«Moving to opportunity voluMe 14, nuMber 2 • 2012 U.S. Department of Housing and Urban Development | Office of Policy Development and Research ...»
194 Moving to Opportunity Increasing the Value of MTO Research for Housing Policy Development Edgar O. Olsen University of Virginia Abstract The Moving to Opportunity (MTO) for Fair Housing demonstration has estimated the effects of two concrete reforms of low-income housing policy on one important group those reforms affected. Reforms of this type have received, and will continue to receive, serious consideration in housing policy debates. At this stage, estimating the effects of the MTO reforms on all the people they affected significantly is not feasible. Estimating the effects of two similar reforms that would have almost the same effect on the families studied in MTO, however, is feasible and desirable. These reforms would have no effect on the number of families who receive housing assistance. Instead, they would affect the nature of the housing assistance offered and the taxpayer cost of providing the assistance. One alternative reform would almost surely have generated cost savings and additional revenue exceeding the cost of the vouchers, thereby providing greater benefits than the current system at a lower taxpayer cost. The MTO results supplemented with estimates of the taxpayer cost of either reform would provide a reasonably comprehensive analysis of its effects. This article suggests how to estimate the taxpayer costs of the alternative reforms. Doing so would significantly increase the value of MTO research for housing policy development.
Introduction The primary purpose of the Moving to Opportunity (MTO) for Fair Housing demonstration was to study the effects of better neighborhoods on a wide range of individual outcomes. However, it also produced results that are relevant for assessing the likely effects of major reforms of low-income housing policy. Literally, the final impacts evaluation (Sanbonmatsu et al., 2011) reported the effects of offering housing vouchers to families with children living in subsidized housing projects in census tracts with high poverty rates as opposed to the effects of alternatives that would have been available to these families in the absence of this offer.
Cityscape 195 Cityscape: A Journal of Policy Development and Research • Volume 14, Number 2 • 2012 U.S. Department of Housing and Urban Development • Office of Policy Development and Research Olsen Political leaders have proposed similar reforms, and some have been implemented. The Clinton Administration proposed comprehensive legislation for phasing out project-based assistance (HUD,
1995) and, in his campaign against President Clinton, Robert Dole also proposed vouchering out public housing. Although the Clinton proposals were not adopted, the Quality Housing and Work Responsibility Act of 1998 (Title V, Public Law 105-276) mandated demolishing public housing projects and providing housing vouchers to their residents under certain circumstances and allowed these actions under other circumstances. Furthermore, HOPE VI grants and other sources have funded the redevelopment of many public housing projects in census tracts with high poverty rates, and the families whose projects have been demolished have almost always been offered housing vouchers.
Although MTO has produced evidence on important effects of its reforms on members of the families offered vouchers, this evidence is inadequate to judge the desirability of the reforms because MTO also affected taxpayers and the families who replaced MTO voucher recipients in subsidized housing. To maximize the benefit from MTO’s enormous investment in data collection, it is important to supplement the results in the final impacts evaluation with evidence on the effects on other individuals.
This article argues that, at this stage, producing the information needed to estimate the effects on all groups of the exact reforms MTO implemented is not feasible, but that the MTO results are applicable for assessing several similar reforms whose primary effects on other people could be estimated with reasonable effort. MTO provided substantial benefits to the families who moved into the public housing units vacated by those who accepted an MTO voucher, but it did not collect information relevant for assessing the effects of the reform on this group, and it is not feasible to assemble it now. It is feasible, however, to estimate the effects of two similar reforms that would have almost the same effect on the families studied in MTO.
The alternative reforms would have made the same voucher offer to the same families as MTO.
They would have differed primarily regarding what was done with the units vacated by the public housing tenants who accepted the offered vouchers. Under one option, the units vacated by families with MTO vouchers would have been rented at market rates. Under the other option, buildings with a number of units equal to the number vacated by MTO voucher recipients, plus a fraction of the previously vacant units in the projects involved, would have been demolished and the land sold to the highest bidder. Under both alternatives, when the initial recipients give up their vouchers, those vouchers would be offered to occupants of subsidized housing projects in high-poverty census tracts. Therefore, unlike MTO, these reforms would have had no effect on the number of families who received housing assistance. They would instead have affected the nature of the housing assistance offered and the taxpayer cost of providing it. Evidence on the cost-effectiveness of different types of housing assistance strongly suggests that the second reform would have substantially reduced taxpayer cost and hence would have benefited both voucher recipients and taxpayers.
MTO research focuses heavily on the effects of the experimental group treatment that required voucher recipients to live for the first year in a census tract with a 1990 poverty rate of less than 10 percent. Because no Congress is at all likely to change the voucher program to limit participation to families willing and able to live in a low-poverty neighborhood, the MTO results on the effects of offering traditional Section 8 housing vouchers are more important for an analysis of the reforms under consideration. The points made in this article, however, are equally applicable to the lowpoverty voucher option.
196 Moving to Opportunity Increasing the Value of MTO Research for Housing Policy Development MTO Housing Policy Reforms For purposes of this article, certain features of the MTO reforms are especially important. Under MTO, tenants from the public housing waiting list filled the public housing units that MTO voucher recipients vacated. Public housing authorities (PHAs) received the same operating and capital subsidies and HOPE VI grants to redevelop some of their projects that they would have received in the absence of MTO. Therefore, MTO did not affect the number of families in public housing units or their taxpayer cost. Indeed, it did not affect the budget of any other low-income housing program. It provided additional subsidies to serve the families who accepted the MTO voucher offer to leave their public housing units.
