Hospital Mortgage Insurance Program: Program and Risk Management 
Could Be Enhanced (28-FEB-06, GAO-06-316).			 
                                                                 
Under its Hospital Mortgage Insurance Program, the Department of 
Housing and Urban Development's (HUD) Federal Housing		 
Administration (FHA) insures nearly $5 billion in mortgage loans 
for the renovation or construction of hospitals that would	 
otherwise have difficulty accessing capital. In response to a	 
requirement in the 2005 Consolidated Appropriations Conference	 
Report, GAO examined (1) the design and management of the	 
program, as compared with private insurance, (2) the nature and  
management of the relationship between HUD and the Department of 
Health and Human Services (HHS) in implementing the program, (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.						 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-316 					        
    ACCNO:   A47918						        
  TITLE:     Hospital Mortgage Insurance Program: Program and Risk    
Management Could Be Enhanced					 
     DATE:   02/28/2006 
  SUBJECT:   Facility construction				 
	     Facility maintenance				 
	     Facility repairs					 
	     Financial analysis 				 
	     Hospitals						 
	     Interagency relations				 
	     Mortgage loans					 
	     Mortgage programs					 
	     Mortgage protection insurance			 
	     Performance measures				 
	     Private sector					 
	     Program evaluation 				 
	     Program management 				 
	     Risk management					 
	     Subsidies						 
	     Program implementation				 
	     FHA Hospital Mortgage Insurance Program		 
	     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-06-316

     

     * Report to Congressional Committees
          * February 2006
     * HOSPITAL MORTGAGE INSURANCE PROGRAM
          * Program and Risk Management Could Be Enhanced
     * Contents
          * Results in Brief
          * Background
          * FHA's Selection Process Includes Additional Steps Compared with
            Private Insurers, but Monitoring Techniques Are Similar
          * Agencies Coordinate Key Activities, but FHA Does Not Track Most
            Performance Measures
               * FHA and HHS Coordinate Their Client Screening, Application
                 Review, and Loan Monitoring Activities
               * FHA Has Not Used Performance Measures to Manage the Program
               * Program Guidance Is Not Up to Date
          * Potential Risks Exist Although FHA Has Mitigation Strategies in
            Place
               * The Hospital Program Accounts for a Relatively Small Share
                 of the Broader GI/SRI Fund, and Has Recovered a Majority of
                 All Claims
               * Loan Performance and Market Trends Reveal Sources of
                 Potential Risks
               * Industry Trends Pose Risks to Hospitals, Including Those
                 with FHA Insured Mortgages
               * FHA Uses Tools to Mitigate Risk
          * HUD's Model for Estimating Credit Subsidy Costs Excludes
            Potentially Relevant Factors
          * Conclusions
          * Recommendations for Executive Action
          * Agency Comments and Our Evaluation
     * Objectives, Scope, and Methodology
     * FHA and HHS' Responsibilities in FHA's Hospital Mortgage Insurance
       Program Loan Cycle
     * FHA Assessed Performance Using 2 of 22 Performance Measures Included
       in the 2002- 2005 Memorandum of Agreement
     * Comments from the Department of Housing and Urban Development
     * Comments from the Department of Health and Human Services
     * GAO Contact and Staff Acknowledgments

Report to Congressional Committees

February 2006

HOSPITAL MORTGAGE INSURANCE PROGRAM

Program and Risk Management Could Be Enhanced

Contents

Table

Figures

February 28, 2006Letter

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 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 Department of Housing and Urban Development (HUD), through the Federal
Housing Administration's (FHA) Hospital Mortgage Insurance Program,
insures loans to finance the renovation or construction of hospitals.
Through an interagency agreement, the Department of Health and Human
Services (HHS) administers certain aspects of this program based upon its
health care and hospital expertise. The program is intended to protect
lenders against losses they might incur if hospitals fail to make their
mortgage payments. As of December 31, 2005, FHA reported that it insured
nearly $5 billion in outstanding mortgages under the program.

The Hospital Mortgage Insurance Program is one of several programs
included in FHA's General Insurance/Special Risk Insurance (GI/SRI) fund;
other programs in the GI/SRI fund are much larger and include mortgage
insurance for various types of multifamily housing projects and for
nursing homes. Pursuant to the Federal Credit Reform Act of 1990, for
budget purposes HUD must annually estimate the credit subsidy for the
program.1 The credit subsidy cost for loan guarantees is the present value
of cash flows over the life of the loan from estimated payments by the
government (for defaults, delinquencies, and other payments) minus
estimated payments to the government (for loan origination and other fees,
penalties, and recoveries); it excludes administrative costs. Such
estimates are important indicators of the full cost of programs to the
government.

The 2005 Consolidated Appropriations Conference Report mandated that we
review two FHA insurance programs-those for hospitals and nursing homes.
This report provides the results of our evaluation of the Hospital
Mortgage Insurance Program.2 For this report, we reviewed (1) the design
and management of the program, as compared with private insurance; (2) the
nature and management of the relationship between HUD and HHS in
implementing the program; (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.

To address these objectives, we reviewed program manuals and documentation
of loan processing procedures and analyzed program financial data, which
we determined to be reliable for the purposes of our review. We also
reviewed documentation of HUD's credit subsidy model and applicable
program laws, regulations, and policy statements. We interviewed officials
from FHA's Office of Insured Health Care Facilities and the Division of
Facilities and Loans within HHS' Health Resources and Services
Administration. We also interviewed health care and hospital associations,
mortgage and investment banking firms, rating agencies, and private bond
insurers. Our review of the hospital program did not include an evaluation
of underwriting criteria, construction monitoring, or the need for the
program. See appendix I for more detailed information on our objectives,
scope, and methodology.

We conducted our work in New York, New York; Chicago, Illinois; Paterson,
New Jersey; Rockville, Maryland; and Washington, D.C., between February
2005 and January 2006 in accordance with generally accepted government
auditing standards.

Results in Brief

The Hospital Mortgage Insurance Program insures the mortgages of hospitals
that are financially riskier than those that can obtain private insurance,
but it shares some management techniques with private insurers. While
FHA's process for reviewing mortgage insurance applications includes more
steps, such as a preapplication meeting and review by an independent
consultant, and generally takes longer, FHA officials believe these extra
steps are justified given the generally riskier nature of the hospitals
applying. Once it insures a hospital mortgage, FHA monitors the loan using
many of the same techniques that private insurers use. For example, both
FHA and private insurers identify the riskiest loans in their portfolios
for closer monitoring. They also periodically review hospital financial
statements and management activities and can require hospitals
experiencing financial difficulties to use consultants for needed
expertise.

FHA and HHS work together throughout the process of reviewing mortgage
insurance applications and monitoring active loans, guided by a multiyear
Memorandum of Agreement between the two agencies. However, FHA has not
used the agreement's performance measures to manage the program, and its
program guidance is not up to date. FHA and HHS use joint working groups
to carry out certain activities. For example, client service teams, which
can be composed of HHS staff, FHA staff, or both, review application
materials. Also, senior FHA and HHS officials meet weekly to discuss
insurance applications, as well as insured hospitals that are experiencing
difficulties. While the agencies coordinate in program implementation, FHA
does not collect data to track most of the performance measures, including
those for coordinated or HHS-delegated tasks. For example, one performance
measure is designed to capture the soundness of the analysis of a
hospital's application, but FHA does not collect data to assess this
performance measure. Other performance measures are not measurable or
objective. According to FHA, the measures are intended to communicate
expectations, and it has not tracked most performance measures because
there have not been enough problems to warrant a tracking system. Finally,
FHA has not updated the program handbook, which contains program
eligibility requirements, policies, and procedures, since 1984.

The Hospital Mortgage Insurance Program is small compared with other
programs in the GI/SRI fund-accounting for about 7 percent of the amount
of all active loans insured by the fund-and the losses from claims have
been relatively low. The program has not experienced a claim since 1999.
Despite the program's relatively small size, some program and market
trends may pose risks. For example, the geographic concentration of
insured hospitals located in the state of New York, while decreasing,
makes the program vulnerable to state policies and regional economic
conditions. In addition, literature on hospital industry market trends
generally predicts reductions in hospital revenues along with increasing
capital needs-conditions that could increase the risk of FHA insured loans
going to claim. FHA has established mitigation strategies to address some
potential risks. For example, FHA requires hospitals to establish cash
reserve funds equal to 2 years worth of mortgage payments. These funds may
be used to help hospitals through temporary financial crises and prevent
their lenders from filing an insurance claim. Also, FHA's identification
of its riskiest hospitals enables the agency to prioritize those hospitals
that need additional monitoring and assistance. In addition, since 1999,
FHA has had goals for geographically diversifying the hospital mortgage
insurance portfolio. Though it has made efforts to diversify the
portfolio, FHA does not have a formal strategy for reaching its
diversification goals.

