«Moving to opportunity voluMe 14, nuMber 2 • 2012 U.S. Department of Housing and Urban Development | Office of Policy Development and Research ...»
Because so many units in the housing projects redeveloped under HOPE VI were not available for occupancy due to their failure to meet the program’s minimum housing standards, it would have been possible to demolish buildings with many more units than the number of families that accepted MTO vouchers without reducing the number of assisted households. Popkin et al. (2004) reported that about one-third of the units in projects redeveloped under HOPE VI were vacant at the time of the grant award. Demolishing buildings with more units than the number occupied by voucher recipients and selling the land would have generated additional revenue to reduce the taxpayer cost of assisting the same number of families.
To continue to assist the same number of households over time under either alternative reform, something must be done to replace the initial voucher recipients who give up their vouchers.
By the time of data collection for the final impacts evaluation, only 45 percent of households in the Section 8 group continued to receive vouchers. To offset attrition among the initial voucher recipients, the alternative reforms would have offered vouchers to households that moved into the same housing projects after the initial voucher recipients departed or who lived in housing projects in census tracts with only slightly lower poverty rates. Those households would have benefited and taxpayers would have incurred additional costs on their behalf. We could use the MTO data to estimate these magnitudes because the recipient benefits and taxpayer costs should be about the same as for initial voucher recipients with the same observed characteristics.
Both alternative reforms would have generated additional revenue to offset the cost of the vouchers.
For the first alternative, the additional revenue would have come from the new public housing tenants who pay market rents. For the second, it would have come from selling public housing land.
The second alternative also would have led to considerable cost savings. It would have avoided the cost of operating some public housing projects before their redevelopment and the considerable construction cost of HOPE VI redevelopment.2 Indeed, as mentioned in the final impacts evaluation and discussed in more detail later in this article, the second alternative would almost certainly have generated cost savings and additional revenue exceeding the cost of the vouchers. In this case, the cost of providing the benefits to families who used vouchers would have been negative; that is, the reform would have provided greater benefits than the current system at a lower taxpayer cost.
Turner et al. (2007) reported a construction cost of $160,400 per dwelling unit for the 192 HOPE VI projects initiated before 2004.
Applicability of the MTO Results to the Alternative Reforms The alternative reforms might have had somewhat different effects on the studied outcomes than the MTO reforms. Because they would have affected the desirability of remaining in public housing to some extent, the alternative reforms would have affected treatment group members’ decisions about whether to remain in public housing and, hence, the MTO results would be less applicable.
This section argues that these differences likely would have been very small.
Under the MTO reform, families from the top of the public housing waiting list initially occupied units vacated by families with MTO vouchers. Under the first alternative reform, they would have been rented at market rents, that is, to the households willing to pay the most for them.
Therefore, the alternative reform would have led to a somewhat different set of neighbors for each subsidized family in public housing than the MTO reform. However, because the vacated units are in bad neighborhoods and usually in bad condition, the families willing to pay the most to live in them would surely have had very low incomes, like the families on public housing waiting lists.
Furthermore, only 15 percent of the units would be rented at market rents. The differences in the characteristics of the families who would occupy public housing units under MTO and under the first alternative reform are likely to be modest. They would have little effect on families’ decisions about whether to use the offered voucher.
In deciding whether to use the MTO voucher, a farsighted eligible person would consider the future housing assistance that would be available later if he or she remained in public housing.
This consideration, however, would not lead to different decisions under the alternative reforms compared with under MTO because the future options are about the same under all reforms. For example, many families offered vouchers under MTO lived in projects that were subsequently demolished, and their occupants were offered the option of a housing voucher or a vacant unit in another public housing project at that time. The same options would have been available under the first alternative reform. This reform assumes the same pattern of public housing redevelopment. Its only difference from the MTO reform is that enough occupants of public housing units pay market rents so that the voucher offer does not affect the total number of assisted households.
The MTO evidence is unambiguously applicable to a version of the second alternative reform that retains the same timing of the demolition of public housing projects as occurred under MTO. If the same projects would have been demolished at the same time under either reform, the effect on the occupants of these projects would have been essentially the same. Public housing tenants would have had the same options at each point in time. In this case, the MTO results should provide excellent estimates of the effects of the second reform on the families offered vouchers.
If the projects had been sold at the earliest feasible time after voucher lease up, occupants of the projects sold would have been offered the option of a housing voucher or a unit in another public housing project earlier than under MTO. MTO evidence would be somewhat less applicable to this reform because the accelerated demolition of public housing projects would make staying in public housing more attractive to some families in the two treatment groups and less attractive to others.
For the reasons mentioned previously, however, these changes in the desirability of remaining in public housing are likely to be small, and the bias in the MTO estimates of effects on the individuals in families initially offered vouchers would be modest.
200 Moving to Opportunity Increasing the Value of MTO Research for Housing Policy Development In short, the two alternative reforms should have about the same effects as the MTO reform on families living in public housing projects in high-poverty census tracts who are offered vouchers, so the MTO evidence applies to the effects of the alternative reforms on these families.