Feasibility of Comprehensive Analysis of MTO Housing Policy Reforms The MTO final impacts evaluation indicated many beneficial effects and a few unintended negative consequences of the two reforms for the individuals in families who were offered and used MTO vouchers. Because it was not designed to provide a comprehensive analysis of the reforms, the report is silent on the reforms’ effects on two other groups—namely, taxpayers and the additional families who received the housing assistance that would have gone to MTO voucher recipients in the absence of the demonstration.
Because MTO provided additional subsidies to serve the families who accepted the voucher offer, the reforms had a cost to taxpayers. Using MTO administrative data to determine this cost in every year would surely be possible. The additional taxpayer cost of assisting the Section 8 group was the cost of providing its members with MTO vouchers. For the experimental group, the cost of mobility counseling and search assistance would be added to the cost of their vouchers. After initial lease up, some voucher recipients relinquished their vouchers to move to a publicly or privately subsidized project. Because MTO did not affect the budgets of existing low-income housing programs, however, the assistance that treatment families received from these programs is not an additional taxpayer cost.1 It is very important to recognize that the additional taxpayer cost provided benefits not only to families who accepted MTO vouchers but also to the additional families who replaced those with MTO vouchers in existing programs. MTO did not collect information relevant for assessing the effects of the reform on the latter group, and this omission is important from the viewpoint of assessing the consequences of the reforms. Immediately after the initial lease up, these additional families were about as numerous as the families who used the offered vouchers. They replaced the public housing tenants who used MTO vouchers. Furthermore, the benefits to these additional families could easily have been larger than the benefits to those who used MTO vouchers, because almost all of them came from unsubsidized housing. Initially, the net benefit to the families who Taxpayers who care about the voucher recipients also benefit from the reforms. Because of our inability to estimate these altruistic benefits, comprehensive empirical analyses of the benefits and costs of the reforms are not possible. These benefits should not be forgotten, however, in assessing the desirability of the reforms.
replaced MTO voucher recipients in public housing may have been modest, because they moved to public housing projects in bad neighborhoods and often in poor condition. The immediate benefit may have resulted primarily from spending less on housing and hence consuming more of other goods. A substantial fraction of these families ultimately received larger benefits, however. About 42 percent of units in MTO housing projects were demolished under HOPE VI or some other redevelopment initiative before data collection for the final impacts evaluation. Their occupants were offered the choice between a housing voucher and a vacant unit in another (likely better) public housing project. By the time of final data collection, about one-fourth of the members of the control group (to whom the MTO demonstration did not offer vouchers) had housing vouchers. By this time, about the same fraction of the families who replaced the MTO voucher recipients at the outset of the experiment almost surely had them. Because MTO did not collect the relevant data, it is not possible at this stage to estimate the effects of the MTO reforms on the many additional families who received housing assistance in existing programs as a result of the MTO reforms.
Alternative Reforms Although it is not possible to conduct a comprehensive analysis of the MTO reforms implemented, it is feasible to produce a broader analysis of the effects of two similar reforms that would have yielded almost the same outcomes for members of the experimental and Section 8 groups. The MTO results would be part of a comprehensive analysis of these reforms.
Unlike the MTO reforms, both alternative reforms are designed to serve the same number of families as the current system. They would have made the same offer to the same families living in the same housing projects. Other subsidized households, however, would not have filled the units vacated by families who used the offered voucher. Under the first alternative reform, households paying market rents would have occupied those units. The second reform would have demolished buildings in subsidized projects with a number of units equal to the number vacated by MTO voucher recipients plus a fraction of the vacant units in the projects involved, sold the land to the highest bidder, and reduced the public housing budget by the amount that would have been spent on these units with a continuation of the current system.
The public housing parcels sold under the second alternative would have been a subset of those that were redeveloped with HOPE VI grants and funding from other sources during the course of MTO. In essence, selling these parcels would be an alternative use of some, but not all, of the properties that were redeveloped during MTO. As mentioned previously, about 42 percent of public housing units involved in MTO were demolished and replaced with new housing before data collection for the final impacts evaluation. The MTO results suggest that if all public housing tenants living in projects in high-poverty neighborhoods had been offered traditional Section 8 vouchers, only 15 percent of the families would have used them—about 25 percent of eligible families enrolled in MTO and 61 percent of enrollees in the Section 8 group used the offered vouchers. Therefore, the number of units that would have been sold under the second alternative is much less than the number redeveloped during MTO.
198 Moving to Opportunity Increasing the Value of MTO Research for Housing Policy Development Under the second alternative, the public housing properties would have been sold at the earliest reasonable time after the departure of families with vouchers. In almost all cases, this would have happened earlier than redevelopment occurred in the MTO projects. This alternative would have avoided expenditures on operating these units under the status quo until their redevelopment and generated the revenue from selling them sooner. It would also have avoided leaving vacant for long periods many additional units in public housing projects that would eventually be redeveloped.
The accelerated demolition would almost surely have benefited some and harmed other residents of these housing projects. Because it affects only the timing of the demolition of their units, however, the magnitudes of these benefits and costs arguably would be modest.