HUD uses a model for estimating annual credit subsidies, or program costs
over the life of the outstanding loans insured, that does not explicitly
consider the impacts of some potentially important factors. HUD's model-an
important tool for estimating the cost of the hospital program to the
government-incorporates factors and assumptions about how the loans will
perform, including estimated claim and recovery rates, which are
consistent with guidance issued by the Office of Management and Budget
(OMB). Although the number of claims paid since the program's inception is
small, HUD assumes that the lenders for some active hospitals will file
claims for insurance and, therefore, increases its estimated claim rate.
In 10 of the 14 years that HUD has been estimating the cost of the
hospital program under credit reform, HUD has calculated a negative
subsidy rate, meaning that estimated cash inflows (including fees,
premiums, and recoveries on defaulted loans) have been greater than
estimated cash outflows (including claims and certain program expenses).
HUD's model does not explicitly consider the potential impacts of
prepayment penalties or restrictions, which can influence cash flows
through the timing of prepayments, collection of premiums, and claims.
HUD's model also does not consider the initial debt-service coverage
ratios of hospitals (an indicator of a borrower's ability to make regular
mortgage payments) at the point of loan origination when estimating future
claims, although, according to some economic studies, this ratio is an
important factor in modeling default risk. HUD does not include these
factors in its model because, according to HUD, it does not collect data
on prepayment restrictions and because debt-service coverage ratios, among
other things, do not vary. We found that data on prepayment restrictions
are readily available to FHA. Further, our analysis of projected
debt-service coverage ratios for hospitals that applied for mortgage
insurance between 2002 and 2005 found that these ratios varied (with the
highest being over twice that of the lowest) and thus can be useful in
assessing relative risk.

This report contains recommendations designed to improve FHA's management
of the Hospital Mortgage Insurance Program and reduce risks associated
with the geographic concentration of the portfolio. We provided draft
copies of this report to HUD and HHS. In its response, HUD's Assistant
Secretary for Housing concurred with our recommendations and noted actions
that it plans to take. HUD also stated that the report did not adequately
emphasize the program's accomplishments. HHS concurred with HUD's
observations.

Background

In 1968, the Congress added Section 242 to the National Housing Act
establishing the Hospital Mortgage Insurance Program to address a serious
shortage of hospitals and the need for existing hospitals to expand and
renovate. Through this program, FHA insures the loans lenders make for the
construction and renovation of hospitals.3 Since the inception of the
program, FHA has insured 341 hospital mortgages for $11.9 billion in 42
states and Puerto Rico. As of the end of calendar year 2005, FHA was
insuring 74 hospital mortgages totaling nearly $5 billion.4 The number of
loans insured annually has increased in recent years, from 2 in fiscal
year 2001 to 11 in fiscal year 2005 (see table 1). According to the House
report accompanying the Hospital Mortgage Insurance Act of 2003, which
revised the standards for determining the need and feasibility for
hospitals, as well as eligibility requirements for small, rural hospitals,
hospitals face significant financial challenges when providing care to
patients who are covered by Medicare and Medicaid, as well as those that
are uninsured. At the same time, improvements in technology and health
care knowledge necessitate capital improvements such as additions and
renovations to existing buildings.

Table 1: Number of New Loans Insured through the Hospital Mortgage
Insurance Program Since 2001

Fiscal year                                                  Loans insured 
2005                                                                    11 
2004                                                                     6 
2003                                                                     7 
2002                                                                     1 
2001                                                                     2 

Source: FHA.

FHA's Office of Insured Health Care Facilities and HHS' Division of
Facilities and Loans coordinate to implement the hospital program. HUD has
statutory responsibility for the program based on the FHA's experience
with promoting housing construction through housing mortgage insurance
programs. As such, HUD is fully responsible for management of the program,
including developing and proposing legislation, policy development,
strategic planning, and approval of applications and loan documents. The
House Committee on Banking and Currency, in recommending that HUD be given
this responsibility, expected HUD to draw upon HHS's hospital expertise to
devise standards for insuring hospitals' mortgages. Through an interagency
agreement, HUD formally delegates authority to HHS to assist in the review
of applications for mortgage insurance and the monitoring of insured
loans. HHS is also given full responsibility for construction monitoring.
See appendix II for additional information about FHA and HHS's loan
processing responsibilities.

FHA's Hospital Mortgage Insurance Program generally serves the segment of
the market consisting of hospitals that are too risky to obtain private
bond insurance but are strong enough to pass FHA's underwriting tests.
Mortgage insurance, like private bond insurance, guarantees that lenders
will be paid if the hospital stops making payments on its loan. In
addition, both mortgage insurance and private bond insurance are forms of
credit enhancement and improve the credit rating of the underlying debt
for the insured entity, resulting in a lower interest rate for the loan.
Hospitals with FHA-insured mortgages automatically receive
investment-grade ratings (AA or AAA) because the reliability of the cash
flows from the mortgage note are rated on the insurer's, not the
hospital's, ability to repay the debt. Both FHA and HHS officials and
private insurers agree that FHA's Hospital Mortgage Insurance Program
serves a different market than private insurers. According to FHA and HHS
officials, FHA insures loans that are too risky, too small, or too large
for private insurers, or are located in a market not served by private
insurers.

For the Hospital Mortgage Insurance Program, if a hospital fails to make
any payment due under the mortgage, the mortgage is in default. If a
default continues for 30 days, the lender is entitled to receive benefits
from FHA. FHA may pay claims in either cash or debentures.5 

Federal agencies that provide direct loans or loan guarantees are required
by the Federal Credit Reform Act of 1990 to estimate the expected cost of
programs by estimating or predicting their future performance and
reporting the costs to the government in their annual budgets. Such
estimates are important in that they more accurately measure the
government's costs of federal loan programs and permit better cost
comparisons among different programs. Under credit reform procedures, the
cost of loan guarantees, such as mortgage insurance, is the net present
value of all expected future cash flows, excluding administrative costs.6
For guarantees, cash inflows consist primarily of fees and premiums
charged to insured borrowers and recoveries on assets, and cash outflows
consist mostly of payments to lenders to cover the cost of claims.
Agencies discount projected future cash flows to the year in which the
guaranteed loan was disbursed. The discounted cash flows are the estimated
budgetary cost or gain of the cohort of loans obligated in a given
fiscalyear.7 The net present value of each cohort's estimated cash flows
is expressed as a percentage of the volume of guaranteed loans in the
cohort-that is, a subsidy rate. Agency managers are responsible for
accumulating relevant, sufficient, and reliable data on which to base
their credit subsidy estimates. OMB has final responsibility for
determining subsidy estimates, in consultation with agencies.

FHA's Selection Process Includes Additional Steps Compared with Private
Insurers, but Monitoring Techniques Are Similar

FHA requires hospitals to take certain steps, both before they apply for
mortgage insurance and as a part of the application process, that private
insurers do not mandate. These additional steps are used because FHA
insures mortgages that are generally riskier than those using private bond
insurance. For example, before they apply for mortgage insurance, FHA
advises hospitals to compare their financial status with the program's
minimum requirements. If they meet these requirements, FHA requires
hospitals to submit market and financial information so that FHA can make
a preliminary assessment about the project and determine whether to
conduct a preapplication meeting with the applicant to discuss the
project. None of the private insurers that we met with have similar
preapplication processes.

After these preapplication steps are met, FHA's application process
includes additional steps compared with those of private bond insurers.
FHA requires hospitals to submit a financial feasibility study containing
historic and forecasted financial statements and ratios, a financing plan,
and information about market demand, among other things. In addition, FHA
hires consultants to evaluate the feasibility of each proposed project as
an additional, independent check on the viability of the project. While
the private bond insurers that we met with review the types of information
included in feasibility studies, they do not require hospitals to submit
such studies and do not hire consultants to assess the feasibility of
proposed projects.

FHA's application process also includes a final level of review that is
absent from private bond insurer processes. After an application for
mortgage insurance has gone through underwriting and been reviewed by an
independent consultant, it is considered by the program management group,
a group of senior-level FHA and HHS staff. FHA also refers to this group
as its "credit committee." Similarly, private bond insurers also consider
applications within a credit committee structure. However, while private
bond insurers make final insurance decisions through their credit
committees, FHA has an additional layer of review. Based upon input from
the program management group, the Director of FHA's Office of Insured
Health Care Facilities makes a recommendation to the FHA Commissioner, who
then makes the final decision.

