Welfare Reform: Effect on HUD's Housing Subsidies Is Difficult to
Estimate (Letter Report, 12/07/98, GAO/RCED-99-14).

Pursuant to a congressional request, GAO provided information on welfare
reform's financial impact on the Department of Housing and Urban
Development's (HUD) budget, focusing on what: (1) studies have been done
on welfare reform's financial impact on public and assisted housing; and
(2) methodological and data issues, if any, arise when researchers
estimate welfare reform's financial impact on low-income housing.

GAO noted that: (1) officials at housing agencies and researchers at
government agencies, universities, trade associations, and a consulting
firm have estimated welfare reform's financial impact on some components
of HUD's housing subsidy programs; (2) GAO identified 13 studies that
estimated this impact; (3) these studies of welfare reform's financial
impact on HUD's housing subsidy programs varied in their geographic
scope, focus and assumptions, methods, and findings; (4) some studies
also estimated welfare reform's impact under alternative scenarios and
therefore developed a range of estimates of welfare reform's cost for
HUD and housing agencies; (5) the estimates in the studies GAO reviewed
generally varied with the issues on which they focused and the
assumptions on which they were based; (6) some of the authors of the
studies GAO reviewed told it that their estimates might not hold up over
time because some federal and state welfare laws have changed since the
estimates were first developed and the economy has been more robust than
anticipated; (7) experts with whom GAO spoke generally agree that
several issues complicate efforts to forecast welfare reform's financial
impact on HUD's housing subsidy programs; (8) these issues include not
only those encountered in predicting welfare reform's impact on the
recipients and providers of public assistance, but also those specific
to estimating welfare reform's financial impact on the residents of
assisted housing, providers of subsidized housing, and HUD; (9) in
general, wide variations in state welfare plans and their implementation
complicate the estimation of welfare reform's impact; (10) the
employment and wage prospects for welfare recipients depend, in part, on
future local and national economic health and on recipients' behavior;
(11) housing experts generally agree that estimating welfare reform's
impact on housing programs is more complex than estimating welfare
reform's impact overall because of possible differences in the behavior
of welfare recipients with and without housing assistance, as well as
variations in policies adopted by housing agencies and landlords; and
(12) a lack of reliable data further hampers researchers' efforts to
predict welfare reform's financial impact on HUD's housing programs.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-99-14
     TITLE:  Welfare Reform: Effect on HUD's Housing Subsidies Is 
             Difficult to Estimate
      DATE:  12/07/98
   SUBJECT:  Federal aid for housing
             Housing programs
             Public housing
             Welfare benefits
             Low income housing
             Rent subsidies
             Public assistance programs
             Workfare
             Projections
IDENTIFIER:  HHS Temporary Assistance for Needy Families Program
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed 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.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Subcommittee on VA, HUD, and Independent Agencies,
Committee on Appropriations, House of Representatives

December 1998

WELFARE REFORM - EFFECT ON HUD'S
HOUSING SUBSIDIES IS DIFFICULT TO
ESTIMATE

GAO/RCED-99-14

Estimating Welfare Reform's Financial Impact

(385742)


Abbreviations
=============================================================== ABBREV

  AFDC - Aid to Families with Dependent Children
  CBO - Congressional Budget Office
  DMG - David M.  Griffiths & Associates, Ltd. 
  HHS - Department of Health and Human Services
  HUD - Department of Housing and Urban Development
  MDRC - Manpower Demonstration Research Corporation
  OMB - Office of Management and Budget
  TANF - Temporary Assistance for Needy Families

Letter
=============================================================== LETTER


280875

December 7, 1998

The Honorable Jerry Lewis
Chairman
The Honorable Louis Stokes
Ranking Minority Member
Subcommittee on VA, HUD, and
 Independent Agencies
Committee on Appropriations
House of Representatives

The Personal Responsibility and Work Opportunity Reconciliation Act
of 1996 replaced an entitlement program for poor families with a
program that establishes work requirements for recipients and limits
the time they can receive federal cash assistance.  Because about
one-fourth of the residents of public and assisted housing rely on
cash assistance for some or all of their income, welfare reform could
have a financial impact on the Department of Housing and Urban
Development (HUD), which subsidizes housing agencies' operations and
tenants' rents, as well as on the housing agencies and private
landlords that depend on HUD subsidies.  As a result of changing
welfare and housing policies and the complexity of these programs,
this financial impact has not yet been estimated with any certainty. 

Because of your concern about welfare reform's financial impact on
HUD's budget, you asked us to determine (1) what studies have been
done on welfare reform's financial impact on public and/or assisted
housing and (2) what methodological and data issues, if any, arise
when researchers estimate welfare reform's financial impact on
low-income housing. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

Officials at housing agencies and researchers at government agencies,
universities, trade associations, and a consulting firm have
estimated welfare reform's financial impact on some components of
HUD's housing subsidy programs.\1 We identified 13 studies that
estimated this impact.  These studies of welfare reform's financial
impact on HUD's housing subsidy programs varied in their geographic
scope, focus and assumptions, methods, and findings.  For example,
while several studies indicate that welfare reform would result in
modest changes in the amounts of the HUD subsidies needed, the
findings of one national study indicate that HUD would need to
increase its annual subsidies to housing agencies by almost 42
percent to offset expected decreases in public housing rents, and
another study's findings indicate that HUD could decrease its annual
subsidies to a particular housing agency by almost 20 percent when
welfare reform fully takes effect.\2 Some studies also estimated
welfare reform's impact under alternative scenarios and therefore
developed a range of estimates of welfare reform's cost for HUD and
housing agencies.  The estimates in the studies we reviewed generally
varied with the issues on which they focused and the assumptions on
which they were based.  Some of the authors of the studies we
reviewed told us that their estimates might not hold up over time
because some federal and state welfare laws have changed since the
estimates were first developed and the economy has been more robust
than anticipated.  In addition, the authors of some studies expressed
uncertainty about their findings because the studies rely heavily on
factors that are difficult to predict, such as the degree to which
welfare recipients will change their behavior when faced with new
welfare policies. 

