National Academies Press: OpenBook

Rebuilding the Research Capacity at HUD (2008)

Chapter: 9 The Relationship Between Research and Policy Development

« Previous: 8 Dissemination
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 149
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 150
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 151
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 152
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 153
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 154
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 155
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 156
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 157
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 158
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 159
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 160
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 161
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 162
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 163
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 164
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 165
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 166
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 167
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 168
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 169
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 170
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 171
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 172
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 173
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 174
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 175
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 176
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 177
Suggested Citation:"9 The Relationship Between Research and Policy Development." National Research Council. 2008. Rebuilding the Research Capacity at HUD. Washington, DC: The National Academies Press. doi: 10.17226/12468.
×
Page 178

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

9 The Relationship Between Research and Policy Development Previous chapters have discussed PD&R’s research and policy develop- ment activities by function. Although this is useful analytically, in practice external research, in-house research, data collection, and policy develop- ment occur simultaneously and are closely related. This chapter describes and illustrates the relationships among the activities: in particular, how research feeds into policy development and how policy development influ- ences research. The examples show that PD&R research (both external and in-house) has played an important role in the evolution of policy across all of HUD’s program areas, including its regulatory responsibilities. Before discussing those examples in HUD’s major policy areas, we briefly consider the policy process in general. HOW POLICIES EVOLVE The input of any office to the policy-making process must be timely if it is to be given consideration. If a decision has to be made on a set date, offices generally cannot develop new information from a standing start according to an ideal plan. Rather, they can summarize the information that is already available and perhaps analyze it in new ways or from new perspectives. And policy makers can only make use of the research that is available. They can draw on the findings of research that has been com- pleted or perhaps is in process. Research thus informs policy incrementally and over time. Results may be available to answer a narrow question that a secretary or assistant secretary is asking at a particular moment. Or results may not be available: Sometimes issues come up unexpectedly, and 149

150 REBUILDING THE RESEARCH CAPACITY AT HUD sometimes research takes longer than originally contemplated when the project was undertaken. For example, the committee was informed of a quick-turnaround request by Secretary Martinez for an evaluation of the small Youthbuild Program: The program had not been evaluated during its 8-year history and a full formal evaluation could not be produced in the few months between the request and the beginning of the next budget cycle. More fundamentally, however, research informs policy by shaping a shared understanding of the nature of a housing problem and how various programs work in attempting to cope with this problem—what is known, what is not known, and what challenges remain unaddressed. Research on major housing issues and programs tends to be ongoing, with new research projects being developed on the basis of findings of earlier research and on program outcomes. Each HUD administration is able to draw on that body of research, each is able to add to it during its term, and each leaves behind it a body of completed studies and studies in process that are intended to be of use to its successor. The research is ongoing because HUD has had the same basic missions for many years, as well as many of the same programs. Yet although the missions have seldom changed and the major programs have long histories, some programs have been terminated, and there have been modifications in all of them. Some of the modifications have occurred in response to changes in policy priorities and some to address program management or other problems identified in the course of program operations. Research contributes to the decision to undertake new programs and the design of the programs. Experience with the programs, once they have begun operations, often raises issues of program effectiveness or cost, and identifies problems that need attention. Research is often undertaken to address these issues, to evaluate the effectiveness of the programs, answer specific questions about them, and suggest modifications. It is an iterative process. Thus, it is necessary to take a long perspective on the contributions of PD&R to the policy development process. This is particularly true of the research program, and most particularly true in recent years as the size of the research budget has been constant or shrinking. Research is undertaken to answer questions or resolve problems; when the budget is limited, fewer questions or problems can be addressed. This reduction in research can have negative consequences for HUD policy makers and the public; use- ful information is not available when it is relevant. The cost is real, albeit indirect and easily overlooked. The research activities in any given year will not cover all of the major program areas of HUD. But over time research has covered nearly all of them. Policy development in any given year draws on the research activities of the last several years, and even longer.

RESEARCH AND POLICY DEVELOPMENT 151 The remainder of this chapter illustrates this ongoing, iterative process and provides examples of these interrelationships, drawn from all of the major program areas: Section 8 new construction; tenant-based assistance, with a focus on cost; housing vouchers, with a focus on program outcomes; the Community Development Block Grant (CDBG) formula; housing mar- ket discrimination; and regulation of the government sponsored enterprises, Fannie Mae and Freddie Mac. There is some overlap between the first two examples, since both programs were authorized under Section 8 in 1974, and some of the major research projects over the next several years covered both. Also, the second and third both concern the programs that provide assistance to households, tracing the development of policy along different dimensions and over somewhat different periods of time. The emphasis on assistance programs reflects the strong policy interest and controversies in the area of low-income housing during the years since PD&R was estab- lished. Housing assistance has routinely constituted well over one-half of the HUD budget. Section 8 New Construction The Section 8 New Construction Program was enacted in 1974, at the same time as the Section 8 Existing Housing (Certificate) Program. It differed from previous project-based subsidy programs in that the subsidy was explicitly based on the income of the assisted household. Households paid 25 percent (later raised to 30 percent) of their income toward the cost of the unit, which included the project owner’s mortgage payment and operating costs. The commitment to an income-conditioned subsidy was derived in part from Housing in the Seventies (U.S. Department of Housing and Urban Development, 1974), the major 1974 study of previous subsidy programs. That in-house study was undertaken by the first assistant secre- tary for policy development and research and largely staffed by individuals who became part of PD&R when it was created during the course of the study. (The study did not recommend enacting a program like Section 8 new construction, but its recommendation for income-conditioned subsidies became part of the program.) By the late 1970s it was becoming clear to policy makers that the Section 8 New Construction Program was exceptionally expensive. The 1981 evaluation by Abt Associates (Wallace et al., 1981), comparing costs and outcomes for Section 8 new construction and the certificate program, showed that per-unit costs were about twice as high in the new construc- tion program. With respect to outcomes, the evaluation found that the new construction program primarily served white elderly households and that few minority households participated; in contrast the certificate program was generally representative of the eligible population. Both programs

152 REBUILDING THE RESEARCH CAPACITY AT HUD served households in deficient housing. In 1982, the President’s Commis- sion on Housing recommended terminating the Section 8 New Construc- tion Program and making tenant-based assistance the primary housing subsidy program. The commission drew extensively on the Abt study and on other data and analysis from PD&R, including the American Housing Survey. The recommendations of the commission report were adopted by the administration, and in 1983, Congress repealed the Section 8 New Construction Program. The repeal applied to further projects. The inventory of Section 8 projects remained as assisted housing. By the early 1990s, the question of whether and how to preserve these projects for their low-income residents became an important public policy concern. Like its predecessor program (under Section 236), the subsidy contracts for Section 8 new construction had a 20-year term, after which the owners could opt out of the program. PD&R conducted a survey and analysis of Section 8 projects insured by the Federal Housing Authority (FHA) as of 1990-1991, which included estimates of per-unit annual subsidy cost, per-unit backlog of needed repairs to bring the units up to market standards, and per-unit annual accrual of repair needs. This study, with results published in 1992 and 1993, provided the most extensive data yet available on the Section 8 inventory. It became a basic resource in the policy debates over “preservation” during 1995-1997. In addition, PD&R produced a number of “Issue Briefs” and provided other analyses and data as contributions to the debate. In 1997 Congress enacted the Mark to Market Program to preserve as much of Section 8 inventory as financially reasonable and provide housing assistance for the residents of those projects whose owners chose to convert them to market-rent housing. Rents were marked down to the fair market rents for existing housing in the local market, thus lowering the subsidy. In addition, the project mortgage was restructured so that the subsidies and tenant rents were sufficient to cover the payments on a new first mortgage; the remainder of the original mortgage became a second mortgage on which payments were to be made if funds were available or when the project was sold. A sunset date of 2006 was established for the Mark to Market Pro- gram, and PD&R funded a major evaluation covering the period through 2003, which was published in 2005. Congressional staff who met with the committee identified this evaluation as an important resource for policy makers in the deliberations that led to the reauthorization of the program in 2006. This evaluation was the most recent in a series of major PD&R research projects, and numerous smaller-scale in-house analyses, that con- tributed to the policy process over the life of the Section 8 New Construc- tion Program.

