited were aware of
current program requirements. For example, four offices were unaware of
required addendums to the programs' standard regulatory agreement.
Further, while individual offices had developed useful practices for loan
underwriting and monitoring, they lacked a mechanism for systematically
sharing such practices with other offices. Also, field office officials
were concerned about adequate current or future levels of staff
expertise-a critical factor in managing program risk in that health care
facility loans are complicated and require specialized knowledge and
expertise.

FHA requires a review of the most recent annual state-administered
inspection report for state-licensed facilities applying for program
insurance, and recommends, but does not require, continued monitoring of
such reports for facilities once it has insured them. Four of the five HUD
field offices GAO visited do not routinely collect annual inspection
reports for their insured facilities. While the reports are but one of
several monitoring tools, they provide potential indicators of future
financial risk. HUD has proposed revising its standard regulatory
agreements to require insured facility owners or operators to submit
annual inspection reports and to report notices of violations. However,
the proposed revisions have been awaiting approval since August 2004, and
the implementation date is uncertain.

The Section 232 program accounts for only about 16 percent of the GI/SRI
Fund's total unpaid principal balance, but program and industry trends
pose potential risks to the Section 232 program and to the GI/SRI Fund.
For example, in recent years the program has insured increasing numbers of
assisted living facility loans and refinancing loans, for which there are
limited data available to assess long-term performance. Other potential
risk factors include increasing prepayments (full repayment before loan
maturity) and loan concentration in several large markets and among
relatively few lenders. Projected shifts in demand for residential care
facilities could affect currently insured facilities and the overall
market for the types of facilities that HUD insures under the program.

To estimate the program subsidy cost, HUD uses a model to project cash
flows for each loan cohort (the loans originated in a given fiscal year)
over its entire life. HUD's model does not explicitly or fully consider
certain factors, such as loan prepayment penalties, interest rate changes,
or differences in loans to different types of facilities, and uses some
proxy data that is not comparable to Section 232 loans. The model's
exclusion of potentially relevant factors and it use of this proxy data
could affect the reliability of HUD's credit subsidy estimates.

GAO's Mission

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

Obtaining Copies of GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( www.gao.gov ). Each weekday, GAO posts newly
released reports, testimony, and correspondence on its Web site. To 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, Waste, and Abuse in Federal Programs

Contact:

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

Congressional Relations

Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [email protected] (202) 512-4800 U.S.
Government Accountability Office, 441 G Street NW, Room 7149 Washington,
D.C. 20548
*** End of document. ***