Experts with whom we spoke generally agree that several issues
complicate efforts to forecast welfare reform's financial impact on
HUD's housing subsidy programs.  These issues include not only those
encountered in predicting welfare reform's impact on the recipients
and providers of public assistance, but also those specific to
estimating welfare reform's financial impact on the residents of
assisted housing, providers of subsidized housing, and HUD.  In
general, wide variations in state welfare plans and their
implementation complicate the estimation of welfare reform's impact. 
Additionally, the employment and wage prospects for welfare
recipients--key factors affecting their incomes--depend, in part, on
future local and national economic health and on recipients'
behavior, both of which are difficult to predict.  Housing experts
generally agree that estimating welfare reform's impact on housing
programs is more complex than estimating welfare reform's impact
overall because of possible differences in the behavior of welfare
recipients with and without housing assistance, as well as variations
in the policies adopted by housing agencies and landlords.  Finally,
a lack of consistent and reliable data further hampers researchers'
efforts to predict welfare reform's financial impact on HUD's housing
programs. 


--------------------
\1 Throughout this report, we use the term "studies" to describe the
analyses we found of the impact of welfare reform on HUD and housing
agencies.  While some of these analyses were formal studies, others
were informal estimates. 

\2 The results of the studies we reviewed were typically presented in
dollar amounts--that is, how much less (or more) rental revenue would
be collected or how much more (or less) in subsidies HUD would need
to provide in thousands or millions of dollars.  Because some studies
have a national focus while others examine the impact of welfare
reform on housing agencies of various sizes, dollar estimates were
not readily comparable across studies.  To provide more comparability
for the purpose of this review, we converted all dollar estimates to
percentages reflecting the change that would occur in HUD subsidies
compared with the total relevant program subsidies in 1997.  Our
methodology for these conversions is discussed later in this report. 


   BACKGROUND
------------------------------------------------------------ Letter :2

The Personal Responsibility and Work Opportunity Reconciliation Act
of 1996 (P.L.  104-193) replaced the entitlement program, Aid to
Families with Dependent Children (AFDC), with Temporary Assistance
for Needy Families (TANF).  TANF provides $16.5 billion annually to
the states in the form of block grants through 2002.  Under TANF,
recipients are required to work and can receive federal cash
assistance for only a limited period of time.  TANF's requirements
vary from state to state because the 1996 act gave the states more
control over the design of their own programs.  While TANF is
generally administered at the state level, the Department of Health
and Human Services (HHS) is the primary federal agency providing
oversight of states' welfare programs. 

Through its public housing and Section 8 programs, HUD provides
housing assistance to about 4.3 million low-income households.  In
fiscal year 1997, HUD's outlays for Section 8 subsidies and for
public housing modernization, development, and operating subsidies
amounted to about $22.6 billion.  About $7.5 billion of this amount
was allocated through the tenant-based Section 8 certificate and
voucher programs, under which housing agencies provide rent subsidies
to private landlords.  About $7.9 billion went directly to private
landlords as part of the project-based Section 8 program, and about
$7.2 billion went for the modernization, development, and operation
of Public and Indian Housing.  Included in this latter amount is $2.6
billion in appropriations that was distributed through HUD's
formula-based performance funding system to state, county, and local
housing agencies for the operation of public housing. 

In 1996, approximately a quarter of the households receiving HUD
subsidies also received cash assistance.\3 In general, families
receiving housing assistance are required to pay 30 percent of their
cash income (adjusted for certain items, such as child care and
medical expenses) in rent, while HUD provides subsidies to housing
agencies and private landlords to make up the difference between
tenants' rental payments and the cost of operating public housing
units or the rents charged by the landlords.  Because rental payments
are linked to household income, rental revenues will fall if families
receiving assistance are unable to replace lost welfare benefits with
wage income, and additional HUD subsidies will then be needed.  But
if assisted families' employment and earnings increase under welfare
reform, then the amount of the required rental payments may rise,
reducing the need for subsidies.  For subsidy needs to decline,
residents would have to earn more than they formerly received in cash
assistance, and working residents would need to either remain in
public and assisted housing after gaining employment or be replaced
with employed residents.  These conditions are less likely to hold in
areas where cash benefit levels are high and nonsubsidized housing is
available and affordable. 

HUD has established policies that can influence the impact of welfare
reform on housing programs.  For example, housing agencies and HUD
can set minimum rents of up to $50 for tenants who live in public
housing or have Section 8 certificates or vouchers, while owners of
project-based Section 8 properties are required to charge minimum
rents of $25.\4 In addition, recent legislation has expanded housing
agencies' authority to exclude some wage earnings from rental payment
calculations in an effort to retain working families in public
housing units.  Similarly, while housing agencies and subsidized
private landlords were formerly required to give preference in
admission to very poor families, they are now allowed to give some
preference to working families with wage income.  In addition to
these rent and admission policies, HUD provides programs, some of
which originated in the mid-1980s, to deliver employment-related
services to the tenants of public and assisted housing.  These
programs have provided job training, counseling, and placement
services; child care; and transportation. 


--------------------
\3 In addition, about half of the families with children residing in
public and assisted housing also received cash assistance in 1996. 

\4 Until recently, housing agencies were encouraged to grant
exceptions to the minimum rent requirements for public housing
residents and recipients of certificates and vouchers because of
hardship.  However, the Quality Housing and Work Responsibility Act
of 1998 (P.L.  105-276), enacted in Oct.  1998, now requires housing
agencies to grant hardship exceptions for, among other things, the
loss of federal cash assistance.  Exceptions to the project-based
minimum rents are provided for the elderly, the handicapped or
disabled, and working families with an adjusted monthly income below
$75.  These exceptions to the minimum rent requirements limit the
ability of housing authorities to use minimum rents to offset the
impact of welfare reform. 