RESEARCH AND POLICY DEVELOPMENT 153 Cost-Effectiveness of Low-Income Housing Programs Since low-income housing assistance accounts for more than three- fourths of HUD’s budget, research on this topic should be and has been an important part of HUD’s research agenda. One of the most important policy decisions in this area concerns how much of the total budget to allocate to individual programs. Good information about the comparative performance of different programs is essential for making good allocation decisions. This section assesses the past contribution of HUD-funded research on comparative performance to better inform decisions concerning the alloca- tion of the budget for rental housing assistance to particular programs, and it suggests some important opportunities for future contributions. It focuses on the research on the cost of providing equally good housing under different programs. The available evidence, which is largely HUD funded, indicates that this is the largest difference in the performance of different housing programs. Differences in cost-effectiveness are of great significance for policy. When needlessly expensive methods of delivering housing assistance are used, many low-income households that could have been provided with adequate housing at an affordable rent within the current housing assis- tance budget continue to live in deplorable housing, and taxpayers pay unnecessarily high taxes to achieve that outcome. A Brief History Between 1937 and 1965, the U.S. government delivered rental housing subsidies to low-income households exclusively through the construction and operation of housing projects for these households. Local public hous- ing authorities operated all of the projects built during the first 17 years. In 1954, the federal government began to contract with private parties to build and operate projects for low-income households, while still continu- ing to build public housing projects. In 1965 Congress enacted Section 23, a program under which public housing authorities could lease apartments in existing private unsubsidized housing for the use of households that were eligible for public housing. One variant of this small program allowed tenants to locate their own apartments that met the program’s minimum standards. This was the first program of tenant-based rental assistance in the United States.   or F a summary of the evidence on many other aspects of program performance, see Olsen (2003).

154 REBUILDING THE RESEARCH CAPACITY AT HUD In 1974, the Section 8 Existing Housing Program replaced Section 23. Since then, tenant-based Section 8 has become the country’s largest pro- gram of housing assistance. It now serves about 43 percent of the low- income households that receive HUD rental assistance. In 1983, Congress cancelled HUD’s authority to make new commitments under the Section 8 New Construction and Substantial Rehabilitation Program, its last large program to subsidize the construction and major rehabilitation of new housing projects. In 1995, the Clinton administration proposed a sweep- ing reform of programs of low-income housing assistance that involved the gradual replacement of all project-based assistance with tenant-based vouchers. Congress did not adopt these proposals, but the 1998 Quality Housing and Work Responsibility Act of 1998 (QHWRA) required public housing authorities to “voucher out” housing projects under certain cir- cumstances and allowed them to do it in other cases. Interaction of HUD-Funded Research and Housing Policy HUD-funded research on the cost-effectiveness of low-income housing programs was influential in the enactment of the Section 8 Existing Housing Program. In 1973, the National Housing Policy Review Taskforce produced the first estimates of the cost-effectiveness of low-income housing pro- grams. This research indicated that the total cost of the housing provided under the public housing program and Section 236, HUD’s largest program that subsidized the construction of privately owned projects, significantly exceeded the market rents of these units and hence that households with tenant-based assistance could occupy equally good housing in the private market at a lower cost to taxpayers (U.S. Department of Housing and Urban Development, 1974). Secretary James Lynn was thoroughly briefed on these results and the methods used to obtain them. In his cover letter to the report of the task force, he said “the report was the basis for the hous- ing policy recommendations included in President Nixon’s message to Con- gress of September 19, 1973.” These recommendations led to the creation of the tenant-based Section 8 Existing Housing Program in 1974. The 1974 Housing Act eliminated subsidies for additional projects under HUD’s Section 236 Program, but it did not end HUD subsidies for privately owned subsidized projects. Instead, it replaced Section 236 with the Section 8 New Construction and Substantial Rehabilitation Program   he T task force was housed in HUD’s headquarters building and staffed with civil servants on loan to HUD from other departments, academic consultants, and others. Michael Moskow, the head of the task force, became HUD’s first assistant secretary for policy development and research.

RESEARCH AND POLICY DEVELOPMENT 155 as HUD’s primary vehicle for subsidizing the construction and substantial rehabilitation of privately owned projects for low-income households. HUD-funded research in the late 1970s and early 1980s, based on much more detailed data on housing characteristics than underlay the results of the National Housing Policy Review, showed that tenant-based housing assistance provided equally good housing at a much lower total cost than the Section 8 New Construction and Substantial Rehabilitation Program and its predecessor, Section 236 (Mayo et al., 1980; Wallace et al., 1981). This evidence played an important role in persuading Congress to terminate the Section 8 New Construction and Substantial Rehabilitation Program in 1983. Since then, few new units have been authorized under HUD’s remaining active construction programs. The savings to taxpayers from the shift in budget from project-based to tenant-based assistance since 1974 has enormously exceeded the com- bined cost of the Housing Policy Review Taskforce, the HUD-funded cost- e ­ ffectiveness studies of the late 1970s and early 1980s, and indeed the entire budget of PD&R over the past three decades. The lowest estimate of the excess cost of project-based relative to tenant-based housing assistance for providing equally good housing based on detailed data on the housing pro- vided is 35 percent. In 2006, HUD spent about $15 billion on the Section 8 Housing Choice Voucher Program. To serve the families assisted by this program equally well with project-based assistance would have cost taxpay- ers at least an additional $5 billion. Alternatively, if HUD had devoted its entire budget for low-income housing to project-based assistance, it would have served many fewer households. Smaller, but still substantial, savings occurred in earlier years. This shift from project-based to tenant-based assistance has not come at the expense of the recipients of housing assistance. The evidence indicates that recipients of tenant-based vouchers have typically occupied ­ better housing overall than occupants of housing projects (Orr et al., 2003; Olsen, 2008). Voucher recipients also benefit from a much wider range of choice of housing and neighborhood characteristics than families living in hous- ing projects and from the ability to move to another unit without losing their subsidy when a change in their circumstances lead to a change in their preferred housing type or location. During the 1990s, growing concerns about the mounting costs of public and private subsidized housing projects and the poor living conditions in some of them led HUD Secretary Cisneros to propose “vouchering out” almost all subsidized projects. This “reinvention” proposal would have provided vouchers to all households living in these developments, which   any M additional units were built after 1983 due to the long lags between the time that money is appropriated under these programs and the time that projects are completed.

156 REBUILDING THE RESEARCH CAPACITY AT HUD they could use to move elsewhere or to remain in their current units. Public housing authorities and private owners of the assisted stock would then operate more like conventional, private-market landlords, charging market rents and competing for tenants (some of whom would use vouchers to help pay the rent). The expectation was that some developments would not be viable under this plan and would be demolished and replaced, while others would be operated more effectively in order to attract and retain tenants. HUD relied heavily on the accumulated body of rigorous research to design this approach and to argue for its likely benefits. Specifically, evidence on the relative costs of alternative subsidy programs supported analysis show- ing that the reinvention proposal would not increase subsidy expenditures. And research on participation and benefits for voucher recipients supported the argument that this proposal would not disadvantage minorities or other vulnerable populations. Although the reinvention proposal was rejected by Congress, it led to a number of more incremental reforms to both public housing and vouchers, all of which were informed by research evidence. These included adjust- ments to the “take one, take all” requirement (which essentially required that a landlord who agreed to accept any voucher household had to accept an unlimited number), voucher lease requirements, and restrictions on security deposits. And public housing developments with high vacancy rates and high estimated modernization costs were targeted for demolition or conversion, and their residents received vouchers in place of project-based subsidies. In addition, the accumulated evidence about the effectiveness of v ­ ouchers led HUD to offer emergency vouchers to low-income households displaced after the Northridge earthquake in 1994. And research on pro- gram features that discouraged landlord participation informed the decision to waive some of the existing program regulations so that these emergency vouchers would be more widely accepted in the Los Angeles rental market. HUD also funded research on the Northridge emergency voucher program to provide guidance for dealing with future disasters. Housing Vouchers HUD’s Section 8 Housing Choice Voucher Program is the largest low- income housing program in the United States. It costs about $15 billion a year, accounts for almost one-half of HUD’s budget, and serves about 2 million of the poorest families in the country. Given its importance, it is hardly surprising that PD&R has done and funded a considerable amount of research on housing vouchers. This research has produced unusually reliable evidence on program effects, in part because some of the studies have been random-assignment experiments.