It generally takes FHA longer to process applications than it takes
private bond insurers. According to program data, it took FHA an average
of 265 days to process the 11 applications for hospital mortgage insurance
that it endorsed in fiscal year 2005.8 According to the FHA, processing
times vary with the complexity of the project and may be affected by
issues requiring a hospital to rethink or resubmit its application,
including issues that are beyond HUD's control. In contrast, according to
the private bond insurers and investment bankers that we interviewed, it
generally takes private insurers up to 60 days to process an insurance
application, sometimes less. While FHA's average processing time is higher
than private bond insurers, it has decreased from an average of 399 days
in fiscal year 1999. According to FHA, processing times have improved as a
result of implementing the preliminary review process, which disqualifies
hospitals that don't meet the program's minimum requirements.

FHA uses many of the same techniques that private insurers use to monitor
insured hospitals. Both FHA and private bond insurers identify the
riskiest hospitals in their portfolio for closer monitoring. Since
November 1999, FHA has placed on a priority watch list hospitals it
determines are at risk of having a claim filed within the next 12 months.
FHA considers a hospital for inclusion on the priority watch list if
certain financial criteria are not met. For example, if the ratio
measuring a hospital's ability to pay its mortgage payments with cash
generated from current operations (the debt service coverage ratio) falls
below an acceptable level, the hospital may be placed on the watch list.9
A hospital can also be placed on the list if FHA becomes aware of other
conditions at the hospital, such as management or personnel problems. As
of December 2005, FHA data showed that 11 of the 59 insured hospitals are
on this list, representing an unpaid (insured) principal balance of
approximately $762 million. Private insurers also assess the risk of the
hospitals that they insure in order to identify those that should be
monitored more closely. For example, one private bond insurer explained
that they monitor compliance with loan agreements by reviewing financial
statements, documentation of payer mix (i.e., proportion of reimbursement
from Medicare, Medicaid, private insurance, etc.), and notices of
litigation, among other things.

As a part of their monitoring efforts, both FHA and private bond insurers
monitor agreements that exist between themselves and the insured
hospital.10 These agreements specify the requirements that the insured
hospital must comply with in order to maintain the insurance. Agreements
may cover issues such as the debt-service coverage ratio; liquidity, or
the ability to convert assets to cash; and activities that a hospital
cannot do without approval by the insurer. Both FHA and private insurers
require hospitals to request waivers from agreement requirements if they
are not going to meet them. FHA and private insurers monitor hospitals'
compliance with these agreements through various means, such as by
evaluating changes in indicators of financial performance, as reported in
hospitals' financial statements. For example, one private bond insurer
reported that it monitors hospitals' cash on hand, and FHA monitors
hospitals' debt-service coverage ratios. FHA and private insurers monitor
financial statements and other documentation quarterly and annually,
respectively, and more frequently for hospitals that are experiencing
financial difficulty. Both FHA and private insurers require hospitals that
are not in compliance to correct violations within specific time frames.

Both FHA and private insurers can require hospitals experiencing financial
difficulties to hire consultants. In some cases, FHA will pay for
consultants to identify and suggest solutions to hospitals' financial
difficulties. According to FHA, since fiscal year 2000, it has paid $1.3
million for consultant's studies of 27 hospitals.11 However, FHA can also
require hospitals to hire and pay for consulting services on their own.
Similarly, private insurers can require hospitals to hire consultants to
assist them with identifying and addressing problems. The requirement for
a hospital to hire a consultant can be triggered if a hospital is not in
compliance with its loan agreements, according to both FHA and private
bond insurers.12

Agencies Coordinate Key Activities, but FHA Does Not Track Most
Performance Measures

FHA and HHS coordinate key activities, including screening applicants,
underwriting loans, and monitoring insured hospitals. While FHA has
established performance measures for both coordinated tasks and tasks
delegated to HHS through an interagency agreement, it does not collect
data with which to assess most of these measures. FHA's primary guidance
for the program has not been updated in over 20 years and, therefore, does
not reflect key changes in eligibility criteria.

FHA and HHS Coordinate Their Client Screening, Application Review, and
Loan Monitoring Activities

FHA and HHS coordinate to implement the hospital program based upon FHA's
experience with promoting housing construction through its housing
mortgage insurance programs and HHS's hospital and health care expertise.
As previously noted, FHA is responsible for management of the program and
formally delegates certain responsibilities to HHS. A Memorandum of
Agreement (MOA) between FHA and HHS outlines the duties and
responsibilities of each agency in carrying out the Hospital Mortgage
Insurance Program, including coordinated activities and tasks that FHA
delegates to HHS.13 In accordance with this agreement, both FHA and HHS
staff are involved with the screening of applicants during the
preapplication meetings. FHA's policy is to include senior FHA staff and
legal counsel, the account executive and client service team members (both
of which can be either FHA or HHS staff), and engineering staff from HHS,
among others, in such meetings.14 This policy helps insure that
preapplication discussions with applicants are coordinated between FHA and
HHS.

FHA and HHS also coordinate activities during the underwriting review
portion of the application process, which is the process used by FHA to
assess the risk of a potential loan to the GI/SRI fund. The nature of
coordination at this level depends on the staffing of the account
executive and client service team positions, since these positions can be
filled by either FHA or HHS staff or a combination of both. The account
executive and client service team are responsible for underwriting
activities, including analysis of the market and financial feasibility of
the project. In addition, HHS engineers review all design and construction
aspects of the proposed project. Appendix II presents the roles and
responsibilities of each agency in more detail.

FHA and HHS use regular meetings of the program management group to
coordinate additional activities. This group, composed of senior FHA and
HHS staff, meets weekly to assist account executives and client service
teams as they review applications for mortgage insurance and monitor
insured hospitals. Minutes of program management group meetings that we
reviewed show joint FHA and HHS discussion of new applications, as well as
issues associated with the existing portfolio. According to investment
bankers, hospital associations, consulting firms, and selected hospitals
we spoke with, coordination between FHA and HHS is generally seamless.

FHA Has Not Used Performance Measures to Manage the Program

The fiscal years' 2002-2005 MOA between FHA and HHS provides for FHA to
establish performance measures and use them to evaluate tasks.15 While the
MOA between FHA and HHS contains 22 performance measures, FHA has tracked
actual performance for only 2 of these measures, 1 for processing complete
applications within 120 days, and 1 for processing loan modification
requests within 30 days. As a result, it is not possible to evaluate how
well the agencies perform in implementing the program. According to FHA
officials, the agency never intended to track these measures, or use them
as actual measures of performance, but rather to show FHA's expectations
of HHS. Neither HUD's fiscal year 2005 performance plan nor its
performance and accountability report includes other performance measures
for this program. Moreover, OMB did not assess this program as a part of
its fiscal year 2005 Program Assessment Rating Tool (PART), which is used
to assess the performance of federal programs. Appendix III provides more
detailed information about the 22 performance measures contained in the
MOA between FHA and HHS.

Analysis of the two performance measures for which data is collected shows
that FHA is not meeting its performance goals for those measures. Based
upon analysis of data from the Hospital Mortgage Insurance Management
Information System, we determined that FHA did not meet its goal of
processing 75 percent of hospital mortgage insurance applications within
120 days. Although the FHA received no more than 10 applications each year
between fiscal years 2002 and 2005, FHA and HHS never processed more than
2 within 120 days (see fig. 1).

Figure 1: FHA Did Not Process Most Hospital Mortgage Insurance
Applications within Targeted Time Frame

In addition, according to FHA, the agency did not meet its goal of
processing at least 75 percent of loan modification requests within 30
days.16 However, analysis of available data shows that FHA and HHS
improved from processing 45 percent of loan modification requests received
in fiscal year 2002 within 30 days to processing 71 percent in fiscal year
2005.

FHA has not tracked other performance measures related to activities that
are coordinated, or can be done, by both FHA and HHS staff. For example,
according to one performance measure, hospitals with a weakening financial
position should be identified early enough to allow time for the account
executive to provide technical assistance and undertake default prevention
measures. Since such hospitals are identified through the FHA's priority
watch list system, these data are readily available for measurement.
Similarly, another performance measure is designed to capture the
soundness of analysis performed by client service teams, which can include
both FHA and HHS staff, in assessing insurance applications. FHA has also
not tracked this measure.