   STUDIES OF WELFARE REFORM'S
   FINANCIAL IMPACT ON HUD'S
   HOUSING SUBSIDY PROGRAMS VARIED
   WIDELY
------------------------------------------------------------ Letter :3

We identified five studies estimating welfare reform's financial
impact on housing programs nationally, one estimating the impact for
eight housing agencies, and another seven estimating the impact for a
single housing agency.  While some of the studies suggest that
welfare reform will likely cause only modest changes in the amounts
of the HUD subsidies needed, some of the studies indicate more
substantial effects.  For example, one national study indicated that
HUD would need to increase its annual subsidies to housing agencies
by almost 42 percent to offset expected decreases in public housing
rents, while another study indicated that HUD could decrease its
annual subsidies to a particular housing agency by almost 20 percent. 
Differences in the studies' focus and assumptions help explain the
widely varying estimates.  While some researchers focused on a single
feature of a state's welfare reform plan, others examined the
national impact of a broad range of state plans; while some studies
used "worst-case" assumptions about the employment and earnings
prospects of welfare recipients, others used "best-guess"
assumptions.  Moreover, because certain welfare and housing policies
have changed since the estimates were developed, the economy has been
stronger than anticipated, and the effects of welfare reform on
welfare recipients' behavior are difficult to predict, some of the
authors of the studies we reviewed expressed uncertainty about their
estimates. 


      ESTIMATES VARIED WIDELY
      ACROSS AND WITHIN STUDIES
---------------------------------------------------------- Letter :3.1

Five studies we identified estimated the financial impact of welfare
reform nationally (see table 1).  Three of these five studies suggest
that welfare reform will have a relatively modest effect on the need
for HUD subsidies, ranging from a 0.4-percent annual decrease to a
3.3-percent increase in the amount needed.  The two remaining studies
anticipate a greater effect.  For example, the Council of Large
Public Housing Authorities indicated, in the fall of 1996, that the
annual amount needed for HUD subsidies could increase by 19 percent. 



                                     Table 1
                     
                      National Estimates of Welfare Reform's
                        Financial Impact on HUD's Housing
                                 Subsidy Programs

                                                            Estimate of
Estimate by/                                                percentage change in
for             Scope                 Scenario              HUD subsidy
--------------  --------------------  --------------------  --------------------
HUD             Public housing and    Impact of welfare     0.7% increase in
                Section 8 programs    reform, best-guess    1997
                                      scenario              3.3% increase in
                                                            2002

The Johns       All assisted housing  Impact of welfare     Results ranged from
Hopkins         programs\a            reform on AFDC        a
University                            recipients under      2.4% increase to a
                                      alternative welfare   0.4% decrease
                                      plans

Congressional   Section 8             Impact of welfare     0.1% increase in
Budget Office   programs              reform, best-guess    1998
                                      scenario              0.7% increase in
                                                            2008

Council of      Public housing        Impact of time        19% increase
Large Public                          limits, worst-case
Housing                               scenario
Authorities

David M.        Public housing        Impact of welfare     Results ranged from
Griffiths &                           reform with and       a 41.8% increase to
Associates,                           without proactive     a
Ltd.\b/Public                         housing agency        2.8% increase
Housing                               management\c
Authorities
Directors
Association
--------------------------------------------------------------------------------
\a As reported in the AFDC-Quality Control database. 

\b Since this study was completed, David M.  Griffiths & Associates,
Ltd., was purchased by Maximus, Inc., to become DMG-Maximus, Inc. 

\c The author of this study defined proactive housing agency
management as taking an active role in mitigating the impact of
welfare reform on the housing agency's residents and the housing
agency. 

Source:  GAO's analysis of the national studies listed in app.  I. 

Of the other eight studies we reviewed, seven estimated welfare
reform's impact on individual housing agencies in different parts of
the country, and one, by HUD's Office of Policy Development and
Research, covered eight individual housing agencies.\5 The estimates
for these eight studies, which are summarized in table 2, varied
widely, both from one housing agency to another and from one scenario
to another for a single housing agency.  In particular, under
assumptions that the authors characterize as unlikely--that the state
would adopt a harsh welfare reform plan and the housing authority
would not provide employment assistance to affected residents--the
Seattle housing authority's findings indicate that the agency could
need an annual increase of as much as 37 percent of its fiscal year
1997 HUD subsidy to offset welfare reform's impact.  Conversely,
using optimistic assumptions, HUD predicted that rental revenues at
the Dallas housing authority could rise by enough to warrant as much
as a 20-percent reduction in the amount of the HUD subsidy needed. 



                                     Table 2
                     
                        State and Local Housing Agencies'
                     Estimates of Welfare Reform's Financial
                     Impact on HUD's Housing Subsidy Programs

                                                            Estimate of
Estimate by/                                                percentage change in
for             Scope                 Scenario              HUD subsidy
--------------  --------------------  --------------------  --------------------
Los Angeles     Public housing and    Impact of welfare     Results ranged from
                Section 8             reform, different     a 0.8% increase to a
                certificates and      labor market          0.9% decrease
                vouchers              scenarios with and
                                      without exclusions
                                      of earned income
                                      from rental payment
                                      calculations

Seattle         Public housing        Impact of welfare     Results ranged from
                                      reform with focus on  a 36.5% increase to
                                      reductions in         a 1.9% increase
                                      benefits for
                                      unqualified legal
                                      immigrants\a

Miami-Dade      Public housing        Impact of welfare     18.5% increase
                                      reform assuming a
                                      $50 to $100 loss in
                                      each recipient's
                                      rental payment

David M.        Public housing        Impact of welfare     Results ranged from
Griffiths &                           reform with and       a 7.2% increase to a
Associates,                           without proactive     8.2% decrease
Ltd./District                         agency management\b
of Columbia

Minneapolis     Public housing and    Impact of local       2.8% increase
                Section 8             provision reducing
                certificates and      each recipient's
                vouchers              TANF benefits by
                                      $100, worst-case
                                      scenario

St. Paul        Public housing and    Impact of local       4% increase
                Section 8             provision reducing
                certificates and      each recipient's
                vouchers              TANF benefits by
                                      $100, worst-case
                                      scenario

New Jersey      Section 8             Impact of losing      1.8% increase
                certificates          TANF income for
                                      those whose sole
                                      source of income is
                                      TANF, worst-case
                                      scenario

HUD/Eight       Public housing        Impact of welfare     Results ranged from
housing                               reform; optimistic    a 21.8%\d increase
agencies\c                            and conservative      to a 19.6% decrease
                                      scenarios and a
                                      worst-case scenario
                                      with and without
                                      minimum rents
--------------------------------------------------------------------------------
\a The original provisions of welfare reform that affected immigrants
have been significantly modified, and this study was completed
shortly after the passage of the original reforms. 

\b The author of this study defined proactive housing agency
management as taking an active role in mitigating the impact of
welfare reform on the housing agency's residents and the housing
agency. 