RESEARCH AND POLICY DEVELOPMENT 157 In its early years, PD&R oversaw the Experimental Housing Allowance Program (EHAP), the country’s largest social experiment. More recently, it has funded three other random-assignment voucher experiments: the Freestanding Housing Voucher demonstration, the Moving to Opportunity (MTO) for Fair Housing demonstration, and the Welfare to Work Voucher demonstration. These studies have estimated the effects of housing vouchers in comparison with no housing assistance and alternative types of housing assistance, and they compared the performance of different types of hous- ing vouchers. They have produced reliable information on a wide range of effects of housing vouchers that is of enduring value for housing policy development. Other studies done by PD&R staff and its contractors have shed light on the validity of widely expressed concerns that vouchers would have negative side effects or perform poorly in some circumstances or for some types of families. Experimental Housing Allowance Program No discussion of housing policy research would be complete without considering the first major experimental study, the EHAP. EHAP was the first study of tenant-based housing assistance. Congress authorized this program in 1970, planning for the experiment occurred in the early 1970s, data were collected during the mid-1970s, and the final reports were com- pleted in the late 1970s and early 1980s. The experiment cost almost $200 million (that is, more than $600 million in 2008 prices). Research and data collection accounted for almost one-half of this amount. The research firms that ran the experiments issued more than 300 reports, technical notes, and professional papers. As a result of these expenditures, more is known about the effects of these experimental programs than any established housing program. The two largest and most important components of EHAP were the supply experiment and the demand experiment. The primary purposes of the supply experiment were to determine the market effects of an entitle- ment program of household-based assistance, such as its effects on the rents of units with specified characteristics and how suppliers alter their units in response to the program. The experiment involved operating entitlement   ntroductions I to this vast literature can be found in the final reports of the supply exper- iment, the demand experiment, and the integrated analysis (Kennedy, 1980; Struyk and B ­ endick, 1981; Lowry, 1983); an edited volume containing summaries of the findings by the major contributors to EHAP research (Friedman and Weinberg, 1983); an edited volume containing evaluations of this research by outside scholars (Bradbury and Downs, 1981); a monograph containing some of the more technical results on consumer behavior from the demand experiment (Friedman and Weinberg, 1982) and HUD’s summary report (U.S. Depart- ment of Housing and Urban Development, 1980).

158 REBUILDING THE RESEARCH CAPACITY AT HUD housing allowance programs in the Green Bay (WI) and the South Bend (IN) metropolitan areas. Families were offered a cash grant on the condition that they occupy housing meeting certain standards. The demand experiment, conducted in the Pittsburgh (PA) and Phoenix (AZ) metropolitan areas, was primarily intended to see how recipients would respond to different types of household-based housing assistance and, for a given type, to different program parameters. The most influential demand experiment research went beyond a comparison of different types of household-based assistance: it compared the effects of the minimum-standards housing allowance pro- gram with the major established housing programs in existence at the time, namely, public housing, Section 236, and Section 23 programs. EHAP produced many results that have been influential in housing policy debates. One of the most important results of the supply experiment was that the entitlement housing voucher program tested had a minimal effect on the rents and prices of housing units with unchanging charac- teristics. This funding allayed concerns that the smaller nonentitlement Section 8 housing voucher program would have significant effects on the rents of unsubsidized units. Another important finding of the supply experiment was that the pro- gram induced a substantial increase in the supply of units meeting the program’s minimum standards. Despite the modest subsidies provided, the entitlement housing allowance program led to a 9 percent increase in the supply of apartments meeting minimum housing standards over its first 5 years. This increase resulted from upgrades to the existing stock of hous- ing (not from the production of new rental housing) and was entirely in response to tenant-based assistance that required families to live in apart- ments meeting the program’s standards in order to receive the subsidy. The most influential finding of the demand experiment was that r ­ ecipient-based assistance is more cost-effective than the types of project- based assistance that existed at the time. This result played an important role in persuading Congress to rely more heavily on tenant-based housing assistance to deliver housing subsidies. Freestanding Housing Voucher Demonstration The Section 8 Housing Certificate Program, enacted in 1974, differed from the EHAP model in several respects. Most notably, certificate recipients   is worth noting that the research tools and skills developed as part of EHAP laid the foun- It dation for much of the World Bank’s Housing Policy Research Program (see, e.g., Malpezzi and Mayo [1987], Renaud [1999], Malpezzi [2001], and Buckley and Kalarickal [2005]).   he subsidy amounted to about one-half of the taxpayer cost per household under the T Section 8 Housing Choice Voucher Program after accounting for inflation.

RESEARCH AND POLICY DEVELOPMENT 159 were limited to units renting for less than specified amounts (the fair ­market rent for units of a given size in a particular market). This requirement limited the choices for lower-income families and led housing authorities to press for higher fair market rents to accommodate families who wanted to occupy better units. In contrast, EHAP participants could choose units renting for more than the fair market rent, if they were willing to pay the difference from their own resources. The subsidy formula differed in other ways. In the certificate program, if the family occupied a unit that rented for the upper limit, it would receive the maximum subsidy. If the rent of its unit was less than the ceiling rent, its subsidy was reduced by the amount of the difference. EHAP voucher recipients could occupy units with rents greater than the certificate ceiling and receive the same subsidy, regardless of the rent of the unit. In addition, EHAP participants were more mobile, using their assistance to move across political boundaries to a greater extent than certificate holders. Drawing on the EHAP results, the President’s Housing Commission recommended the EHAP approach in preference to the certificate program (U.S. President’s Commission on Housing, 1982). In response to this recom- mendation, Congress enacted the Section 8 Housing Voucher Program in 1983 as an experiment, to be operated alongside the certificate program. At the same time, Congress mandated the Freestanding Housing Voucher demonstration. PD&R contracted with Abt Associates to conduct the dem- onstration as a random assignment experiment. This research showed that these different types of tenant-based assistance had very similar effects (Leger and Kennedy, 1988, 1990a, 1990b). It also showed that both types of housing assistance led recipients to live in substantially better housing and somewhat better neighborhoods, and both led to substantial reductions in recipients’ out-of-pocket housing expenditure and, hence, more money to spend on other goods. These findings eventually led policy makers to consolidate the two programs into a single program (the Housing Choice Voucher Program, enacted in 1998) that combined features of the older certificate and voucher programs. Voucher Success Rates One of the major policy concerns regarding the effectiveness of both certificates and vouchers has been the fact that a significant fraction of families offered certificates or vouchers do not use them. Another concern was that certain types of families might find it especially difficult to use them. In response, PD&R funded a series of studies on success rates in the certificate and voucher programs (Kennedy and Wallace, 1983; Kennedy and Finkel, 1994; Finkel and Buron, 2001). The Freestanding Housing

160 REBUILDING THE RESEARCH CAPACITY AT HUD Voucher demonstration also studied this issue as a part of a broader study of program performance. National success rates have varied somewhat over time. Over the past 25 years, the success rate has been under 70 percent in some years and over 80 percent in others (Finkel and Buron, 2001, pp. 2-3). The variation in success rates across areas is much greater. Finkel and Buron (2001) find success rates between 37 and 100 percent for the 48 housing authorities in their sample. These studies show that differences in success rates across housing authorities have many causes. For example, Kennedy and Finkel (1994) and Finkel and Buron (2001) find that among localities that are the same with respect to other factors, those with the lowest vacancy rates have the lowest success rates. Obviously, it is more difficult for anyone to locate a suitable unit when the vacancy rate is low; however, housing market tightness does not explain most of the variation in success rates. Success rates also vary with family characteristics and program p ­ arameters. For example, families who are eligible for larger subsidies due to lower incomes or higher payment standards have a higher success rate, presumably because they have a greater incentive to find a unit that meets the program’s standards. These studies indicate some differences in success rates for different types of individuals and households, but they allay many concerns about this matter. For example, after controlling for other factors, Finkel and Buron (2001) find no significant difference in success rates by the race, ethnicity, or gender of the head of the household or disability status of members of the household. In addition to informing the debate about the performance of the voucher program, the success rate studies have led to specific policy reforms. For example, the finding of the earliest study that larger families had lower success rates than smaller families that were the same with respect to other characteristics led HUD to increase fair market rents for larger units in the voucher program in 1983.   A local housing authority’s success rate is the percentage of the families authorized to search for a unit that occupy a unit meeting the program’s standards within the housing authority’s time limit and thus receive housing assistance from the voucher program. It is important to distinguish between the success rate and the participation rate of different types of families. Families with low voucher success rates can have high voucher participation rates if they are overrepresented among those offered vouchers. It is also important to distinguish between a housing authority’s success rate and its voucher utilization rate. Its utilization rate is the fraction of all vouchers in use during a period of time. By over issuing vouchers, housing authorities have been able to use all of their vouchers even though success rates were well below 100 percent.