Estimating the Costs to Taxpayers of the Alternative Reforms As explained previously, the MTO data could be used to estimate the taxpayer cost of the vouchers offered to households under the alternative reforms. This section addresses the estimation of the additional revenue and cost savings from the alternative reforms that would be available to offset the cost of these vouchers. Existing evidence on the relative cost-effectiveness of public housing and housing vouchers gives us good reason to expect that the former exceeds the latter for the second alternative reform. That is, the reform would have achieved benefits to voucher recipients at a lower taxpayer cost than the current system.
The first alternative reform would have charged market rents for the public housing units vacated by the families who accepted the initial voucher offer. Therefore, it would have generated additional revenue that could have been used to defray a part of the cost of the vouchers. The additional revenue is the difference between the market rents of the units vacated and the rents that the voucher recipients would have paid for them. MTO administrative data on voucher recipients could be used to estimate the latter. Market rents of the public housing units might be predicted in one of two ways. Since the early 2000s, PHAs have been required to estimate the market rents of their units to give their tenants a choice between a flat rent and an income-based rent. To the best of my knowledge, no one has ever studied the accuracy of these estimates. Nevertheless, they could be used for the years available, and the earliest available PHA estimate could be adjusted for real depreciation and inflation to produce estimates for earlier years. Alternatively, the data on MTO voucher units could be used to estimate a hedonic relationship between the rents paid to landlords and the characteristics of the dwelling units and their neighborhoods, and this estimated equation could be used to predict the market rents of public housing units. Previous research has indicated that the rents paid to landlords of voucher units are very close to the rents of unsubsidized units with the same characteristics (Leger and Kennedy, 1990; ORC Macro, 2001; Wallace et al., 1981; Weinberg, 1982).
The second alternative reform would have generated substantial cost savings and additional revenue compared with the MTO reform. The cost savings is the money that was spent to operate and redevelop the public housing projects that would have been sold under the second reform.
Data on these magnitudes are or were in PHA records. However, some considerable assembly may be required, and it may be necessary to impute values for the early years because records have been discarded. Because the new public housing built under HOPE VI and other public housing redevelopment initiatives during the years of the MTO study will provide benefits beyond the time of data collection, the entire cost of this redevelopment should not be subtracted from the cost of the vouchers during this period. Applying an interest rate to the redevelopment cost yields a simple approximation of the cost savings in each year.3 Olsen (2009) suggests a more refined approach.
Selling the projects would have generated the additional revenue. Although many of the structures have little or no market value, the land often has great value. Many projects are located near the center of large metropolitan areas. Because of their proximity to employment, their land would often command a high price if only for its value in future development. Large parcels are particularly valuable. Many public housing sites are in or near gentrifying areas (Holin et al., 2003;
Popkin et al., 2004). In gentrifying areas, the highest bidder for vacant land might be a developer of an upscale condominium complex to house people who work downtown and want to avoid lengthy commutes to their jobs. Or the highest bidder might be the developer of a retail store that would serve these people or the developer of an office building. Over time, business uses of land in central locations of metropolitan areas have squeezed out residential uses. In areas that are not currently experiencing gentrification, the highest bidder for vacant land might be an entity that will hold it for future use. In either type of area, the highest bidder might be the developer of a lowincome housing tax credit project. The tax credit program provides a substantially higher subsidy to developers of projects in census tracts with poverty rates in excess of 25 percent. Estimation of the market value of the land could be based on the sales prices of nearby properties whose structures were demolished in order to convert the land to another use, and on the cost of demolishing the public housing project (Dye and McMillen, 2007; Weber et al., 2006).
Evidence on the performance of low-income housing programs suggests that the cost savings and additional revenue resulting from this reform would have been substantial. The best study that compares the cost-effectiveness of public housing with that of housing vouchers pertains to public housing projects built before 1975. More than 75 percent of public housing units had been completed by that time. The results imply that the excess total costs of public housing compared with those of housing vouchers for equally desirable housing in the two cities studied (Phoenix and Pittsburgh) were 64 and 91 percent, respectively, and the excess taxpayer costs were 97 and 150 percent, respectively (Mayo et al., 1980).
As with all the best cost-effectiveness analyses of housing programs, this study compared the total cost of providing the housing under each program, with estimates of their market rents based on detailed information about the characteristics of the units and their neighborhoods. For tenantbased vouchers and certificates, the approach is straightforward because all the costs associated with providing the housing during a period occur in that period and they are all in the records of the administering agency. Dealing with project-based assistance is more difficult because the time path of costs bears no particular relationship to the time path of the market rents of the units and all project-based assistance involves indirect costs that are not in the records of the administering agency. The ideal measure of cost-effectiveness for project-based housing assistance is the ratio of the present value of the rents paid by tenants and all direct and indirect costs incurred by federal, state, and local governments to the present value of the market rents of the units over the period that the units house subsidized families. If a government owns the project at the time that it stops housing subsidized families, the present value of the project’s market value at that time should be subtracted from the present value of the costs.
A U.S. General Accounting Office study based on a less complete accounting of the cost and much less information about the housing provided found that the HOPE VI public housing redevelopment program was the least cost effective of the active construction programs, with a 27-percent 202 Moving to Opportunity Increasing the Value of MTO Research for Housing Policy Development excess total cost compared with the cost of housing vouchers (GAO, 2001). This estimate clearly understated the total cost of providing housing in the redeveloped projects because it omitted the opportunity cost of the land and the cost of preparing the site from the cost of HOPE VI projects.