\c The eight housing agencies are in Dallas, Texas; Richmond and
Norfolk, Virginia; Los Angeles and San Francisco, California; and
Cleveland, Columbus, and Toledo, Ohio. 

\d Based on a preliminary California welfare plan introduced by the
governor in Jan.  1997.  This plan is more restrictive than the one
adopted in Aug.  1997. 

Source:  GAO's analysis of the state and local studies listed in app. 
I. 


--------------------
\5 Welfare Reform Impacts on the Public Housing Program:  A
Preliminary Forecast, HUD, Office of Policy Development and Research,
Division of Policy Studies (Mar.  1998). 


      DIFFERING FOCUS LEADS TO USE
      OF DIFFERENT KEY ASSUMPTIONS
---------------------------------------------------------- Letter :3.2

In addition to differences in geographic scope, the studies we
reviewed differed in other key aspects of their focus, and these
differences often dictated the assumptions used in the studies.  Some
of these studies took a worst-case approach to welfare reform,
imposing very conservative assumptions about the employment and
earnings prospects of welfare recipients with housing assistance. 
Generally, these analyses were designed to heighten the awareness of
the welfare and housing assistance communities to the worst possible
implications of some aspects of welfare reform and to prompt these
communities to take appropriate action.  For example, at the national
level, the Council of Large Public Housing Authorities used a
worst-case approach in the summer and early fall of 1996 to look at
what would happen if all residents receiving cash assistance lost
that assistance 5 years after the implementation of welfare reform
(the federally mandated time limit) and if none of the affected
families were able to replace any of these benefits with wage
earnings.  This analysis, which was designed to motivate the public
housing community to take action, estimated that required rental
payments could fall by 30 percent (the percentage of income that
residents generally pay in rent) of the total amount of cash
assistance lost.  At the local level, the Minneapolis and St.  Paul
housing authorities designed studies to show the maximum potential
effect of Minnesota's decision to reduce by $100 the monthly cash
benefits for TANF recipients with housing assistance.  To estimate
the maximum impact of the $100 reduction on the housing agency and
HUD, these studies ignored any possibility that residents receiving
TANF benefits might have additional earnings to offset some of the
$100 loss in benefits.  Thus, the monthly rental payments of
subsidized households would be $30 ($100 times 30 percent) less than
they otherwise would have been.  Under this scenario, the rental
payments of public housing residents and Section 8 certificate and
voucher holders would decline annually by $817,000 for Minneapolis
and over $1 million for St.  Paul.\6

Other studies, designed to forecast HUD's actual budget needs,
attempted to make best-guess estimates of the financial impact of
welfare reform on HUD and housing agencies.  HUD's national study and
one by the Congressional Budget Office (CBO) used more elaborate
methods to predict the employment and earnings prospects of welfare
recipients.  For example, relying on various studies of the earnings
of former welfare recipients,\7 CBO assumed that households with
Section 8 assistance would be able to replace two-thirds of their
lost welfare benefits with earnings when time limits take effect. 
These studies generally found that welfare reform would have a more
modest impact than studies designed to look at the worst-case
outcomes of imposing time limits.  Still other studies, including the
Johns Hopkins study, HUD's multisite study, and the Los Angeles
study, were designed to determine a range of likely effects of
welfare reform on HUD and housing agencies and to identify the
factors that might influence the extent of these effects.  Most of
these studies focused on the potential impact of a broad range of
factors--including welfare policies (e.g., time limits and employment
sanctions), housing policies (e.g., minimum rents and exclusions of
earned income from rental payment calculations), and local and
national economic conditions--to determine which factors would have
the largest impact on the subsidies needed.  The studies used varying
assumptions and models to estimate welfare reform's impact under
different scenarios.  For example, HUD's multisite study and the Los
Angeles study estimated the impact of welfare reform under both
optimistic and conservative assumptions about the employment and
earnings prospects of welfare recipients with housing assistance. 
Additionally, both HUD's multisite study and the Johns Hopkins study
varied their assumptions about welfare reform's rules by examining
the potential effects of different state welfare programs. 

Some of the studies discussed above focused on the impact of certain
housing policies.  For example, HUD's multisite study found that
because rental payments do not fall to zero when tenants lose their
cash income but are required to pay minimum rents, the imposition of
such minimum rents could offset much of the potential decline in
rental revenues resulting from welfare reform.\8 The Los Angeles
study attempted to measure how much of the potential increases in
rental payments the housing agencies would forgo because of policies
excluding new income from rental payment calculations.\9 Finally,
studies designed by David Griffiths & Associates, Ltd.  (DMG),
focused on the impact of significant involvement by housing agency
managers in helping tenants move into the labor market.  DMG assumed,
in its studies for both the District of Columbia and the Public
Housing Authorities Directors Association, that a high level of
involvement by housing agency managers would significantly improve
the income prospects of welfare recipients. 


--------------------
\6 The Minnesota legislature has postponed the imposition of the $100
reduction in benefits until July 1999. 

\7 See, for example, Alberto Martini, "Potential Effects of
Congressional Welfare Reform Legislation on Family Income," The Urban
Institute (July 26, 1996); Sandra K.  Danziger and Sherrie A. 
Kossoudji, "When Welfare Ends:  Subsistence Strategies of Former GA
Recipients:  Final Report of the General Assistance Project,"
University of Michigan (Feb.  1995); and Thomas M.  Fraker et al.,
"Iowa's Limited Benefit Plan," Mathematica Policy Research, Inc., and
the Institute of Social and Economic Development (May 1997). 

\8 See footnote 4 for information on recent legislative changes
establishing exceptions to the minimum rent requirements. 

\9 A study by the Atlanta housing authority also examined the effects
of excluding former welfare households' income from rent
calculations.  This study found that such exclusions could wipe out
the potential for significant reductions in the Atlanta housing
agency's need for HUD subsidies.  Without exclusions, the need for
subsidies would decline if the earnings of former welfare households
exceeded the welfare payments they once received. 