RESEARCH AND POLICY DEVELOPMENT 161 Moving to Opportunity for Fair Housing Demonstration Over time, the location choices of voucher recipients, and their conse- quences for the recipients’ well-being, resulted in further program experi- mentation. In 1992, Congress authorized the MTO for Fair Housing demonstration, a major investment in research on the potential benefits to low-income families of receiving housing vouchers instead of public housing in high-poverty neighborhoods. The central goal of MTO was to assess the effects of using vouchers to help families move to much lower poverty neighborhoods. Like EHAP, MTO is a random assignment experiment that yields rigorous evidence of program impacts. Two types of vouchers were tested: one experimental group was offered regular Section 8 housing vouchers; the other experimental group was offered Section 8 vouchers plus mobility counseling assistance on the condition that the family moves to a neighborhood with a poverty rate less than 10 percent and lives there for at least 1 year (known as “special-purpose” vouchers). MTO has studied an unusually large range of program impacts— h ­ ousing and neighborhood characteristics, adults’ and childrens’ health, delinquency and risky behavior among youth, children’s educational out- comes, adult employment, earnings, income, and receipt of public assistance. Among other things, research to date has shown that families receiving s ­ pecial-purpose vouchers move to dramatically safer neighborhoods and experience significant improvements in physical and mental health, com- pared to similar families that remain in public housing or receive conven- tional vouchers (Orr et al., 2003). MTO has not only provided reliable evidence on a wide range of effects of several alternatives to current housing policies, but has also advanced general understanding about how neighborhood environments affect health, educational achievement, employment, and other outcomes. HUD recently launched the final phase of the MTO evaluation, and its results will inform the ongoing debate about how far to go in offering housing vouchers to occupants of the worst public housing projects. The HOPE VI Program has offered housing vouchers to tens of thousands of such families over the past 15 years, and the 1998 QHWRA mandated the vouchering out of additional public housing projects under certain circumstances and allowed it under others. Findings from MTO will also inform further policy discus- sions about whether assisted families should be encouraged to move to particular types of neighborhoods and how the neighborhood environment may affect the long-term well-being of low-income families.

162 REBUILDING THE RESEARCH CAPACITY AT HUD Welfare to Work Voucher Demonstration Later experimentation has concerned the ability to use vouchers to increase the family’s earnings and labor force participation. HUD’s W ­ elfare to Work Voucher demonstration has addressed this issue in par- ticular, while also analyzing a wide range of impacts on welfare fami- lies, again based on a random assignment research design. The findings indicate that, although receiving a voucher does not appear to increase employment or earnings, vouchers (as currently administered) do not create work disincentives as some critics have hypothesized. In addition, the research demonstrates that welfare families who receive vouchers enjoy better housing conditions, greater residential stability, increased disposable income (much of which appears to be spent on food), and a dramatically lower risk of homelessness (Mills et al., 2006). This is the most rigorous evidence to date on the impact of housing vouchers on the lives of low-income families. Voucher Clustering and Neighborhood Property Values HUD has conducted both external and in-house research about the pos- sible clustering of voucher recipients in poor and minority neighborhoods. In a number of neighborhoods around the country, there was significant local opposition to voucher recipients’ moving into a neighborhood. HUD contracted with Abt Associates to investigate eight of the areas in which there were controversies. The analysts found that neighborhoods with conflict were all places that had experienced economic decline (Churchill et al., 2001). The neighborhoods were not necessarily poor, nor the poorest sections of their cities or metropolitan areas. Some of the controversies were extremely local, as small as a few families on a single block. But in each case the controversy occurred in a declining neighborhood, and local residents either blamed the tenant-based programs for the decline or believed that the programs were exacerbating it. The study also reviewed the practices of the local program administrators and noted several common features that contributed to the controversies, including failure to respond to neighbor- hood complaints and failure to help families find housing in a broad range of neighborhoods. In general, the program administrators did not see them- selves as having any responsibility for neighborhood consequences. HUD subsequently conducted in-house research on the geographic distribution of vouchers across the country, as well as the geographic distribution of rental housing at below fair market rents (Devine et al., 2003). This research provides evidence that is central to the ongoing debate about whether voucher recipients can use their assistance to gain access to

RESEARCH AND POLICY DEVELOPMENT 163 affordable rental housing in decent neighborhoods and about the extent of geographic clustering. Among other things, this research finds that the stock of rental hous- ing in which vouchers can potentially be used is widely dispersed. Specifi- cally, in the 50 largest metropolitan areas, the voucher program applies to only about 2 percent of all occupied housing units and 6 percent of all rental units with rents below the applicable fair market rents. Virtually all census tracts contain at least some units with rental housing that is below fair market rents, and 83 percent have at least some voucher recipients living in them. Moreover, vouchers are generally not clustered geographi- cally. In 90 percent of all tracts with any voucher recipients, the program accounts for less than 5 percent of all households. But where vouchers are clustered, the clustering is in high-poverty, mostly minority central-city neighborhoods. Some critics of the voucher program have argued that vouchers under- mine neighborhood property values, especially when large numbers of voucher recipients cluster geographically. To address this question, HUD funded exploratory research on the property value effects of vouchers ( ­ Galster et al., 1999). This study found that a small number of recipients in a neighborhood (fewer than 6 to 8 families within a 500-foot radius) may have a positive effect, while a larger number of recipients concentrated in the same immediate neighborhood may have a negative effect. This finding plays a role in ongoing debates about the need to disperse voucher recipi- ents geographically and programmatic strategies for encouraging public housing authorities to monitor voucher locations and prevent excessive clustering. Community Development Block Grant Formula The interaction between research and policy has been qualitatively different on the “urban development” side of HUD, primarily because the nature of urban development programs changed just as PD&R was created. CDBGs were first put on the congressional agenda in 1971 as part of Nixon’s “new federalism” agenda (then called urban community development special revenue sharing) with Secretary Romney’s enthusiastic support. However, in 1973, at the recommendation of Secretary Romney, President Nixon suspended the urban renewal program at the same time   ationally, the share of tracts where voucher recipients account for more than 10 percent N of households is very small—only 3 percent of all tracts with any voucher recipients living in them. And voucher recipients account for more than a quarter of all households in less than 1 percent of all tracts. But in those tracts, the poverty rate averages 40.4 percent, compared to 19.5 percent where they account for less than 5 percent of households.

164 REBUILDING THE RESEARCH CAPACITY AT HUD that he suspended the major housing subsidy programs. Policy development during the National Housing Policy Review that year led to the recommen- dation that seven of the categorical programs be replaced by a single CDBG Program, which was enacted by Congress in 1974. (As noted above, the National Housing Policy Review began before PD&R was created, but the review staff included many individuals who became part of PD&R, includ- ing the first assistant secretary, Michael H. Moskow, and three of the four deputy assistant secretaries.) CDBG has been the largest and broadest urban development program ever since. The basic policy issue concerning CDBG has always been the formula by which appropriated funds are allocated between jurisdictions. PD&R pioneered much of the thinking on how to target funds to community development need and successfully developed the initial formula that was used to allocate funds in 1975 (Richardson, 2005). The formula included three variables: population (weighted at 25 percent), poverty (weighted at 50 percent), and overcrowded housing (weighted at 25 percent). Over- crowding was used because the Census Bureau had discontinued measuring overall housing quality as of the 1970 decennial census. Further in-house research conducted by PD&R in 1976 showed that the formula did not target funding particularly well toward cities generally considered to be “older and declining” (Bunce, 1976). The study suggested a dual formula, with the awarded allocations based on which formula would provide the higher level of funding. That is, if a jurisdiction would receive more funds under the alternative formula than under the original formula, its grant would be based on the alternative. The alternative for- mula included poverty (weighted at 25 percent), presence of old housing (weighted at 37.5 percent), and growth lag—the extent to which population growth in the jurisdiction lagged behind the growth in other recipient juris- dictions (weighted at 37.5 percent). The addition of growth lag and to a lesser extent old housing, resulted in larger allocations for many of the older cities of the Northeast and Midwest, compared to the original formula. This formula was included in the 1977 Housing and Community Development Act and used for the first time for the fiscal 1978 allocations. The dual formula system has continued since 1978, with no changes other than the substitution of new data from each decennial census. Over that period, however, PD&R researchers have identified a variety of prob- lems with the formula. For example, in the course of the policy develop- ment work on the enterprise zone eligibility formula during 1992, PD&R’s   ecause B under a dual formula system aggregate allocations would always exceed the total amount of CDBG funds appropriated by Congress in a given year, HUD has used a pro rata reduction to conform individual formula allocations with available resources. In fiscal 2002, for example, the pro rata reduction was 11.43 percent (Richardson, 2005, p. viii).