FHA also does not track performance measures of activities that it
delegates to HHS. For example, one measure is designed to capture the
number of complaints and compliments about HHS's timeliness, helpfulness,
courtesy, and understanding. According to FHA, the agency has not tracked
this or other measures because it has not had enough problems with HHS to
warrant establishing a tracking system and that establishment of such a
system would be both an administrative burden and a poor use of their
resources. However, without collecting appropriate information, FHA cannot
quantify the input it receives about HHS. In addition, FHA has not tracked
performance measures related to construction design and monitoring, which
HHS is responsible for. According to FHA, performance measures exist to
indicate FHA's expectations of HHS's performance, even though HHS's
performance is not tracked.

Several of the performance measures contained in the agreement between FHA
and HHS lack the necessary characteristics of performance measures; that
is, they are not measurable or objective. As a result, they do not provide
useful information about the performance of the hospital program. For
example, the measures related to the number of complaints and compliments
about HHS are not measurable in that they do not specify a quantifiable
threshold for expected performance. As a result, even if FHA tracked
complaints, it is not possible to tell whether performance is meeting
expectations. Other goals lack objectivity in that they require subjective
judgment to assess program performance. As an example, one performance
measure indicates that "plans and specifications do not need major
revisions during the construction process because of significant
architectural or engineering errors." Another indicator states that
"preconstruction meetings are thorough and do not precipitate delays in
application processing." In both cases, the performance measures require
subjective judgment, because they do not make explicit what constitutes
"major," "significant," or "thorough." As we have previously reported,

useful performance information is based upon measurable and objective
performance measures.17 If useful performance information is collected,
managers could use it to identify problems, try to identify the causes of
problems, and/or to develop corrective actions.18 (App. III provides a
complete list of the performance measures.)

While FHA does not track most of the performance measures outlined in the
MOA, FHA's Hospital Mortgage Insurance Management Information System
captures a significant amount of quantitative and qualitative data about
the performance of the program, which could be incorporated into
measurable and objective performance measures. This system captures key
loan processing dates, financial performance data over time, and
documentation of internal meetings and actions performed by both agencies
to assist insured hospitals. Incorporation of this readily-available data
into meaningful performance measures would enable FHA to better assess its
management of the program.

FHA and HHS established a new interagency agreement covering fiscal years
2006 through 2010, which includes many of the same measures as the
previous agreement, including those that are not measurable or objective.
The new agreement also includes a requirement that HHS provide FHA with an
annual report detailing its performance against each of the performance
measures in the agreement. However, this interagency agreement does not
specify whether and how FHA will track its own performance against the
measures.

Program Guidance Is Not Up to Date

FHA's primary guidance for its hospital mortgage insurance program has not
been updated in over 20 years and does not reflect changes to the program
over that time. As a result, this document does not contain current
eligibility requirements, which may cause confusion for potential
applicants. In 1973, FHA published the Mortgage Insurance for Hospitals
Handbook and last updated the handbook in 1984. The purpose of the
handbook is to provide complete information about the processing of
hospital mortgage insurance, including basic program features and
requirements, to hospitals, lenders, sponsors, FHA and HHS personnel, and
all other interested parties. According to FHA, the Office of Insured
Health Care Facilities has not had adequate staff to revise the handbook
and is waiting for a proposed regulation to become final before revising
it. Since the handbook has not been updated since 1984, it does not
contain current eligibility requirements, policies, and processing
procedures. As we have previously reported, internal control standards
applicable to federal programs provide that information should be recorded
and communicated in a timely manner.19

The handbook does not reflect key changes that the Hospital Mortgage
Insurance Act of 2003 made to the program. This act revised the existing
requirement that hospitals applying for FHA mortgage insurance have either
a Certificate of Need or a state-commissioned study of market need;
specifically, it provided that FHA would establish the means for
determining market need and feasibility for hospitals.20 In addition, the
2003 act exempted Critical Access Hospitals (CAH) from the requirement
that at least 50 percent of care must be for general acute-care
patients.21  According to one of the mortgage bankers that we met with,
the handbook causes confusion because hospitals are uncertain about
requirements applicable to them.

As we have previously reported, internal control standards provide that
information, such as changes in eligibility requirements and application
processing procedures, should be communicated in a timely manner. While
FHA publicly communicates program changes through Mortgagee Letters,
updating the Applicant's Guide, distributing copies of its minimum
criteria for consideration, and updating its Web page, it has not
incorporated all of this updated information into the program's handbook.
All documentation, including the handbook, should be updated in a timely
manner.22 Maintaining current documentation is an internal control that
would benefit both those interested in the program and those that
administer the program.

Potential Risks Exist Although FHA Has Mitigation Strategies in Place

The hospital program is a relatively small program within the broader
GI/SRI fund and has a record of recovering claims. Despite its small size,
both program and market trends show risks that could affect the hospital
portfolio. FHA has mitigation strategies in place to address some risks
but does not have a formal strategy to geographically diversify the
hospital loan portfolio.

The Hospital Program Accounts for a Relatively Small Share of the Broader
GI/SRI Fund, and Has Recovered a Majority of All Claims

The Hospital Mortgage Insurance Program comprises a relatively small part
of the GI/SRI fund, representing about 2.9 percent of the GI/SRI's fund's
fiscal year 2006 total commitment authority.23 Moreover, the approximately
$5 billion in loans that FHA currently insures through the program is 6.5
percent of the $77 billion in unpaid principal balance of the fund (see
fig. 2). In addition to being a financially small component of the broader
GI/SRI fund, the Hospital Mortgage Insurance Program has a record of
recovering more than two-thirds of all historical claims, and lenders have
not made a claim on an insured loan since 1999. Since the program's
inception in 1968, there have been a total of 22 claims totaling $225
million. Of this amount, FHA recovered 68 percent, or $153 million.

In addition, the majority of the currently insured loans in FHA's hospital
portfolio are less than 10 years old. According to HUD, 70 percent of
claims have historically occurred prior to a loan's tenth year. Currently,
the loans that have been insured for less than 10 years have an aggregate
unpaid principal balance of $2.8 billion, representing about 57 percent of
the aggregate unpaid principal balance (see fig. 3).24

Figure 3: Proportion of FHA Hospital Loans That Have Been Insured Less
Than 10 Years

Comparing FHA data on selected financial indicators with the criteria the
agency uses to determine the financial health of program applicants shows
some favorable trends but also indicates sources of potential financial
risk (see fig. 4). Specifically, our analysis of program data for calendar
years 2000 to 2004 shows that some insured hospitals increased their
ability to meet their monthly and future mortgage payments. For example,
the median debt-service coverage ratio, a measure of a hospital's ability
to pay its mortgage with cash generated from current operations, increased
from 1.54 to 2.18. While a value of 2.18 for this ratio indicates a low
level of risk, according to FHA criteria, other financial indicators
indicate medium levels of risk.25 For example, the median number of days
of cash on hand and the median  current ratio (which compares a hospital's
current assets to its current liabilities) both improved, yet still
indicate a medium level of risk to the program. Finally, the median
operating margin, which is indicative of a hospital's ability to control
costs and expenses, improved between 2000 and 2004, yet indicates a medium
level of risk based on FHA's criteria.

Figure 4: Selected Median Financial Indicators Show Varying Levels of Risk

aThe N for debt-service coverage ratio differs slightly. It is 36 in 2000,
41 in 2001, 42 in 2002, 44 in 2003, and 49 in 2004.

Median financial indicators for the 11 hospitals that FHA has placed on
its priority watch list show much greater levels of risk when compared
with FHA's underwriting guidelines (see fig. 5). For these hospitals,
performance as measured by all four selected indicators declined from 2000
to 2004. Further, in 2004, three indicators showed a high level of risk,
based on FHA's criteria. For example, according to FHA's criteria, an
applicant with an operating margin of less than zero is considered high
risk. The hospitals on FHA's priority watch list had median operating
margin of -2.65 in 2004. Similarly, according to FHA's criteria, an
applicant with less than 15 days of cash on hand is also high risk, and
hospitals on the priority watch list had a median of 3.3 days of cash on
hand in 2004. FHA recognizes that the high risk levels of these selected
financial indicators are among the reasons that these hospitals are on its
priority watch list and are, therefore, subject to closer monitoring to
reduce the risk of a claim.

Figure 5: Selected Median Financial Indicators Show High Levels of Risk
for Priority Watch List Hospitals

Analysis of program data further shows that, while loans are increasingly
being insured outside of the Northeast, the program is still concentrated
in New York (see fig. 6). Though the percentage of the unpaid principal
balance concentrated in New York has decreased from 89 percent in 2000, 61
percent of the unpaid principal balance in the program remains
concentrated in New York in 2005. Of the 30 hospital loans that FHA has
insured since 2000, 21 are outside of New York, and 19 are outside of the
Northeast region. Since 2003, 5 of the loans insured were for CAHs.