      STUDIES' AUTHORS EXPRESSED
      UNCERTAINTY
---------------------------------------------------------- Letter :3.3

The varied assumptions underlying the studies we reviewed reflect
researchers' uncertainties about changes in welfare and housing
policies over time, the future of the economy, and the behavioral
responses of welfare recipients.  The authors of several of the
studies described their estimates as outdated because events (such as
the final version of a state's welfare reform law or the condition of
the national economy) had not played out as the authors had
anticipated when they conducted their studies.  For example, the
representative of DMG who developed the estimates for the Public
Housing Authorities Directors Association and the District of
Columbia told us that if he were developing the estimates today, he
would revise the results of his pessimistic scenario significantly to
take account of (1) modifications to the welfare reform law that have
reduced the cuts in Supplemental Security Income he initially
anticipated, (2) the significant emphasis HUD has placed on programs
to help move people from welfare to work, and (3) the surprisingly
strong economy.  Similarly, the authors of HUD's multisite study told
us that their estimates for Los Angeles are probably too pessimistic
because they assumed a more restrictive welfare program than the one
California actually adopted in August 1997.\10 Other authors were
also concerned about the general difficulty of predicting the future
behavior of TANF recipients.  For example, officials at CBO stated
that because of uncertainties about how welfare reform would be
implemented and how recipients would respond, they recognized that
their estimates could be substantially different from actual
outcomes.  And, as we reported in January 1998, HUD is no longer
standing behind its initial assessment of welfare reform's nationwide
impact, in part, because of difficulties it identified in predicting
how states will implement welfare reform plans and how welfare
recipients will respond to welfare reform. 


--------------------
\10 HUD researchers said that they used the provisions of an early
version of California's welfare reform plan (Jan.  1997), proposed by
the governor, which would have reduced benefits for whole families
when time limits were met or sanctions were imposed.  However, the
California welfare reform law that was adopted in Aug.  1997
continues to pay benefits to children out of state funds regardless
of time limits or sanctions. 


   MULTIPLE ISSUES COMPLICATE
   EFFORTS TO FORECAST WELFARE
   REFORM'S FINANCIAL IMPACT ON
   HUD'S HOUSING SUBSIDY PROGRAMS
------------------------------------------------------------ Letter :4

The experts with whom we spoke generally agree that several
methodological and data issues complicate efforts to forecast welfare
reform's financial impact on HUD's housing subsidy programs.  Some
issues, such as differences in state welfare policies and plans for
implementing welfare reform and uncertainty about the strength of the
economy and the behavior of welfare recipients, make it difficult to
predict the impact of welfare reform itself.  Housing researchers
generally agree, however, that estimating welfare reform's financial
impact on housing programs is more complex than estimating welfare
reform's impact overall because the characteristics of welfare
recipients with housing assistance may be different from those of
other welfare recipients, and housing agencies and landlords may
adopt a broad range of housing philosophies and policies.  Finally,
the lack of consistent and reliable data further hampers researchers'
efforts to predict welfare reform's financial impact on HUD's housing
programs with any certainty. 


      STATE WELFARE REFORM PLANS
      AND IMPLEMENTATION DIFFER
      CONSIDERABLY
---------------------------------------------------------- Letter :4.1

Differences in state welfare policies have always been important in
evaluating the federal welfare program.  Under AFDC, states paid
different levels of benefits to entitled recipients, and HHS
researchers reported that recipients were more likely to leave the
welfare rolls in states with lower benefits than in states with
higher benefits.  Because welfare reform gave the states greater
discretion in setting welfare policy, state policies now differ
across a multitude of dimensions.  For example, the states can now
determine who is eligible for cash assistance and for how long.  In
addition, they can set specific work requirements and establish
sanctions for recipients who violate their state welfare policies. 

Differences in the implementation of welfare plans at the state and
local levels may exacerbate interstate differences in the impact of
welfare reform.  In particular, the manner in which caseworkers relay
information to and interact with recipients affects outcomes under
welfare reform.  For example, in evaluating welfare reform in several
states, the Manpower Demonstration Research Corporation (MDRC) found
that differences in what caseworkers told clients in Florida and
Minnesota help to explain differences in the timing of caseload
reductions in those states.  MDRC found that in Florida, where
recipients could receive cash benefits for a maximum of 2 years,
recipients tended to exhaust their benefits before getting jobs and
therefore caseloads did not decline quickly, while in Minnesota,
where recipients could receive federal cash benefits for 5 years,
caseloads dropped quickly.  MDRC told us that this difference in
behavior seemed to occur, at least in part, because Florida
caseworkers encouraged recipients to remain on TANF and spend their 2
years investing in job skills, while Minnesota caseworkers advised
their clients to become employed as soon as possible and save their
limited benefits for possible future needs. 


      ASSUMPTIONS ABOUT ECONOMIC
      HEALTH AND RECIPIENTS'
      RESPONSES ARE KEY
---------------------------------------------------------- Letter :4.2

The studies we identified varied widely in the assumptions they used
to predict welfare recipients' potential employment prospects and
earnings.  Experts with whom we spoke generally agree that welfare
recipients' employment prospects and earnings depend on the market
for low-skilled workers.  The demand for these workers generally
depends on the overall health of the national and local economy,
while the supply depends on recipients' responses to the level of
wages they could earn and the level of welfare benefits they could
receive. 

In general, some issues make it difficult to predict the demand for
low-skilled workers.  First, because welfare reform has thus far
occurred during a period of strong national job growth, researchers
have little sense of how the demand for low-skilled labor will hold
up during an economic slowdown.  For example, experts have been
unable to agree on how much of the recent decline in the number of
families receiving cash assistance is the result of the very robust
economy in recent years and how much is the result of welfare
reform.\11 In order to shed light on the degree to which economic
conditions affect the impact of welfare reform, HUD researchers, in
their multisite analysis, studied at least two cities with different
economic conditions in each of three states.  Welfare recipients in
the same state were generally subject to the same welfare laws. 
While HUD found that welfare recipients in all three states had more
success in entering labor markets in cities with more robust local
economies, the difference was not consistent across the states. 
Second, some researchers have noted that while it may be possible to
measure the number of low-skilled jobs available at a given point in
time, this type of analysis will not necessarily reveal how many
people can find employment over a period of time because of constant
turnover in employment and other changes in labor market conditions
over time. 