RESEARCH AND POLICY DEVELOPMENT 165 in-house work showed how the poverty variable results in overfunding of “college towns” relative to their per capita need because the Census Bureau does not count parental support as student income; it recommended subtracting college students for formula purposes (Neary and ­Richardson, 1995). PD&R conducted further research on the formulas, published in 1979, 1983, 1995, and 2005, with the purpose of offering alterna- tives that would better serve the stated purposes of CDBG (Bunce and G ­ oldberg, 1979; Bunce, Neal, and Gardner, 1983; Neary and Richardson, 1995; Richardson, 2005). Most recently, in 2006 PD&R contracted with E ­ conometrica, Inc., to develop innovative measures of community need and fiscal capacity that could be operationalized with annually updated Ameri- can Community Survey data, in an effort to develop the most efficient and fair way to distribute CDBG funds (Eggers, 2007b). These alternatives have never been adopted, probably because any change in the formula is a zero-sum game: some jurisdictions would gain funds, others would lose. Politically, the potential losers have always been able to prevent change. The CDBG formula research did affect policy, however; in 1990 Con- gress enacted a second block grant, the HOME Investment Partnerships Program. The statute did not specify a formula for HOME; it was left to HUD to establish the formula by regulation. The formula was developed by a team of PD&R analysts who had worked on the previous CDBG formula s ­ tudies, working with the leaders of both PD&R and the Office of Com- munity Planning and Development, which administered HOME. Drawing on the CDBG formula research, the HOME formula used the “number of families at or below the poverty level” rather than the number of individu- als, to avoid the CDBG formula issue of large allocations to college towns (24 CFR 92.50(c)(5)). This was an instance of cooperation between the policy and program offices, working under a tight deadline to put a new program in place. Fair Housing Since the early 1970s PD&R has produced a consistent and high- q ­ uality volume of research on the scope, nature, and consequences of racial and ethnic discrimination in housing and related markets, such as home mortgage lending and insurance. This research has come from in-house and external work, as well as from panels and papers by scholars organized by PD&R. In all three categories, PD&R efforts have produced substantively important, oft-cited, and sometimes methodologically path-breaking work. Illustrations follow in each category.

166 REBUILDING THE RESEARCH CAPACITY AT HUD In-House Research Probably the single most important piece of research on housing dis- crimination initiated and conducted by PD&R staff is the Housing Market Practices Survey (HMPS) (Wienk et al., 1979).10 This research represented the first effort to use paired testing to gain a nationally representative mea- sure of differential treatment discrimination against blacks as they sought rental and sales housing in metropolitan areas. Paired testing was a path- breaking approach insofar as it was the first attempt to directly measure discrimination in housing. Prior to this, social scientists had attempted to infer discrimination by examining differentials in housing costs, quality, or tenure that could not otherwise be explained by socioeconomic character- istics (e.g., Kain and Quigley, 1975; Galster, 1977; Yinger, 1977). Paired testing allows for a more controlled experiment, in which only the race or ethnicity of the tester teammates differs. This approach allows for differ- ences in treatment to be observed on a variety of outcome measures (cat- egorized by the PD&R researchers in terms of housing availability, courtesy, terms and conditions, information requested, information volunteered). Systematic patterns favoring one group or the other provide the evidence of discrimination. Based on tests in 40 metropolitan areas conducted in 1977, Wienk et al. (1979) concluded that a severe incidence of discrimination existed in rental or sales markets in virtually all areas, a decade after the passage of the Fair Housing Act. This finding provided important information for the debate that let to the strengthened Fair Housing Amendments Act in 1988. Following closely on the heels of the HMPS findings, PD&R spon- sored the Fair Housing Enforcement Demonstration, in cooperation with the National Committee Against Discrimination in Housing (U.S. Depart- ment of Housing and Urban Development, 1983). The demonstration was designed to investigate the degree to which a more formal and fiduciary connection between HUD and the then-emerging private, nonprofit fair housing groups would substantially enhance the effectiveness of fair hous- ing enforcement. Nine groups were selected to receive HUD funds for 2 years to receive and record complaints, conduct tests and other investiga- tions, refer complaints to HUD, and conduct tests to uncover discrimina- tion that may not have been triggered by bona fide complaints. The research found that HUD funds were leveraged at rates between 200 and 300 percent and that they led to a rationalizing and standardizing of all aspects of the fair housing enforcement portfolio of these groups. This demonstration established an evaluative foundation for legislation to 10  he T study also reviewed supplementary resources from external contractors, the National Committee Against Discrimination in Housing and George Schermer Associates.

RESEARCH AND POLICY DEVELOPMENT 167 establish ongoing federal support to private fair housing groups through the Fair Housing Initiatives Program of 1987. The HMPS proved to be the forerunner of two subsequent projects to garner a nationally representative estimate of housing market discrimina- tion against blacks and Hispanics, both undertaken by a team from the Urban Institute. (These studies are also discussed in Chapter 3.) Again, using the paired testing methodology, field work was undertaken in 25 metropolitan areas in 1989; the final report, the Housing Discrimination Study (HDS-1989), was published 2 years later (Turner, Struyk, and Yinger, 1991). This study involved significant advances in the statistical analysis of paired testing data to measure systematic discrimination. It also presented innovations in analyzing: (1) discrimination against not only black but also Hispanic home- and apartment-seekers; (2) variations in discriminatory behavior yielding inferences about motivations; and (3) steering, especially through the use of newly emerging spatial statistical techniques. Though few strict comparisons could be made with results from the 1977 HMPS due to some subtle differences in methods, the HDS-1989 continued to reveal very high incidences of discrimination against both minority groups when they sought apartments or homes. As discussed in Chapter 3, the 2000 replication of HDS (Turner, Ross, et al., 2002) found substantial declines in the incidence of discrimination against minority apartment seekers and against black home seekers, but not against Hispanic home seekers. It also found a worrisome increase in the rate of steering both black and Hispanic home seekers. HDS-2000 offered further advances in the statistical analysis of paired testing data, enhanced by the coincident National Research Council workshop report on the same topic (National Research Council, 2002a). Later phases conducted the first paired testing research into discrimination against Asians and Native Americans in selected sites (Turner and Ross, 2003a, 2003b). External Research In related program evaluation research, PD&R commissioned the Urban Institute to evaluate the Fair Housing Initiatives Program (FHIP), a private enforcement initiative testing program. FHIP was begun as a demonstration with the 1987 Housing and Community Development Act, with a segment of the effort devoted to funding private, nonprofit fair housing groups to conduct paired testing as part of an enhanced fair housing enforcement effort. The Urban Institute conducted: (1) key informant interviews with private and public agencies related to fair housing and with representatives of private-sector housing providers; (2) statistical analyses of complaints and investigations of 31 private fair housing groups; and (3) a legal analysis of evolving federal fair housing court decisions for 1968-1991. The research