Figure 6: FHA Loans Remain Concentrated in the Northeast as of December
2005: Active Loan Dollar Amount and Number of Loans by State

Further, 24 out of 25 mortgage insurance applications in development at
the time of our study are located outside of the Northeast (see fig. 7).26

Figure 7: Applications for Hospital Mortgage Insurance are Geographically
Dispersed as of December 2005: Application Dollar Amount and Number of
Applications by State

Despite these strides, the high concentration of the program's unpaid
principal balance in New York, as well as concentrations with single
borrowers with multiple loans, creates risks. New York hospitals insured
through FHA, like hospitals nationwide, rely heavily upon reimbursement
through Medicare and Medicaid.27  Since a portion of Medicaid funding
comes from states, any cuts made by the state of New York could have an
especially negative impact on the hospital program. Insured hospitals in
New York are also vulnerable to other state policies. For example, a task
force appointed by the Governor is in the process of identifying New York
hospitals for closure or restructuring. The Governor and state legislature
have committed state funds to assist in restructuring efforts, and the
state has had a history of helping its hospitals avoid defaults.
Nevertheless, any recommendations for the closure or restructuring of
FHA-insured hospitals may present the risk of an insurance claim. Further,
some New York hospitals have multiple loans insured through the program,
one with unpaid principal balances totaling approximately $828 million as
of December 2005. According to HUD's comments on the draft of this report,
this hospital is a financially sound, well-endowed institution that poses
a low risk of default.

Industry Trends Pose Risks to Hospitals, Including Those with FHA Insured
Mortgages

The hospital program may also face risks from changes in the health care
industry at large. According to industry literature, decreasing revenue
streams, increases in the number of uninsured patients, increased
competition from specialized facilities, and heightened capital needs are
some of the trends that affect all hospitals, including FHA-insured
hospitals.

We and others have reported that Medicare and Medicaid spending may not be
sustainable at current levels.28 If program cuts occur in Medicaid, for
example, states may take cost containment measures to reduce spending.29
Such measures may include frozen, or reduced, reimbursement rates to
providers and restrictions on eligibility for these programs.30  In
addition, the number of Medicare enrollees is projected to increase as
baby-boomers age and become Medicare-eligible.31  These trends will affect
all hospitals, including FHA-insured hospitals, which generally have
Medicare and Medicaid patients in their payer mix. On average, Medicare
discharges for FHA-insured hospitals represented 29 percent of total
discharges per hospital, and Medicaid discharges represented 19 percent of
total discharges per FHA-insured hospital.32 (See fig. 8 for Medicare and
Medicaid discharges by state.) Stated another way, nearly 50 percent of
the reimbursement that program hospitals receive is through Medicare and
Medicaid.

Figure 8: FHA-Insured Hospitals Have Medicare and Medicaid Payers among
Their Patient Discharges

Note: More than 90 percent of these data have undergone basic edit checks;
however, Centers for Medicare & Medicaid Services (CMS) has not yet
determined whether these data require an audit. These data may change as
they undergo further review by CMS. In addition, CMS does not enforce
dates by which hospitals must report data. Thus, only 49 of the 59 FHA
hospitals active as of October 2005 had provided data for 2003.

Hospitals, including FHA-insured hospitals, must also contend with the
rising number of underinsured and uninsured patients, which place demands
on hospitals to provide care with little to no reimbursement. According to
the U.S. Census Bureau, the number of uninsured persons rose from under 40
million people in 2000 to approximately 45 million people in 2003. This
trend may pose a risk to the program. In addition, hospitals in New York,
where the hospital mortgage insurance is concentrated, serve a high
proportion of uninsured patients.

Credit rating agencies state that competition is increasing in the health
care market as the type of care provided shifts to outpatient and
specialty hospitals, which provide profitable services, such as
cardiology, surgery, orthopedics, and diagnostic imaging.33 Specialty
facilities providing these services can take patients and revenue from
general acute-care hospitals, which supplement revenue shortfalls with
profitable services after providing needed, but unprofitable, services to
the community. The growth of specialty hospitals, such as ambulatory
surgery centers, is strong. The average number of specialty hospital
openings has increased from 5 hospital openings in the 1960s to 27
hospital openings in the 2000-present time period.34

Hospitals throughout the health care sector face increasing capital
demands, yet many have limited access to capital according to hospital
industry literature.35  For example, hospitals face demand for outpatient
services, emergency room upgrades, and technological advancements, which
have significant up-front and maintenance costs. A reputable credit rating
agency estimates that information technology expenditures now range
between 20 to 30 percent of a hospital's capital budget. Financially
weaker hospitals have less access to capital, yet often have pent-up
capital needs. According to a recent rating agency report, New York
hospitals have unmet capital needs as a result of their older
infrastructure and because they are generally financially weaker than the
average hospital.36

FHA Uses Tools to Mitigate Risk

FHA uses a variety of tools to mitigate risk in the hospital program. For
example, during its preliminary assessment of a hospital, FHA reviews the
hospital's ability to pay its mortgage by analyzing its debt service
coverage ratio and determines if this ratio meets FHA's minimum
requirement.37 FHA takes other steps when reviewing applications (as
discussed previously) designed to keep out excessively risky projects and
also imposes requirements on insured hospitals to control risks. These
include

o assessing the viability of projects at preapplication meetings with key
hospital representatives;

o using a comprehensive underwriting process that assesses, among other
factors, past and projected financial performance and the demand for the
hospital's services;

o hiring an independent consultant to evaluate the feasibility of the
proposed project and its potential risk to the FHA;

o requiring insured hospitals to establish a cash reserve fund sufficient
to cover 2 years of mortgage payments;38

o requiring insured hospitals to maintain compliance with key agreements
between the hospital and FHA and monitoring these agreements;39 and

o considering insured hospitals that fail to meet certain financial
criteria for placement on the priority watch list.40

FHA has also made some efforts to address the risks associated with the
geographic concentration of the program in New York. Since 1999, FHA has
had goals for geographically diversifying the hospital portfolio.
Currently, FHA's goals for diversifying the portfolio include reviewing
and processing applications for projects in states other than New York.
While the agency does not have a formal strategy for marketing the program
outside of New York, it has made some efforts to diversify the hospital
portfolio by

o simplifying its application process for CAHs and providing rural
hospital associations with information about the program;

o hiring an expert in rural hospitals;

o visiting hospital association conferences to educate members about the
program; and

o educating HUD field attorneys, mortgage bankers, and consultants about
the program.

HUD has also cooperated with requests for program information from the
trade media and assisted other researchers, which resulted in the
publication of articles and reports that provided information about the
advantages of the hospital program in financing capital projects. A formal
strategy, however, would provide the agency with a tool for
comprehensively planning for and executing activities that would lead to
the geographic diversification of the hospital portfolio. OMB guidance,
for example, requires that agencies include a description of the means and
strategies that will be used to achieve goals in their strategic plans.
Such strategies could include, for example, the processes, skills,
technologies, and various resources that will be used to achieve goals.

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

HUD uses a model for estimating annual credit subsidies that does not
explicitly consider the impacts of some potentially important factors.
HUD's model incorporates factors and assumptions about how loans will
perform, including estimated claim and recovery rates, which are
consistent with OMB guidance. HUD has generally calculated a negative
subsidy rate for the hospital program, meaning that estimated cash inflows
have been greater than estimated cash outflows. However, HUD's model does
not explicitly consider the potential impacts of prepayment penalties or
restrictions when estimating prepayments, or the debt-service coverage
ratios of hospitals at the time of loan origination.

For budgeting purposes, agencies that make loans and provide loan
guarantees must estimate the costs to the government over the life of the
loans that will be insured, commonly referred to as the subsidy cost. In
order to estimate the subsidy cost of the Hospital Mortgage Insurance
Program, HUD uses a cash-flow model to project expected net cash flows for
all these loans over their entire life. HUD's model is a computer-based
spreadsheet that uses assumptions based upon historical and projected data
to estimate the amount and timing of claims, subsequent recoveries from
these claims, as well as premiums and fees paid by the borrower.  In
addition, HUD does not consider prepayment penalties and restrictions when
it estimates the level and timing of prepayments, which affect estimates
of future claims and premiums.

HUD inputs its estimated cash flows into the OMB's credit subsidy
calculator, which 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
cash inflows are estimated to exceed cash outflows. For the hospital
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
and notes held in inventory.