The supply of low-skilled workers will depend, in part, on how
welfare recipients respond to changes in their state's welfare
program.  While some researchers believe that past studies of the
impact of changing wages and benefit levels on welfare recipients'
desire to work could help answer this question, other researchers
believe that the 1996 welfare reforms constitute such a dramatic
shift from earlier welfare policies that past behavior may not be a
good predictor of future behavior.  Thus, there is little consensus
among experts about the future behavior of welfare recipients. 


--------------------
\11 Two recent national studies of the period from 1993 through 1996
point to economic expansion as a key factor in explaining the decline
in the number of families receiving cash assistance.  However, both
studies also conclude that in states such as Oregon and Wisconsin,
which implemented stringent welfare reform provisions during this
period, these provisions also contributed significantly to declines
in the number of families receiving cash assistance.  See Council of
Economic Advisers, Explaining the Decline in Welfare Receipt,
1993-1996 (Washington, D.C.:  Council of Economic Advisers, May 9,
1997) and James Ziliak et al., Accounting for the Decline in AFDC
Caseloads:  Welfare Reform or Economic Growth?  (Ann Arbor, Mich.: 
University of Michigan, Sept.  1997).  A critique of the Council of
Economic Advisers' study disputes the Council's claim to have
explained the decline in the number of people receiving cash
assistance.  See Alberto Martini and Michael Wiseman, Explaining the
Recent Decline in Welfare Caseloads:  Is the Council of Economic
Advisers Right?  (Washington, D.C.:  Income and Benefits Policy
Center, The Urban Institute, July 1997). 


      TENANTS' CHARACTERISTICS AND
      HOUSING POLICIES FURTHER
      COMPLICATE ESTIMATION
---------------------------------------------------------- Letter :4.3

Housing experts with whom we spoke generally agree that estimates of
the effect of welfare reform on the general welfare population may
not apply to the subset of welfare recipients in public and assisted
housing.  As we reported in June 1998, welfare recipients living in
public housing are more likely to have been on welfare longer than
those without housing assistance, and longer spells on welfare have
been associated with less success in obtaining employment.\12 In
addition, experts suggest that welfare recipients with housing
assistance may be less likely to go to work for several reasons: 

  -- Welfare recipients with housing assistance will be able to
     retain a smaller portion of their new earnings because they will
     generally be required to pay 30 percent of those earnings in
     rent. 

  -- Because housing assistance provides a "cushion," welfare
     recipients with public or assisted housing may have less
     incentive to work than other welfare recipients with the same
     job prospects but no housing assistance. 

  -- Recent evidence suggests that job growth is occurring in the
     suburbs while welfare recipients are likely to live in urban
     centers or rural areas.  In particular, welfare recipients with
     project-based housing assistance (including both public housing
     and project-based Section 8 assistance) may face higher
     relocation costs than other welfare recipients because they may
     have to give up their housing assistance to move to locations
     with better job prospects. 

The combination of longer periods on welfare and less incentive to
work may help explain why some researchers have found that welfare
recipients with housing assistance are less successful in moving from
government-sponsored job training programs into long-term private
sector employment.  A recent study by MDRC researchers in Atlanta
showed that of the participants in certain federal job training
programs, those with no housing assistance were most likely, those
with certificates and vouchers slightly less likely, and those in
public housing least likely, to find employment after completing
their training.\13

Because of recent housing policy changes and uncertainty about the
degree to which housing agencies will adopt these changes,
researchers will have difficulty separating the effects of welfare
reform from those of changes in housing policy, just as they have had
difficulty separating the effects of welfare reform from those of
overall economic conditions.  Many of the studies we reviewed, as
well as experts we consulted, recognized the importance of recent
changes in rent and admission policies and programs to move welfare
recipients to work.  For example, the director of the housing
authority in Athens, Georgia, told us that the types of admission
preferences, the effect of management's involvement in helping
tenants obtain employment, and the level of minimum rents could
determine whether his agency's rental revenues rise or fall with
welfare reform.  However, recent legislation may limit the degree to
which housing agencies will be able to use minimum rents to offset
potential declines in rental payments under welfare reform. 


--------------------
\12 Welfare Reform:  Changes Will Further Shape the Roles of Housing
Agencies and HUD, (GAO/RCED-98-148, June 25, 1998). 

\13 Alternatively, one expert argues on the basis of preliminary data
from Minnesota that welfare recipients with housing assistance are
more likely than other welfare recipients to find and retain jobs
because, with housing assistance, they are more stable and less
likely to be forced to move.  The Minnesota data also suggest that
those with housing assistance are receiving higher earnings.  In
addition, a study in Los Angeles has shown that welfare recipients
with Section 8 certificates and vouchers are more likely to be
employed than those in public housing or those without housing
assistance. 


      DATA AVAILABILITY AND
      RELIABILITY POSE ADDITIONAL
      OBSTACLES
---------------------------------------------------------- Letter :4.4

Welfare and housing researchers have used a combination of government
administrative data--data collected by federal, state, or local
officials on a host of factors associated with the recipients of
welfare and housing assistance--and survey data to study the behavior
of those who receive welfare and housing assistance.  Administrative
databases generally provide information over time on the participants
in a program, while survey data generally conform more closely to
research objectives but cover a smaller number of households. 
Because the states have more flexibility to design their own systems
under welfare reform, data elements in administrative and survey
databases may be less consistently defined than in the past. 
Although state welfare agencies have reported administrative data
under HHS' emergency rules, which were phased in over a period of 9
months beginning in July 1997, some states have submitted data that
are not fully consistent with HHS' specifications.  HHS will be
issuing final TANF reporting rules that better define terms, but
according to an HHS official, the states will continue to have
significant flexibility in how they define their programs, making
data assessment more difficult.  Similarly, according to an official
in the Census Bureau's Housing and Household Economic Statistics
Division, the increased interstate variation promoted by the 1996
welfare reforms has placed a significant burden on national
organizations that collect survey data.  For example, obtaining
consistent data across states about the level of cash benefits may be
difficult because the states have given their TANF benefits a variety
of names.  For example, Minnesota calls TANF the Minnesota Family
Investment Plan, while California refers to TANF as CalWORKS.  In
addition, questioners and respondents may not know how to classify
the increasingly common "one-time diversion" payments, which states
use to provide one-time assistance to families in lieu of placing
them on the welfare rolls. 