168 REBUILDING THE RESEARCH CAPACITY AT HUD concluded that, while there was no template for tests that proved most effi- cacious in all circumstances, tests conducted objectively and professionally yielded the most credible evidence, and FHIP helped private fair housing groups undertake tests with such qualities (Wienk and Simonson, 1992). PD&R also took the lead in commissioning the Urban Institute to undertake two research projects involving new forms of paired testing in realms where this technique was in its infancy. In the first project, Wissoke, Zimmerman, and Galster (1998) developed an investigative method for assessing discrimination in home insurance against either individuals or neighborhoods. In this approach, a tester called insurance agents to request a quote about a particular (real) home “s/he was trying to buy (ficti- tiously)”; the closely matched teammate of a different race called the same agent to seek a quote on a closely matched (real) home in a neighborhood that differed only in the racial composition of the neighborhood. Different permutations of races of callers and of neighborhoods were used. In 1995, tests of differential treatment on the basis of Hispanic-white differences were undertaken in Phoenix; comparable black-white tests were undertaken in New York City. The tests revealed no statistically significant differences in treatment of either individuals or neighborhoods, except on a few minor indicators. The second project involved the process of shopping for a mortgage loan.11 Turner et al. (2002) developed protocols for testers posing as pro- spective first-time home buyers who approached lenders to inquire about procedures, available loan products, terms, personal qualifications for loans, etc. Black-white tests and Hispanic-white tests undertaken in Chicago and Los Angeles revealed many cases in which the white person was provided more assistance and “coaching” in the process. Recently, PD&R funded two matched studies of the general public’s awareness of fair housing laws, what constitutes illegal discrimination, and how violations of rights may be addressed (Abravenel and ­ Cunningham, 2002; Abravenel, 2006). The private, nonprofit Advertising Council con- ducted an extensive media campaign about recognizing and reporting dis- crimination between the baseline survey (2001-2002) and in the follow-up (2005). The studies both found vast majorities of people knew of and supported fair housing laws, and support grew during the period of study. However, far fewer people were sure of what constitutes illegal discrimina- tion in particular scenarios, and this did not improve during the period. Moreover, four of every five persons who claimed to have experienced discrimination did not pursue the matter by filing a complaint. The results implied that federal fair housing education efforts should be intensified to counter the public’s remaining misconceptions. 11  his T study was also supported by HUD’s Office of Fair Housing and Equal Opportunity.

RESEARCH AND POLICY DEVELOPMENT 169 Outside Panels and Papers The third source of PD&R-initiated research related to racial discrimi- nation and segregation consists of a series of commissioned conferences and subsequent edited volumes of papers that were organized by PR&R. Emblematic of this effort is the work of John Goering, a PD&R career employee, who for over two decades directed and conducted civil rights research and evaluation studies for PD&R. Collectively, this long-term body of work represents a comprehensive history of the best historical, political, sociological, statistical, and philosophical scholarship related to discrimination in housing. Much of this scholarship likely would not have been published, and certainly not as well integrated with complementary work, were it not for these efforts. In 1978, Goering assembled several emerging scholars in this field who had just published new quantitative analyses of housing discrimina- tion to write a report on the status of research on racial discrimination and segregation and establish a research agenda in this realm for HUD. This work was disseminated through PD&R’s HUD Occasional Papers in Housing and Community Affairs (Yinger et al., 1979). He then undertook an even more ambitious project, assembling 14 papers from a wide range of scholars, editing them, and adding synthesizing and overview chapters. The resultant Housing Desegregation and Federal Policy (Goering, 1986) proved an invaluable resource for thoughtful discussions of housing inte- gration, the quantitative connection between discrimination and segrega- tion, social and attitudinal factors affecting integration, and the role of federal policies and desegregation. By the early 1990s increasing attention was being paid in the civil rights and bank regulatory communities to the issue of discrimination in mortgage lending. In response, Goering organized a conference of scholars, advo- cates, lending institution regulators, and lending industry representatives to discuss the latest research on the topic. The resulting papers appeared in a massive coedited volume, Mortgage Lending, Racial Discrimination, and Federal Policy (Goering and Wienk, 1996). At the time it represented the single most comprehensive and cutting-edge set of discussions and analyses available on the issue. In anticipation of the 30th anniversary of the Fair Housing Act, Goering and an outside scholar, Gregory Squires, convened a panel of experts at the 1996 American Sociological Association meetings to reflect on what those three decades have meant. They supplemented this work with other papers designed to answer many provocative questions. Has the act achieved its goals and, if not, why? What might be needed to push ahead more effec- tively? The 15 papers that resulted were published in a special issue of PD&R’s Cityscape (Goering and Squires, 1999).

170 REBUILDING THE RESEARCH CAPACITY AT HUD Goering recently repeated the successful formula, convening with the support of PD&R in 2004 a multiday conference of scholars, advocates, regulators, and housing and lending industry representatives, the “National Fair Housing Research and Policy Forum.” Papers presented here were organized with additional material to create Fragile Rights Within Cities (Goering, 2007). The effects of the PD&R-initiated conferences and edited volumes on fair housing and lending are difficult to gauge. It seems accurate to suggest that this work provided for three decades a steady flow of high-quality, cur- rent, well-disseminated scholarship to inform policy makers and the public. The enthusiasm of the response appears to have been distinctly cyclical, with a clear intensification of federal fair housing and fair lending enforce- ment activities evinced during the 1990s. The most recent cycle of research has at this writing yielded little in the way of federal initiatives. In sum, the fair housing research undertaken under the auspices of PD&R has been substantial in both volume and scholarly quality over the long term. The cumulative record of paired testing (often called “audit- based”) research is especially noteworthy. This record was systematically evaluated by a panel of the National Research Council (2004, p. 7), which reached two conclusions: Nationwide field audit studies of racially based housing discrimination, such as those implemented by the U.S. Department of Housing and Urban Development in 1977, 1988, and 2000, provide valuable data and should be continued. . . . Because properly designed and executed field audit studies can provide an important and useful means of measuring discrimination in various domains, public and private funding agencies should explore appropriately designed experiments for this purpose. Despite these endorsements, PD&R has not funded paired testing in all contexts of transactions for which they would be appropriate. The three national paired-testing housing discrimination studies funded to date by PD&R have all involved in-person encounters between housing agents and testers. Yet, technologies such as voicemail and the Internet open up new domains for mediated, impersonal discrimination on the basis of naming conventions and linguistic style (see Smith and DeLair, 1999; Massey and Lundy, 2001; Bertrand and Mullainathan, 2004; Squires and Chadwick, 2006). A valuable next step for PD&R to advance research into discrimi- natory actions would be research on the initial contact phases of hous- ing transactions, which are usually over the phone and, increasingly, the Internet.

RESEARCH AND POLICY DEVELOPMENT 171 GSE Regulation Fannie Mae (FNMA) was created by Congress in 1968 as a government agency to create a secondary mortgage market. When HUD was established in 1965, FNMA became part of the department. In 1968 it was split, and the new FNMA became a private corporation to continue serving the sec- ondary market, with various privileges conferring “agency status.” HUD became the regulator of the new FNMA. HUD’s research on FNMA began about 20 years later with an extensive report on the corporation (U.S. Department of Housing and Urban Devel- opment, 1987). This report, mandated by the Secondary Mortgage Market Enhancement Act of 1984, was prepared by PD&R and remains a useful source of information on the early history of FNMA. It discussed possible options for eventual privatization, a policy goal of the administration at the time. More relevant to subsequent policy making, it contained the first inde- pendent analysis of Fannie Mae’s net worth on a “mark-to-market” basis, measuring the value of both assets and liabilities at current market prices rather than historical cost. The report concluded that Fannie Mae had a positive net worth, but that its net worth had been negative every year from 1978 to 1984, a period of large and unexpected interest rate fluctuations, and that it continued to incur substantial risk from future fluctuations. In 1989 the Financial Institutions Reform, Recovery and Enforcement Act converted Freddie Mac (FHLMC) to a similar private entity, vesting regulatory authority in HUD, and requiring annual reports on both FNMA and FHLMC (which became known collectively as government sponsored enterprises [GSEs]). These reports also were prepared by PD&R, for 1989, 1990, and 1991, and a financial institutions regulatory staff was created in PD&R to assist the secretary of HUD in fulfilling the financial regulatory responsibilities. Written in the aftermath of the savings and loan industry collapse and drawing on the methodology developed in the 1986 report, these reports analyzed the adequacy of GSE capital, and contained “stress tests” of the ability to withstand serious economic disturbances. The reports concluded that the GSEs would not be able to earn an AAA rating from the credit rating agencies in the absence of their implicit government guarantee (i.e., would not be able to survive 10 years in the event of a major depression), but that Fannie Mae had improved its financial position so that it was protected against interest rate fluctuations such as those during the early 1980s, which at that time left Fannie Mae with negative net worth. The first 2 years’ reports were significant contributions to the legislative process during 1990-1992, which culminated in the Federal Housing Enterprises F ­ inancial Safety and Soundness Act of 1992 (FHEFSSA), although the

