Encouraging Residential Moves to Opportunity Neighborhoods: An Experiment Testing Incentives Offered to Housing Voucher Recipients

Abstract

Substantial benefits can accrue from living in low-poverty neighborhoods, yet approximately 80% of the 2.2 million Housing Choice Voucher (HCV) recipients rent homes in moderate- or high-poverty census tracts. The Chicago Regional Housing Choice Initiative tested several ways to promote opportunity moves. It included the first experiment that tests whether two types of light-touch incentives induce opportunity moves for HCV recipients who had requested a moving voucher. Based on the 2,005 HCV recipients in the study, we found that neither the offer of a $500 grant nor the offer of a $500 grant coupled with free mobility counseling induced opportunity moves. The receipt of mobility counseling also did not boost opportunity moves. Regardless of the type of offer, 11%–12% of participants moved to opportunity neighborhoods. Despite requesting a moving voucher, half of the study participants remained in place, indicating significant barriers to moving. We offer potential reasons for the results and conclude with two recommended pilots to increase opportunity moves.

Keywords:

• Mobility
• counseling
• vouchers

Acknowledgments

We wish to thank the many people who helped make this study happen. First, we wish to thank The John D. and Catherine T. MacArthur Foundation for funding this research through Grant 101434-0. The work that provided the basis for this publication was also supported by funding under a grant with the U.S. Department of Housing and Urban Development, Office of Policy Development and Research. We also acknowledge the staff from the seven PHAs who administered the moving briefings that were at the core of this study. Housing Choice Partners (Christine Klepper, Executive Director) provided the mobility counseling and data about that counseling for the study. Breann Gala who was then working at the Metropolitan Planning Council organized the regional effort, and Robin Snyderman helped to carry out the regional plan through BRicK Partners. We also thank the Chicago Regional Housing Choice Initiative advisors for the time they devoted to the project. The advisors are: Alex Polikoff, BPI Chicago; Paul Fischer, retired professor, Lake Forest College; Barbara Sard, Center on Budget and Policy Priorities; Ianna Kachoris, John D. and Catherine T. MacArthur Foundation; Ingrid Gould Ellen, New York University; Jeff Lubell, Abt Associates; Juanita Irizarry, Friends of the Parks (formerly with Chicago Community Trust and Office of Illinois Governor Pat Quinn); Kalima Rose, PolicyLink; Mary K. Cunningham, Urban Institute; Phil Tegeler, Poverty and Race Research Action Council; and Stephanie DeLuca, Johns Hopkins University. Cole Sutera provided excellent programming support and was a critical member of the research team. Lydia Tagavi and Ndeye Jackson matched study data to PIC data. Kaitlyn McClain, formerly with the Metropolitan Planning Council, transcribed housing authority sign-in sheets into usable, digital data. We thank Jennifer O’Neil for her input on mobility counseling programs. Thank you to Rolf Pendall for generously sharing Urban Institute’s Neighborhood Advantage Score data set. The substance and findings of the work are dedicated to the public. The authors and publisher are solely responsible for the accuracy of the statements and interpretations contained in this publication. Such interpretations do not necessarily reflect the views of the government, The MacArthur Foundation, RAND Corporation, Housing Choice Partners, or the Metropolitan Planning Council.

Notes

1. The HCV program is a federally funded program costing approximately $19 billion per year that provides housing subsidies to approximately 2.2 million low-income households in the United States. The voucher defrays costs to a low-income participant to rent a home in the private market. Tenant-based rental assistance, which has evolved over time into the HCV program, started in 1974, and voucher recipients contribute 30% of their income after certain adjustment to the combination of rent and utilities. The federal government pays the balance of the rent, up to a price cap that is set by state and local agencies based on HUD-determined metropolitan or county fair market rent (FMR), which is usually at the 40th percentile of recently rented nonluxury units. There are exceptions, such as in Chicago where the price cap is set at the 50th percentile. At least 75% of new families admitted each year to an agency’s HCV program must have income at or below 30% of the area median income, which is approximately the federal poverty line. Program rules also stipulate that the rental home that the voucher recipient selects must pass a physical inspection to meet housing quality standards (HQS).

2. HUD publishes and annually updates metropolitan-wide FMR and county FMR in nonmetropolitan areas, which are usually set at the 40th percentile of recently rented nonluxury units. Housing agencies administering the HCV program may cap the value of HCV subsidy at 90%–110% of the FMR, or else obtain HUD approval to set the cap higher or lower. Starting in FY 2013, HUD has tested via a demonstration program setting capped voucher subsidy levels by small area FMR, which has the effect of increasing subsidy in higher cost zip codes and lowering them in low-cost ones (Kahn & Newton, 2013). HUD has since published hypothetical small area FMR levels by zip code to allow PHA to adjust their FMR levels if the PHA so wishes for its tenant-based vouchers, so long as the adjusted rate remains within 90% to 110% of the metro-wide FMR. See Fischer (2015) for more detail on the small area FMR.