We and others have questioned the reliability of the existing
national administrative and survey data on the residents of public
and assisted housing.  HUD collects administrative data on the
residents of public and tenant-based assisted housing in its
Multifamily Tenant Characteristics System database.  Housing agencies
are supposed to provide information for this database to HUD
electronically in a specified format, but some agencies, especially
the larger ones, do not report data for all of their residents, and
the data contain errors as well.  A HUD official told us that in
recent years, HUD has concentrated on improving the response rate for
these data, but the greater response has been accompanied by an
increase in the number of data entry errors.  HUD collects similar
data on the residents of properties with project-based Section 8
assistance in the Tenant Rental Assistance Certification System
database.  According to HUD officials, this database suffers, on a
smaller scale, from reporting problems and data errors such as those
affecting the multifamily database.  The reliability of survey data
on housing assistance is also questionable because respondents to
surveys with questions about this assistance often misreport their
status.  HUD has documented probable misreporting in the Current
Population Survey and the American Housing Survey.  For example, in a
paper presented in May 1996, HUD economists reported that the
majority of those receiving housing assistance who said they lived in
public housing actually do not.\14 Furthermore, they reported that
the majority of the families receiving other housing assistance do
not accurately identify the way they are assisted, and perhaps
one-fifth of those who report receiving a housing subsidy do not
actually receive one.  In addition, the interim director of the Johns
Hopkins University's Center for Policy Studies has noted similar
reporting discrepancies in survey responses and administrative data
and has suggested methods for improving the reliability of the survey
responses.  Although the Census Bureau and others are developing and
testing questions to improve survey responses on welfare benefits and
housing subsidies and HUD has worked to improve its data as well, it
is still too early to obtain adequate data to test assumptions about
the outcomes of recent welfare and housing reforms. 


--------------------
\14 Mark Shroder and Marge Martin, New Results From Administrative
Data:  Housing the Poor, or, What They Don't Know Might Hurt Somebody
(May 1996). 


   AGENCY COMMENTS
------------------------------------------------------------ Letter :5

We provided a draft of this report to HUD for review and comment. 
HUD stated that the report is fair and straightforward and provided
some technical corrections.  HUD's comments appear in appendix III. 
In addition, we provided excerpts of a draft of this report to the
authors of each of the studies we reviewed.  Several of the
researchers and housing agencies provided us with comments that we
incorporated as appropriate. 


   SCOPE AND METHODOLOGY
------------------------------------------------------------ Letter :6

To identify studies that estimated welfare reform's financial impact
on housing assistance programs, we contacted known experts and
officials from a variety of organizations and government agencies. 
In particular, we spoke with experts in housing and welfare research,
representatives of major trade associations and advocacy groups,
officials at 30 of the largest local housing authorities and 10 of
the largest state housing agencies, and officials from 10 private
management companies of various sizes that are managing properties
with Section 8 subsidies.\15 Although we attempted to identify as
many studies as possible, we recognize that the 13 studies we
identified (see app.  I) may not include all such studies that were
performed.  It should also be noted that it was beyond the scope of
this review to assess the quality of the research underlying the
individual estimates we reviewed.  To consistently present the
different studies' estimates of dollar changes in rental revenues,
costs, and subsidies, we presented each study's findings as the
percentage change in the subsidy relative to the total performance
fund and/or housing assistance payments HUD says the agency received
in 1997.  We ignored the facts that HUD does not always provide 100
percent of the subsidy needed for public housing, as calculated under
the performance funding system formula, and that HUD's outlays may
lag behind changes in rental revenues by 2 to 3 years.  In addition,
although the results presented here are based on the assumption that
HUD will provide 100 percent of the needed subsidy, the studies we
reviewed made different assumptions about the percentage of the
needed subsidy that HUD would be likely to provide.  These
assumptions ranged from 85 percent to 100 percent.  To the extent
that the subsidy is funded at less than 100 percent, more of the cost
of welfare reform will be borne by local housing agencies and less
will be borne by HUD.  See appendix I for a list of the individual
studies we reviewed. 

To better understand the methodological and data issues that arise
when estimating welfare reform's financial impact on HUD's housing
programs, we also contacted known experts in welfare and housing who
represented a broad range of views.  We questioned them about the
importance of specific methodological and data concerns using a
semistructured questionnaire.  We also gathered their research and
analyzed the information collected from the interviews and research
documents to develop common themes.  Appendix II lists the experts
with whom we spoke about methodological and data issues. 

We conducted our work from May 1998 through November 1998 in
accordance with generally accepted government auditing standards. 


--------------------
\15 The names of these private companies were provided to us by the
acting executive director of the National Affordable Housing
Management Association. 


---------------------------------------------------------- Letter :6.1

If you or your staff have any questions about this report, please
contact me at (202) 512-7631.  Major contributors to this report were
Amy Abramowitz, Nancy Barry, DuEwa Kamara, and Lara Landeck. 

Judy A.  England-Joseph
Director, Housing and Community
 Development Issues


IDENTIFIED STUDIES
=========================================================== Appendix I


   NATIONAL STUDIES
--------------------------------------------------------- Appendix I:1

"Background Materials for Baseline Projections of Spending Under
Current Law." Congressional Budget Office, Washington, D.C.:  Mar. 
1998. 

"The Fiscal Impacts of Welfare Reform:  An Early Assessment." Council
of Large Public Housing Authorities.  Issue brief.  Fall 1996. 

"Impact of Welfare Reform in Public Housing." David M.  Griffith &
Associates, Ltd.  Sponsored by the Public Housing Authorities
Directors Association.  Unpublished.  Spring 1997. 

Newman, Sandra and Joseph Harkness.  "The Effects of Welfare Reform
on Housing:  A National Analysis." Presented at the Policy Research
Roundtable on the Implications of Welfare Reform for Housing,
sponsored by the Fannie Mae Foundation in collaboration with the
Institute for Policy Studies at The Johns Hopkins University (work in
progress).  July 22, 1997.  For updated information, see "The Effects
of Welfare Reform on Housing:  A National Analysis" in Newman, Sandra
J.  (ed.) The Home Front:  Implications of Welfare Reform for Housing
Policy, Washington, D.C.:  The Urban Institute Press, forthcoming. 