172 REBUILDING THE RESEARCH CAPACITY AT HUD legislation as enacted did not use the same approach in setting capital standards. FHEFSSA split the regulatory authority between HUD and a new Office of Federal Housing Enterprise Oversight (OFHEO). HUD retained “mis- sion regulation” authority and general regulatory authority; OFHEO was responsible for financial safety and soundness regulation. Affordable Housing In particular, HUD was responsible for establishing and enforcing annual “affordable housing goals.” The provisions of FHEFSSA establish- ing the goals were influenced by the PD&R analysis of the GSEs’ loans. Three such goals were created in FHEFSSA, each as a percentage of the mortgage business of each GSE. HUD was required to establish the quan- titative level of each goal and change the levels every few years in response to GSE activity and market conditions. Under the FNMA Charter Act, the HUD secretary had authority to require that a reasonable portion of FNMA’s mortgages purchases serve low- and moderate-income families. Interest in FNMA’s affordable housing activities dates back to 1978 when Secretary Harris issued regulations requiring FNMA to devote 30 percent of its purchases to homes with prices at or below 2.5 times the median family income for the local market area. At that time, FNMA did not have data on home buyers’ income for the mortgages it purchased, so the standard in terms of median income was chosen as a reasonable proxy. By 1990, however, the GSEs did have such data, and analysis by PD&R showed that the 1978 criterion was not a very good proxy: large numbers of borrowers in the lower half of the income distribution bought homes with prices above 2.5 times the local median income; similarly large numbers of borrowers in the upper half of the income distribution bought homes with prices below 2.5 times the local median. This data analysis contributed to Congress’ decision to establish explicit affordable housing goals and to establish them largely in terms of the income of the home buyer or renter. Until the passage of the Housing and Economic Recovery Act of 2008, HUD has had the responsibility to analyze the mortgage and housing markets and GSE activity, and based on these analyses to set new goals by regulation every few years. More than a dozen working papers have been prepared by PD&R staff, primarily in the Office of Economic Affairs, documenting the extent to which the GSEs fund affordable loans, loans to minority home buyers, loans to first-time home buyers, and loans for multi­ family housing projects. PD&R has also funded some external research, including studies of the ultimate effect of GSE purchases on lower-income households and on lower-income neighborhoods.

RESEARCH AND POLICY DEVELOPMENT 173 A particular concern that has been raised as a result of PD&R’s analy- ses is the extent to which the GSEs purchase the loans of first-time home buyers. Analysis of GSE data, combined with market data from several sources, showed that the GSEs were serving disproportionately few first- time buyers, particularly minority first-time buyers, relative to the share of these households in the conventional mortgage market. These analyses contributed to the regulatory decision to establish subgoals for first-time buyers as shares of the various affordable housing goals as of 2005. All these reports have contributed to the establishment of new afford- able housing goal levels in 1996, 2000, and 2004, in each case becoming effective at the beginning of the following calendar year. They are incor- porated into the economic and regulatory analyses required as part of the rule-making process. These analyses are also prepared by PD&R’s Office of Economic Affairs. They are necessarily voluminous and required extensive and intensive staff effort. Legislation to revise the regulatory structure and authority, which ulti- mately resulted in the Housing and Economic Recovery Act, was introduced in Congress since 2003. Among the issues being considered are changes in the affordable housing goals. Legislation under consideration in 2005-2006 included a new goal for first-time home buyers. PD&R staff also prepared a substantial analysis of the goals established in the major bill introduced into the House of Representatives in 2007; as a result, the proposed legisla- tion was changed to incorporate single-family rental housing in the goals, and the GSEs’ “ability to lead the market” was added as a factor to be considered in establishing specific single-family housing goals. The Housing and Economic Recovery Act of 2008 contains a requirement for the GSEs to report on the number of rental units affordable to low-income families contained in mortgage purchases of 2-4 unit owner-occupied properties, and permits the new regulator to establish requirements for such units. HUD’s regulatory activity does not figure noticeably in the HUD budget, but in many respects is as important to the quality and affordability of U.S. housing as its programs. GSE regulation was a prime example of PD&R’s ability to strengthen and complement the capabilities of program offices while adding an important measure of independence and objectivity to the analysis and the development of regulations. Over the past 15 years, GSE regulation has been conducted by a consortium of several HUD offices, including the Office of Housing, the Office of the General Counsel, and the Office of Fair Housing and Equal Opportunity, as well as the Office of Policy Development and Research, under the leadership of the assistant secretary for housing. GSE regulation was funded from the overall HUD budget, par- ticularly the PD&R appropriation. This is unlike the regulators of all other financial institutions, and unlike HUD’s regulation of the manufactured housing industry, which are funded by fees levied on the regulated entities.

174 REBUILDING THE RESEARCH CAPACITY AT HUD The role of PD&R has been extremely important, and its importance has been recognized by the other offices and the secretary. GSE regulation, and regulatory legislation has been influenced by the PD&R analyses of GSE activity since the 1980s. Subprime Mortgages PD&R’s research on subprime lending started in the mid-1990s, when subprime loans were a very small share of the mortgage market and largely unknown to policy makers or the public. In 1994 PD&R developed a list of subprime lenders, based on data collected under the Home Mortgage Disclosure Act (HMDA), trade publications, and industry sources. Not all subprime loans were originated by specialized subprime lenders, but the extent of market segmentation was pronounced, and identifying subprime lenders was an essential first step toward describing and analyzing the market for subprime mortgages. The PD&R list became a standard refer- ence for tracking subprime lending through HMDA data, used by Federal Reserve Board analysts and advocacy groups, among others. By 1998 the list included 200 lenders who specialized in subprime loans (Scheesele, 1998a). By about 2002, the list was becoming less useful as prime lenders began making more subprime loans, as PD&R pointed out whenever the list was updated, but it was still the most authoritative source. PD&R used the list for two purposes: (1) to create information and facilitate analysis of the subprime mortgage market, and (2) to help in establishing the statutory “affordable housing goals” for Fannie Mae and Freddie Mac. Both purposes resulted in policy initiatives. Analyzing the Subprime Market PD&R used the list to document the rapid growth of subprime lending during the 1990s and to explore the degree of concentration of subprime mortgages, both geographically and demographically. Studies found that minority borrowers in low-income neighborhoods were disproportionately likely to be subprime borrowers (Scheesele, 1998b) and that subprime loans were more common in low-income than in high-income neighborhoods, more common in black than in white neighborhoods, and indeed more common in high-income black neighborhoods than in low-income white neighborhoods (U.S. Department of Housing and Urban Development, 2000d). These patterns were attributed to the absence of prime lenders in those neighborhoods. The later study included a brief but vigorous national summary statement of the problems and dangers of subprime lending and was followed by separate, more detailed, studies of five large metropolitan areas. These studies attributed the patterns to the absence of prime lenders