Since the hospital program's inception, FHA has paid a total of 22
hospital mortgage insurance claims. The last claim was filed in 1999.
Because of the small number of claims, HUD determined that claim rates
based solely upon the program's historical claims experience would not be
reliable.41 As a result, HUD uses a methodology initially developed by OMB
to increase its estimated claim rate by assuming that the lenders for some
active hospitals would file claims for insurance. HUD refers to this
methodology as an artificial default.42 In determining which loans to
artificially default, HUD focuses on hospitals that generally have a
higher risk of default, and are therefore on FHA's priority watch list.43
According to OMB officials, the use of this artificial default accounts
for the risk that exists due to the low number of large size loans
insured, potential changes in Medicare or Medicaid reimbursement rates,
and the geographic concentration of the program in New York, which make
the program vulnerable to regional economic conditions.

In 10 of the 14 years that HUD has been estimating the cost of the
Hospital Mortgage Insurance Program under credit reform, HUD has estimated
that the present value of cash inflows from fees, premiums, and recoveries
from loans and properties sold would exceed the outflows from claim
payments and other expenses related to properties and notes held in
inventory. As a result, HUD calculated a negative credit subsidy rate for
the hospital program for these 10 years. In the other 4 years, HUD
estimated positive or no credit subsidy costs for the program. Figure 9
shows changes in the credit subsidy rate from 1992 to 2005.

Figure 9: Credit Subsidy Rates for the Hospital Mortgage Insurance Program
Have Generally Not Indicated a Need for Subsidies

While HUD's model includes assumptions that are consistent with OMB
guidance, such as assumptions on estimated claim and recovery rates and an
artificial default methodology to supplement the claim experience, 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 collection of premiums. Inclusion of initial
debt-service coverage ratios, as a factor predictive of defaults and claim
rates into HUD's cash-flow model for the hospital program, could
potentially enhance HUD's estimate of the subsidy cost of the program.
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.44 In turn,
prepayments affect claims because if a loan is prepaid it can no longer go
to claim. According to FHA officials, FHA does not place prepayment
penalties on FHA-insured hospital loans. However, according to the
hospital program's regulations, a mortgage loan made by a lender that has
obtained the funds for the loan through bonds can impose a prepayment
penalty charge and place a prepayment restriction on the mortgage's term,
amount, and conditions.45

According to FHA officials and mortgage bankers, prepayment restrictions
on hospital loans are generally in the form of 10-year restrictions on the
prepayment of bonds. While FHA does not maintain data specifically on
insured hospitals' bond-financing terms, prepayment restrictions are
specified on the mortgage note, which is available to FHA. Moreover,
according to the Mortgage Insurance for Hospitals Handbook, FHA has access
to bond-financing terms because, upon completion of bond issues,
applicants are required to submit bond-related documents to FHA so that
FHA can verify that the fees, charges, and other costs previously approved
with respect to debt restructuring. Incorporation of such data into the
hospital 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. However, by not explicitly incorporating prepayment
penalties or restrictions into the cash-flow model, HUD's model is less
able to estimate the impact of changes in prepayment patterns of current
and future  cohorts.

HUD's cash-flow model also does not consider the initial debt-service
coverage ratio of hospital 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.
(HUD's cash-flow model does consider the current debt-service coverage
ratio of insured hospitals through its artificial default methodology,
which, as previously explained, includes hospitals that are on FHA's
priority watch list. This list may include insured hospitals if, based
upon the last available full year of data, their debt-service coverage
ratio is below 1.10.)

According to the HUD official responsible for HUD's cash-flow model, the
initial debt-service coverage ratio of a hospital at the point of loan
origination is not included as a part of the cash-flow model for the
hospital program 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, 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.  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. They can also be used to assess the financial health of
either an applicant or a hospital in the existing portfolio.

According to officials from FHA's Office of Insured Health Care
Facilities, the projected debt-service coverage ratio is most meaningful
for the third or fourth year projected, when construction is most likely
to be complete. Our analysis of projected debt-service coverage ratios,
which include the amount of new debt being insured, shows that these
ratios varied from 1.48 to 3.11 during the fourth year projected.46 All
other factors being equal, loans with a debt-service coverage ratio of
3.11 are generally considered to have less risk than a loan with only a
1.48 debt-service coverage ratio.

Finally, we also found that economic studies show mixed results regarding
the significance of the impact of debt-service coverage ratios upon
commercial mortgage defaults. Some studies find initial debt-service
coverage ratios to be statistically insignificant in modeling commercial
mortgage defaults.47 Other studies indicate that initial debt-service
coverage ratios are meaningful factors in modeling default risk and are
helpful in predicting commercial mortgage terminations.48 Analysis of
initial debt-service coverage ratio information, which is available in
underwriting documents, may be used to identify trends or shifts in the
overall risk of the portfolios that should be considered when making
credit subsidy estimates. Further, current credit reform guidance calls
for agencies to use the best available data when preparing their credit
subsidy estimates.

Conclusions

The Hospital Mortgage Insurance Program plays an important role by
insuring loans for capital improvements at hospitals that, due to their
greater financial risks, would otherwise face difficulty in accessing
capital. FHA's process for reviewing applications for mortgage insurance,
while somewhat lengthier and involving more steps compared with those of
private bond insurers, appears to be a reasonable response to the
generally riskier nature of the applicants. Further, the agency's
techniques for monitoring insured hospitals are quite similar to those
used by private insurers, and the program has operated for several years
without experiencing an insurance claim.

FHA and HHS appear to work together reasonably well in carrying out their
respective roles in administering the program. However, it is difficult
for us, FHA's managers, or the Congress to assess how well the agencies
perform in implementing the program because FHA has not established a set
of meaningful program performance measures or collected the information
needed to assess performance. We have previously reported on the
importance of agencies' collecting useful performance information. If FHA
collected useful performance information, such as information based on
measurable and objective performance measures, the agency's managers could
use it to identify problems, try to identify the causes of problems,
and/or to develop corrective actions. Many program activities, including
those delegated to HHS, are recorded in FHA's Hospital Mortgage Insurance
Management Information System, and data from this system could be used to
establish and monitor useful performance measures. In addition, because
FHA has not updated the program handbook since 1984, hospitals, lenders,
investment bankers, health care financing agencies, and other interested
parties do not have ready access to a consolidated source of current
program eligibility requirements, policies, and procedures. The lack of a
consolidated source of current information may cause confusion and delay
hospitals' ability to prepare applications that meet FHA's criteria.
Further, outdated guidance in federal programs is an internal control
weakness.

Although it represents a relatively small part of HUD's GI/SRI fund, the
hospital program insures multimillion dollar loans that currently total
nearly $5 billion. The continued geographic concentration of insured
hospitals in the state of New York poses a source of financial risk to the
program. While this concentration has decreased from its high of 89
percent of outstanding insured principal balance in 2000, the current 61
percent represents a continuing concentration of credit risk. As a result,
the program is vulnerable to New York State policies, such as the
governor's call to restructure hospitals, as well as regional economic
trends. While FHA has taken steps in the right direction, it does not have
a formal strategy or plan for geographically diversifying the hospital
portfolio, which could enhance current efforts to reach this goal.

HUD's cash-flow model used to estimate annual credit subsidy rates appears
to be consistent with applicable OMB guidance; however, it does not
explicitly take into account potentially useful factors such as prepayment
penalties and restrictions or the initial debt-service coverage ratio of
new loan cohorts. Although the program has not experienced a claim for
insurance since 1999, the increasing size of loans insured, geographic
concentration in New York and the Northeast, and other factors pose risks
to the program. Including additional factors into HUD's model could
potentially enhance the agency's estimates of the subsidy cost of the
program, provide HUD and congressional decision makers with better cost
data to assess the program, and help assure that the program adequately
addresses financial risks.

Recommendations for Executive Action

To improve management of the Hospital Mortgage Insurance Program and
reduce potential risks to the GI/SRI fund, we recommend that the Secretary
of Housing and Urban Development direct the FHA Commissioner to take the
following three actions:

o Establish measurable and objective performance measures for the hospital
program and collect appropriate information to regularly assess
performance against the measures.

o Update the program handbook to make publicly available current
eligibility requirements, policies, and procedures.

o Develop a formal strategy to geographically diversify its portfolio of
insured hospitals, including such elements as the processes, skills,
technologies, and various resources that will be used to reach
diversification goals.

To potentially improve HUD's estimates of the program's annual credit
subsidy rate, we recommend that the Secretary of Housing and Urban
Development explore the value of explicitly factoring additional
information, such as prepayment penalties and restrictions, as well as the
initial debt-service coverage ratio of hospitals, as they enter the
program into its credit subsidy model.