"Technical Paper:  Welfare Reform Budgeting." U.S.  Department of
Housing and Urban Development (HUD).  Washington, D.C.:  Oct.  4,
1996. 


   STATE AND LOCAL STUDIES
--------------------------------------------------------- Appendix I:2

"Estimated PHA Income Loss Due to Proposed AFDC Cuts." St.  Paul
Public Housing Agency, Unpublished.  Feb.  28, 1997. 

"Impact of Welfare Reform on Program Costs." New Jersey Department of
Community Affairs' Section 8 Housing Program.  Unpublished.  Jan. 
1998. 

Nguyen, Mai, Charles Kastner, and Ashley Lommers-Johnson.  "Welfare
Reform:  Status of Washington State's Welfare Reform Plan; Effects on
Residents, the Seattle Housing Authority, and Neighborhoods; and
Prospects for Employment and Rent Income." Presented at HUD
headquarters, Washington, D.C.  Dec.  17, 1996. 

"Potential Impact of CalWORKS." Housing Authority of the City of Los
Angeles.  Unpublished.  Nov.  1997. 

"Welfare Act Impact Analysis:  District of Columbia Housing Authority
(DCHA)--Final Report." David M.  Griffith & Associates, Ltd.  Apr. 
1997. 

"Welfare Reform Impact Study." Minneapolis Public Housing Authority. 
Unpublished.  Spring 1997. 

Welfare Reform Impacts on the Public Housing Program:  A Preliminary
Forecast.  U.S.  Department of Housing and Urban Development.  Office
of Policy Studies and Research.  Rockville, MD., Mar.  1998. 

"Welfare Reform Program and Financial Analysis." Miami-Dade Housing
Agency.  Unpublished.  Oct.  1997. 


METHODOLOGICAL AND DATA EXPERTS
GAO INTERVIEWED
========================================================== Appendix II


   GOVERNMENT EXPERTS
-------------------------------------------------------- Appendix II:1

Paul Cullinan, Chief, Human Resources Cost Estimate Unit,
Congressional Budget Office (CBO)

Shelia Dacey, Analyst, CBO

Debra Devine, Social Science Analyst, Office of Policy Development
and Research, U.  S.  Department of Housing and Urban Development
(HUD)

Katherine L.  Meredith, Program Examiner, Housing Branch, Executive
Office of the President, Office of Management and Budget (OMB)

Charles Nelson, Assistant Division Chief for Income and Poverty,
Housing and Economic Household Statistics Division, Bureau of the
Census, U.  S.  Department of Commerce

Don Oellerich, Acting Deputy Chief Economist, Office of the Assistant
Secretary of Planning and Evaluation, U.  S.  Department of Health
and Human Services (HHS)

Carla Pedone, Analyst, CBO

F.  Stevens Redburn, Chief of the Housing Branch, Executive Office of
the President, OMB

Les Rubin, Social Science Analyst, Office of Policy Development and
Research, HUD

Ron Sepanik, Director, Housing and Demographic Analysis Division,
Office of Policy Development and Research, HUD

Mark Shroder, Economist, Policy Development Division, Office of
Policy Development and Research, HUD

Rueben Snipper, Analyst, Office of the Assistant Secretary of
Planning and Evaluation, HHS

Christopher Snow, Analyst, Office of the Assistant Secretary of
Planning and Evaluation, HHS


   ACADEMIC EXPERTS
-------------------------------------------------------- Appendix II:2

Neil Bania, Associate Director, The Center for Urban Poverty and
Social Change, Mandel School of Applied Social Science, Case Western
Reserve University

Claudia Coulton, Lillian F.  Harris Professor and Co-Director, The
Center for Urban Poverty and Social Change, Case Western Reserve
University

Joseph Harkness, Research Statistician, Institute for Policy Studies,
The Johns Hopkins University

Jon Jacobson, Senior Research Analyst, Mathematica Policy Associates

Laura Leete, Research Associate, The Center for Urban Poverty and
Social Change, and Assistant Professor of Economics, Case Western
Reserve University

Sandra Newman, Interim Director, Institute for Policy Studies, The
Johns Hopkins University

LaDonna Pavetti, Senior Research Analyst, Mathematica Policy
Associates


   PUBLIC POLICY ANALYSTS
-------------------------------------------------------- Appendix II:3

Marty Abravenel, Senior Research Associate, The Urban Institute

Anna Kondratas, Senior Associate, The Urban Institute

Cynthia Miller, Research Associate, Manpower Demonstration Research
Corporation

James Riccio, Senior Research Associate, Manpower Demonstration
Research Corporation

Barbara Sard, Director of Housing Policy, Center on Budget and Policy
Priorities

John Weicher, Senior Fellow, Hudson Institute

Sheila Zedlewski, Director, Income and Benefits Center, The Urban
Institute


   OTHER PUBLIC AND ASSISTED
   HOUSING EXPERTS
-------------------------------------------------------- Appendix II:4

Jim Brigle, Director of Government Affairs, Public Housing
Authorities Directors Association

George C.  Caruso, Acting Executive Director, National Affordable
Housing Mangement Association

Connie Campos, Policy Analyst for Housing, National Association of
Housing and Redevelopment Officials

Major Galloway, Policy Analyst for Housing, National Association of
Housing and Redevelopment Officials

Debra Gross, Research Director, Council of Large Public Housing
Authorities

John Hiscox, Executive Director, Macon Housing Authority

Walter Huelsman, Vice President and National Director of Housing
Consulting, DMG-Maximus, Inc. 

Will Jones, Housing and Community Development Research Officer,
National Association of Housing and Redevelopment Officials

Rich Parker, Executive Director, Athens Housing Authority




(See figure in printed edition.)Appendix III
COMMENTS FROM THE DEPARTMENT OF
HOUSING AND URBAN DEVELOPMENT
========================================================== Appendix II



(See figure in printed edition.)

The following are GAO's comments on the Department of Housing and
Urban Development's letter dated November 10, 1998. 


   GAO COMMENTS
-------------------------------------------------------- Appendix II:5

1.  After additional discussions with HUD to clarify the information
provided in comment 1, we included the data from HUD's suggested
paragraph with certain appropriate modifications. 

2.  We deleted the footnote as suggested. 

3.  We changed the reference as requested. 


*** End of document. ***