RESEARCH AND POLICY DEVELOPMENT 175 in such neighborhoods, and they also warned about high foreclosure rates and a rise in predatory lending practices. PD&R continued to conduct research on predatory lending in the grow- ing subprime market, particularly for minority households and in minority neighborhoods (Bunce et al., 2000; Fishbein and Bunce, 2000; Scheesele, 2002). The studies were undertaken by the Office of Economic Affairs, sometimes using proprietary data (such as the Loan Performance data base), as well as FHA and HMDA data; they exemplify the internal research conducted by PD&R. In addition, several private organizations, including the Center for Community Change (Bradford, 2002), the National Commu- nity Reinvestment Coalition (2007, 2008), and the Consumer Federation of America (Fishbein and Woodall, 2006), began using the same methodology to describe subprime lending patterns by metropolitan area. More recently, PD&R has extended this line of research to investigate the differences in risk characteristics between subprime, FHA, and prime mortgages, and the extent to which these markets overlap (Rodda, Schmidt, and Patrabansh, 2005). It has also funded a large study of the reasons why households choose to make use of subprime mortgages and other uncon- ventional financial services, such as payday loans (Apgar and Herbert, 2006). The assistant secretary for policy development and research also convened a conference in 2006 to obtain information about recent changes in subprime mortgage instruments and a second conference in 2007. This body of research contributed substantially to the work of a joint HUD-Treasury task force on predatory lending and subsequent policy ini- tiatives. PD&R staff also provided most of the HUD staffing for the task force, including the research leader, and played a major role in preparing the report (U.S. Department of Housing and Urban Development and U.S. Department of the Treasury, 2000). The report of the task force offered a series of recommendations for legislative and regulatory actions to reduce predatory lending practices, while at the same time maintaining access to mortgages loans by lower income borrowers. Neither HUD nor Treasury has regulatory authority over most mort- gages or mortgage lenders.12 The recommendations of the task force were 12  Federally chartered financial institutions are regulated by the Federal Reserve Board, the Federal Deposit Insurance Corporation (FDIC), and the Office of the Comptroller of the Cur- rency; state-chartered institutions are regulated by the FDIC as well as the states; and mortgage brokers are regulated by the states. The Office of Thrift Supervision within Treasury regulates community banks; HUD regulates lenders insofar as they make FHA loans but does not have authority over loans not insured by FHA. HUD also has regulatory authority over the GSEs, as discussed above, and HUD regulates real estate settlement procedures for all loans, under the Real Estate Settlement Procedures Act (RESPA). But it has no role in HMDA data collection, for example; by statute, HMDA is the province of the Federal Financial Institutions Examina- tion Council (FFIEC), consisting of the five financial regulatory agencies.

176 REBUILDING THE RESEARCH CAPACITY AT HUD thus directed largely at Congress and the Federal Reserve Board. Specific recommendations for HUD action included additional funding for hous- ing counseling; prohibition of “flipping” with FHA-insured loans (flipping occurs when a home is sold twice within a short period of time, with the second sale at a markedly higher price than the first); closer monitoring of appraisers and mortgage brokers; and clearer, earlier, and binding disclo- sures of settlement costs under RESPA regulations. Action on each of these issues occurred over the next few years. Beginning in 2002, the HUD budget requested a substantial increase in housing counseling funds, and these requests were approved by Congress. Funding tripled over the next several years, from $15 million in 1999 to $20 million in 2001 and $45 million by 2005. In addition, Congress has held several hearings on housing counseling. The Housing and Economic Recovery Act of 2008 (P.L. 110-289) further increased funding for counsel- ing and also required HUD to conduct a demonstration counseling program for home buyers with low down payments. Based on this interest, PD&R initiated an evaluation of HUD-approved counseling agencies in September 2007 (mentioned in Chapter 3). The evaluation, which is being conducted by Abt Associates, will describe the current state of the counseling industry and then evaluate the effectiveness of pre-purchase counseling in forestall- ing mortgage default, including both an analysis of the counseling services received by currently delinquent homeowners and a controlled experiment of future home buyers as part of the evaluation. To date, the analysts have collected information through interviews with counseling agencies and other interested entities. FHA established an anti-flipping rule by 2002, denying FHA mortgage insurance on loans when the home had been sold twice within 6 months unless there was evidence of substantial repairs and rehabilitation. PD&R staff provided analysis of local market data that helped to determine the cutoff dates for mortgage insurance eligibility, to minimize the extent to which legitimate resales were inadvertently denied insurance and predatory “flips” were inadvertently eligible. FHA also began to track the performance of loans by appraiser and release information about default rates on these loans, through a pro- gram known as “Appraiser Watch.” This went into effect in 2002. FHA subsequently instituted rules to improve the monitoring of mortgage brokers. HUD issued a proposed RESPA rule in 2002, improving disclosure along the lines recommended by the task force. The rule was not adopted, and the proposal was withdrawn in 2004, owing partly to strong bipartisan congressional opposition and criticism from several industries that provide settlement services, such as title insurers, appraisers, and realtors. HUD proposed a revised rule in 2008.

RESEARCH AND POLICY DEVELOPMENT 177 In each case the recommendation of the Task Force on Predatory Lend- ing was not the sole impetus for the reform; other factors also contributed. Research by Freddie Mac found evidence that pre-purchase counseling reduced mortgage defaults (Hirad and Zorn, 2002) and was cited by HUD to support its budget proposals; the anti-flipping rule, the Appraiser Watch Program, and broker monitoring were strongly advocated by the Baltimore Predatory Lending Task Force, an organization of local public officials and community groups; RESPA reform became a major initiative of HUD and the administration after a federal appellate court ruling in early 2001 appeared to prohibit common lending practices. The task force recom- mendations, however, were an important contributing factor. The task force also led to a roundtable on predatory lending in 2001, cochaired by assistant secretaries from Treasury and HUD and a member of the Board of Governors of the Federal Reserve, bringing together consumer advocates and industry representatives for vigorous discussion. Setting GSE Affordable Housing Goals The HUD list of subprime lenders has also been used for more than a decade to establish the affordable housing goals for Fannie Mae and Freddie Mac. PD&R used the list to analyze the GSE role in the subprime market in a regular series of reports on GSE affordable lending, beginning in the mid-1990s (Bunce and Scheesele, 1996). This analysis led HUD to define the GSEs’ “market” to include loans classified as “Alt-A” or “A-minus”—the highest quality of subprime loans—as well as conventional conforming “prime” or “A” loans, in the affordable housing goals, begin- ning in 1996. FHEFSSA requires Fannie Mae and Freddie Mac to “lead the market,” which in turn necessitates a definition of the market that the GSEs are expected to serve. HUD has continued to study subprime lending in the context of the GSE affordable housing goals in its series of working papers in housing finance. In addition, its responsibility for “mission” regulation of Fannie Mae and Freddie Mac led to further research on subprime lending in local markets. A PD&R-funded research project at the Urban Institute docu- mented the growing GSE interest in subprime mortgages, to some extent in response to the affordable housing goals established by HUD in 2000, and discussed the extent to which the GSEs were lowering costs for sub- prime borrowers and the extent to which the GSEs were taking higher risks ( ­ Temkin, Johnson, and Levy, 2002). The subprime mortgage market is an instance in which PD&R research over a long period of time has contributed to changes in HUD programs and regulations. In addition, and perhaps more importantly, it has provided information about the broader mortgage market for which HUD has no

178 REBUILDING THE RESEARCH CAPACITY AT HUD statutory authority. Not all the policy recommendations arising from the research were adopted; policy makers and the public gave less attention to predatory lending during the housing boom in the mid-2000s, when house prices were rising rapidly, than before or since. For example, an administra- tion proposal contained in each year’s budget during 2003-2005, to allow FHA to compete with subprime lenders, was not approved by Congress. Conclusion The relationship between research and policy development is complex and ongoing. The examples in this chapter are not intended as a compre- hensive list, but as illustrative. At the same time, they necessarily involve some overlap with the discussions of research projects and policy develop- ment activities in the preceding three chapters. The long-term relationship between research and policy development has been fruitful and valuable for both activities. Looking at the allocation of the PD&R budget for a given year, or even over a few years, does not show the policy relevance of PD&R’s research activities, especially because some of the external research projects are multiple year studies, appearing in the PD&R budget only in the year in which they are funded. As discussed in the next chapter, research budget reductions in recent years are likely to hamper informed policy development in the future.

Next: 10 Loss of Capacity and Its Consequences »
Rebuilding the Research Capacity at HUD Get This Book
×
Buy Paperback | $69.00 Buy Ebook | $54.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Today, the nation faces an array of housing and urban policy challenges. No federal department other than HUD focuses explicitly on the well-being of urban places or on the spatial relationships among people and economic activities in urban areas. If HUD, Congress, mayors, and other policy makers are to respond effectively to urban issues, they need a much more robust and effective Office of Policy Development and Research (PD&R).

PD&R conducts independent research and program evaluation, funds data collection and research by outside organizations, and provides policy advice to the Secretary and to other offices in HUD. Most of PD&R's work is of high quality, relevant, timely, and useful. With adequate resources, PD&R could lead the nation's ongoing process of learning, debate, and experimentation about critical housing and urban development challenges.

Rebuilding the Research Capacity at HUD makes seven major recommendations about PD&R's resources and responsibilities, including more active engagement with policy makers, formalizing various informal practices, strengthening surveys and data sets, and more. Acknowledging that the current level of funding for PD&R is inadequate, the book also makes several additional recommendations to help enable PD&R to reach its full potential.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!