Agency Comments and Our Evaluation

We provided a draft of this report to HUD and HHS for their review and
comment. In written comments from HUD's Assistant Secretary for
Housing-Federal Housing Commissioner, which incorporated comments from
HHS, HUD concurred with our four recommendations. However, the agency
disagreed with our presentation of certain aspects of the program,
commenting that the report's "critique of procedural and technical
matters" overshadowed the program's accomplishments. The Assistant
Secretary's letter appears in appendix IV, and a letter from HHS appears
in appendix V.

HUD expressed general agreement with the recommendations and noted actions
that it plans to take. Specifically, the agency agreed

o to develop appropriate performance measures and implement data
collection procedures to evaluate both program and contract
administration;

o with the need to consolidate updated eligibility requirements, policies,
and procedures into an updated handbook, and stated its intention to have
the handbook finalized by the end of 2006;

o to develop a formal strategy to geographically diversify its portfolio
of insured hospitals, including such elements as the processes, skills,
technologies, and various resources that will be used to reach
diversification goals; and

o to explore the value of explicitly factoring additional information,
such as prepayment penalties and the initial debt-service coverage ratio
of hospitals as they enter the program, during its annual review of cash
flow modeling techniques for the hospital program.

In disagreeing with our presentation of FHA's efforts to diversify the
hospital portfolio, HUD commented that diversification has been a top
program goal for many years. Our draft report acknowledged that FHA has
had goals for geographically diversifying the portfolio since 1999 and
provided examples of FHA's diversification efforts. However, in response
to the comments, we included additional examples of FHA's efforts. HUD
also commented that the report does not appropriately emphasize the
success that HUD and HHS have had in working together to implement the
hospital program. Our draft report acknowledged the agencies coordinated
involvement with key meetings, underwriting, and monitoring. Further, as
the letter from HHS observes, our draft report concluded that the two
agencies appear to be working reasonably well together. Because we believe
the report accurately characterizes the relationship, we did not change
it. Finally, HUD commented that the report infers that (1) it has not
maintained current policies and procedures and (2) indicates that current
eligibility requirements, policies, and procedures are unavailable to the
public. Our draft report stated that the handbook does not contain current
eligibility requirements, policies, or processing procedures, and
acknowledged that FHA publicly communicates program changes through
Mortgagee Letters. Nevertheless, in response to HUD's comments, we revised
the report to include additional examples of FHA's efforts to communicate
changes in eligibility requirements, policies, and procedures. We also
continue to emphasize the value of updating all program documentation,
including the handbook.

HUD also offered comments regarding the report's presentation of risks
facing the hospital program, including potential cuts in reimbursement
from Medicare and Medicaid, the potential for closures of hospitals in New
York stemming from a commission appointed by the Governor, and the large
size of some loans. We recognize that potential cuts in reimbursement from
the Medicare and Medicaid programs are a risk factor for hospitals in all
states; however, New York is unique among states in accounting for over
half of the hospital program's insurance portfolio. We revised the report
to clarify that, due to this concentration, any cuts that the state of New
York makes to its Medicaid program could have an especially negative
impact. Regarding the New York Governor's commission, we are aware that
state funds are available to assist in restructuring efforts, and that the
Dormitory Authority of the State of New York is committed helping its
hospitals avoid defaults. However, since there is no guarantee that
FHA-insured hospitals will be protected, we continue to believe that a
recommendation for their closure or restructuring may present the risk of
an insurance claim. Finally, we revised the report as HUD suggested to
note that the largest single exposure of $828 million is for a hospital
that, according to HUD, poses a low risk of default.

HUD commented that GAO's presentation of processing times for applications
is misleading because it does not mention that there can be periods of
time in which HUD cannot continue to process applications due to factors
that applicants must address and are thus beyond HUD's control. Because
HUD's system for tracking application processing times does not capture
such periods of time, it is not possible for GAO to quantify their impact.
Further, the report notes that processing times vary with the complexity
of the project and may be affected by issues outside of HUD's control.

HUD took exception with our conclusion that it is difficult for us, FHA's
managers, or the Congress to assess how well the agencies perform in
implementing the program because FHA has not established a set of
meaningful performance measures and stated a belief that program results
indicate that the program is fulfilling its purpose. While our report
acknowledges that the program has had a good performance history, the
creation and use of performance measures can be used by agency managers to
improve a program's results. As we note in the report, analysis of
performance information helps managers identify problems, identify the
causes of problems, and develop corrective actions. In addition,
performance information can be used to develop strategies, identify
priorities, make resource allocation decisions, and identify more
effective approaches to program implementation.

HUD disagreed with our suggestion that it include such factors as initial
debt-service coverage ratio into its credit subsidy modeling and noted
that two of the studies that we cited found this ratio to be statistically
insignificant in predicting commercial mortgage defaults. Our draft report
in fact stated that economic studies have shown mixed results regarding
the significance of the impact of debt-service coverage ratios on
commercial mortgage defaults. However, we revised the report to explicitly
footnote studies that show initial debt-service coverage ratios to be
statistically insignificant and those that indicate that this ratio is a
meaningful factor in modeling default risk. We also note that the two
studies that found initial debt-service coverage ratios to be
statistically insignificant were both based on the same, small data set.
We acknowledge that HUD's cash-flow model considers the current
debt-service coverage ratio of insured hospitals through its artificial
default methodology. However, our recommendation is to include the
debt-service coverage ratios at origination, so that the risk of loans at
origination will be reflected in the credit subsidy rates for the cohort.

Finally, our draft report stated that FHA estimates that hospital loans
are most likely to experience a claim during their tenth insured year. In
its comment, HUD stated that historically 70 percent of claims occurred
prior to a loan's tenth year. The statement in our draft report was based
on actual historical conditional claim rate data. HUD subsequently
provided additional information, which explained that the conditional
claim rate peaked due to a single claim with multiple notes. As a result,
we revised the report to reflect additional information.

We are sending copies of this report to the Secretaries of the Departments
of Housing and Urban Development (HUD) and Health and Human Services
(HHS). 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 the report are listed in appendix VI.

David G. Wood, Director Financial Markets and Community Investment

Objectives, Scope, and Methodology Appendix I

Our objectives were to review (1) the design and management of the
program, as compared with private insurance; (2) the nature and management
of the relationship between the Department of Housing and Urban
Development (HUD) and the Department of Health and Human Services (HHS) in
implementing the program; (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.

To review the design and management of the Hospital Mortgage Insurance
Program we interviewed officials at both the Federal Housing
Administration's (FHA) Office of Insured Health Care Facilities and the
Division of Facilities and Loans within HHS' Health Resources and Services
Administration, and reviewed program policies, documentation of
application processes, laws, and regulations. To compare the program's
design with that of private insurers, we met with private bond insurers
and the Association of Financial Guaranty Insurers, credit rating
agencies, mortgage and investment banking firms, hospital associations,
and state health care financing agencies in New York and New Jersey.1

To describe how FHA and HHS coordinate the implementation of the hospital
program, we interviewed FHA and HHS officials about the responsibilities
for each agency in implementing the program. We also reviewed the
Memorandum of Agreement between FHA and HHS that describes the division of
duties and responsibilities between the two agencies and organizational
charts that depict FHA's organization, HHS's organization, and the FHA-HHS
interrelationship in program administration. We analyzed the extent to
which performance measures related to interagency coordination were met by
obtaining available data from FHA and analyzing time frames for processing
applications and loan modification requests from the Hospital Mortgage
Insurance Management Information System (HMIMIS). We compared performance
measures with our criteria on performance measures and compared
performance measures in the 2002-2005 Memorandum of Agreement with the
performance measures in the 2006-2010 Interagency Agreement between FHA
and HHS to identify any changes.