3. Affordable rentals means ones with rents less than the FMR value that HUD publishes.

4. Once a moving voucher recipient has identified a rental unit she wishes to rent, she must complete a request for lease approval form which the landlord must sign to indicate his/her willingness to participate in the program. Once the PHA receives the signed form, the PHA schedules an inspection of the unit, which may or may not trigger reinspections if shortcomings are identified for remedy. Once the unit has passed inspection, the PHA issues a contract to the voucher holder, who must get it signed by the landlord and then resubmit it to the PHA. Then once that is submitted, families can move in. Also, the Housing Assistance Payment contract between the PHA and landlord must be processed before landlords begin to receive the subsidized portion of the rent.

5. As a yardstick for car ownership rates among voucher recipients, in MTO, 37% of the participants owned cars (MTO participants were living in urban public housing), whereas 65% of Welfare to Work participants owned cars (Pendall et al., 2014).

6. Two other salient distinctions between MTO and CRHCI is that CRHCI eligibility was not restricted to families with children, and that CRCHI examines a group of HCV recipients living in nonopportunity areas who requested a moving voucher as compared with MTO which targeted residents living in public housing in high-poverty neighborhoods.

7. The seven housing authorities are: Chicago Housing Authority, DuPage Housing Authority, Housing Authority of Cook County, Lake County Housing Authority, McHenry County Housing Authority, Oak Park Housing Authority, and Waukegan Housing Authority.

8. The Chicago Community Trust contributed $200,000 to HCP for implementation, and the Cook County CDBG program supplemented counseling costs at $187,953. Separately, The John D. and Catherine T. MacArthur Foundation also funded RAND to lead an independent evaluation of the third component of CRHCI. Included in the RAND evaluation budget are funds for the Metropolitan Planning Council to audit briefings and collect data and for HCP to collect counseling data.

9. In the HCV program, households can request to move after living in their current apartment for 12 months as long as they have not violated their lease and are in good standing with their PHA’s HCV program. HCV participants start the moving process by requesting moving papers from the PHA that administers their voucher and then notifying their landlord that they intend to move. HCV clients can request moving papers up to 60 days before lease expiration in the current apartment.

10. PHA submit electronically to the PIC System at least annual updates for households in Public and Indian Housing assisted programs, of which HCV is one.

11. We do include Chicago Housing Authority households in analyses testing whether the $500 grant incentive had different effects on opportunity moves than the $500 grant and counseling.

12. In particular, the primary source of data loss—missing personally identifiable information from sign-in sheets—is not correlated with the outcomes of interest described below or the randomly assigned briefing type.

13. See more details about compliance in the supplementary Technical Appendix.

14. We also conduct the analyses ignoring noncompliance altogether by only considering the type of offers made to households per the sign-in sheets to test possible statistical significance of offers made. However, these analyses should not be interpreted as causal given that noncompliance does not appear to be random.

15. See the supplementary Technical Appendix for a detailed description of the secondary outcome measures.

16. To estimate permanent income, we averaged household-level income for all time points prior to the randomized briefing that was present in the HUD longitudinal data. On average, we had four prebriefing time points for a given household to derive an average income that we called permanent income.

17. As described in Section 4, briefings administered did not always match the type of moving briefing that was randomized for that day.

18. We also tested whether the offer of the $500 grant or the offer of the $500 grant coupled with counseling induced households to move at all or to move to a nonopportunity neighborhood, and found that the differences in move rates between the control group and the two treatment groups are not statistically significant.

19. These regressions were estimated with two-stage least squares, and the coefficient on the instrument in the first stage regression is statistically significant.

20. In fact, the coefficient on grant-only for the neighborhood advantage score is negative and statistically significant at the 5% level. It is unclear how to interpret the negative coefficient, because the score is a composite of multiple indicators, and an overall negative score could imply improvements in some component scores, which were outweighed by lower scores on other components. Also, correcting for multiple-hypothesis testing for the three secondary outcomes makes this coefficient statistically not significantly different from zero.

21. A journalistic investigation of all Chicago and Cook County Craigslist rental listings in June 2015 found that anti-Section 8 bias remains, despite the source-of-income laws making such statements illegal (Yousef, 2015).