To identify the financial implications of the program to the GI/SRI fund,
we interviewed and obtained documentation from FHA and HHS program
officials and analyzed FHA data on program portfolio characteristics,
including number and amount of loans by cohort, current
insurance-in-force, and geographic concentration of loans, claims, and
recoveries. Specifically,

o To obtain the number and amount of active and terminated loans, we
created a report from the HMIMIS database, which is updated monthly. To
assess the reliability of the HMIMIS data, we reviewed relevant
documentation, interviewed agency officials who worked with this database,
and conducted electronic testing of the data, including frequency and
distribution analyses. We determined the data to be sufficiently reliable
to obtain the number and amount of active loans. We corroborated these
data with the FHA's 2004 report to the Congress.2 As of December 2005, the
administrators provided data from HUD's F-47 database, a multifamily
database, to show (1) that there were 59 active hospitals with 74 active
loans in the Hospital Mortgage Insurance portfolio and (2) that there had
been 341 loans in the portfolio since the inception of the program. To
assess the reliability of data from HUD's F-47 database, we reviewed HUD's
Hospital Mortgage Insurance Program Functional Requirements Document,
Procedures for Maintaining Group Records, and other relevant
documentation, interviewed agency officials who worked with this database,
and conducted electronic testing of the data, including frequency and
distribution analyses. Our assessment showed that two loan records were
lacking state data, and one record was lacking hospital name data but were
identified by a unique project number. FHA administrators verified that
these loans were endorsed long before electronic loan records were
maintained and that they were unable to provide additional information.
None of our analyses utilized the missing data elements for the two
projects; therefore, there was no impact on this report. We determined the
data to be sufficiently reliable to describe the geographic concentration
of loans in the program.

o To determine the proportion of the Hospital Mortgage Insurance Program
to the larger GI/SRI fund, we reviewed a spreadsheet provided by HUD's
Office of Evaluation dated June 2005 on insurance-in-force for the
hospital program to that of the GI/SRI fund.

o To determine the risk posed by insurance claims to the Hospital Mortgage
Insurance Program, we analyzed spreadsheets with historic claims and
recoveries data provided by HUD's Office of Evaluation and dated August
2005.

o To determine the geographic concentration of loans and loan prepayment
history in the program, we analyzed data current as of December 31, 2005,
in an extract of HUD's F-47 database. While we obtained extracts from
HUD's F-47 database in April 2005, October 2005, and December 2005, all
analyses from F-47 data in the report utilize the December 2005 extract
only.

We also compared data on four financial ratios including debt service
coverage, days cash on hand, current, and operating margin ratios provided
from HMIMIS, current as of December 2005, with applicant criteria stated
in the Manual of the Hospital Insurance Program.

To determine how FHA manages program risks, we interviewed FHA and HHS
program officials and reviewed the Mortgage Insurance for Hospitals
Handbook and manual to determine steps taken by the agency during the
application and monitoring phases of the insurance process. We analyzed
cash inflows to the program from income from notes held in inventory,
rental income from properties held in inventory, sales income from notes,
and properties sold from inventory. We also reviewed documentation of cash
outflows, such as claim payments and expenses related to properties and
notes held in inventory. To assess risk based on geographic concentration,
we identified the state with the highest unpaid principal balance insured
by the program. Finally, to assess risk posed by the geographic
concentration of the program's unpaid principal balance, we extracted 2003
data from the Centers for Medicare & Medicaid Services (CMS) database on
Hospital Mortgage Insurance Program hospitals active in 2005. We used CMS
data to identify the number of discharged patients whose services were
paid for through the Medicare/Medicaid programs from hospitals that have
loans insured through the program. More than 90 percent of these data have
undergone basic edit checks; however, CMS has not yet determined whether
these data require an audit. These data may change as they undergo further
review by CMS. In addition, CMS does not enforce dates by which hospitals
must report data. Thus, at the time of this report, only 49 of the 59
active FHA hospitals had provided data for 2003.

We conducted a literature review and interviewed numerous officials of
rating agencies and hospital associations to obtain information on risks
due to health care market trends. We conducted the following academic
literature searches: (1) Google's Scholar search engine using the terms
"hospital mortgage insurance," "nursing home mortgage insurance,"
"hospital and default and FHA," "nursing home and default and FHA"; (2)
PubMed Web site using the terms "hospital mortgage insurance" and "nursing
home mortgage insurance;" and (3) HUDuser.org Web site using the terms
"hospital mortgage insurance," "nursing home mortgage insurance," and
"Section 242." We also searched for Inspectors General and agency reports
through HUD and HHS Web sites using the terms "Hospital mortgage
insurance" and "Section 242." Finally, we conducted a search on our
internal Web site to identify previous work on the Section 242 program.
The terms "hospital," "mortgage insurance," and "Section 242" were used
for the period of January 1995 through March 2005.

To determine how HUD estimates the annual credit subsidy rate for the
program, we interviewed program officials from HUD's Office of Evaluation
and program auditors from the Office of Management and Budget (OMB),
reviewed documentation of HUD's credit subsidy estimation procedures, and
reviewed the cash-flow model for the program. 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
FHA on program hospitals' projected debt-service coverage ratios (at the
time of their loan application). HUD's Budget Office provided the
program's annual credit subsidy rates for 1992 and 1993, and we obtained
this rate for years 1994-2005 from the Federal Credit Supplement of the
United States Budget.3

Our review did not include an evaluation of underwriting criteria,
construction monitoring, or the need for the program. We conducted our
work in Albany, New York; Chicago, Illinois; New York, New York; Paterson,
New Jersey; Rockville, Maryland; and Washington, D.C., between February
2005 and January 2006 in accordance with generally accepted government
auditing standards.

FHA and HHS' Responsibilities in FHA's Hospital Mortgage Insurance Program
Loan Cycle Appendix II

Source: FHA Hospital Mortgage Insurance Program staff.

aThe Account Executive can be an FHA or HHS staff member. The review team,
or Client Service Team, can consist of FHA and/or HHS staff. The PMG
consists of both senior FHA and HHS staff.

bFHA's Director of the Office of Insured Health Care Facilities.

FHA Assessed Performance Using 2 of 22 Performance Measures Included in
the 2002-2005 Memorandum of Agreement Appendix III

Source: HHS, Office of Special Programs, Memorandum of Agreement between
the Department of Housing and Urban Development (HUD) and the Department
of Health and Human Services (HHS), 2001.

Note: Some performance measures refer to tasks that can be done by either
FHA or HHS officials, which we refer to as "coordinated tasks." Other
performance measures apply to tasks specifically for HHS, which we refer
to as "HHS-delegated tasks."

Comments from the Department of Housing and Urban Development Appendix IV

Comments from the Department of Health and Human Services Appendix V

GAO Contact and Staff Acknowledgments Appendix VI

David G. Wood (202) 512-8678

Individuals making key contributions to this report included Alison
Martin, Lisa Moore, David Pittman, Minette Richardson, Paul Schmidt, and
Julie Trinder.

(250239)

www.gao.gov/cgi-bin/getrpt? GAO-06-316.

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-06-316, a report to congressional committees

February 2006

HOSPITAL MORTGAGE INSURANCE PROGRAM

Program and Risk Management Could Be Enhanced

Under its Hospital Mortgage Insurance Program, the Department of Housing
and Urban Development's (HUD) Federal Housing Administration (FHA) insures
nearly $5 billion in mortgage loans for the renovation or construction of
hospitals that would otherwise have difficulty accessing capital. In
response to a requirement in the 2005 Consolidated Appropriations
Conference Report, GAO examined (1) the design and management of the
program, as compared with private insurance, (2) the nature and management
of the relationship between HUD and the Department of Health and Human
Services (HHS) in implementing the program, (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.

What GAO Recommends

GAO recommends that the HUD Secretary ensure that program performance
measures are useful, update the program handbook, develop a formal
geographic diversification strategy, and explore adding factors to HUD's
credit subsidy model. HUD agreed with GAO's recommendations but said that
the report did not adequately emphasize the program's accomplishments.

The Hospital Mortgage Insurance Program insures the mortgages of hospitals
that are generally riskier than those that can obtain private bond
insurance. While FHA's process for reviewing mortgage insurance
applications includes more steps and generally takes longer, the agency
monitors active loans with many of the same techniques that private bond
insurers use.

Under a Memorandum of Agreement, FHA and HHS work together in a variety of
ways to review mortgage insurance applications and monitor active loans.
However, FHA does not collect data to assess program performance against
most performance measures specified in the memorandum, some of which are
not objective. Further, FHA has not kept its program handbook of policies
and procedures for applicants, lenders, and others up-to-date.

The hospital program is small compared with other programs in the GI/SRI
fund, and the losses from claims have been relatively low. Despite the
program's relatively small size, some program and market trends may pose
risks. For example, 61 percent of the program's total insured, outstanding
loan amount is concentrated in New York, which makes the program
vulnerable to state policies and regional economic conditions. While FHA
has goals to diversify the hospital insurance portfolio and has made
efforts to do so, it does not have a formal strategy to achieve these
goals.

To estimate the credit subsidy cost, or program costs, over the life of
the outstanding loans insured, HUD uses a model that incorporates factors
and assumptions about how loans will perform, including estimated claim
and recovery rates, which are consistent with federal guidance. However,
HUD's model does not explicitly consider some factors, such as the
potential impacts of prepayment penalties or restrictions, which according
to some economic studies, are important in modeling default risk.

FHA Hospitals Remain Concentrated in the Northeast as of December 2005:
Active Loan Dollar Amount and Number of Loans by State
*** End of document. ***