By Maria Alva (Georgetown University), Natnaell Mammo (The Lab @ DC), Ryan T. Moore (The Lab @ DC), and Sam Quinney (The Lab @ DC)
The District of Columbia piloted and evaluated a shallow rent subsidy to answer two questions: Do shallow flexible rental subsidies promote housing stability? And, can they be a vehicle to further stretch the existing housing resources to serve more people? These questions are important to growing metropolitan areas like D.C. that face severe challenges in making housing affordable and preventing homelessness. Similar to New York or San Francisco, most D.C. residents are renters, 70% of whom spend more than 30% of their gross income on rent. Approximately 1 out of every 125 residents in D.C. is in emergency shelters, in transitional housing, or is unsheltered. Housing vouchers, representing “deeper subsidies,” have historically been in short supply and, necessarily, targeted at the most vulnerable households. In 2017, D.C.’s Department of Human Services (DHS) decided to test a model that could serve more residents by targeting a shallow subsidy to families experiencing housing instability but not homelessness. To this end, DHS began piloting the Flexible Rent Subsidy Program (D.C. Flex).
D.C. Flex is innovative on several accounts. Unlike other housing programs, the subsidies go directly to participants rather than their landlords, and the program does not require participation in case management. During its initial pilot, the shallow subsidy was fixed at $7,200 per year and flexible—people can access funds as needed, using up to one month’s rent. The program functions more like basic income than a rigid housing subsidy.
Our study shows no effect of D.C. Flex on homelessness. However, we find that D.C. Flex decreases utilization of other homelessness-related services and offers potential cost savings depending on who takes up the program. At a minimum, our results show no harm—when participants are given a shallow subsidy and more autonomy in budgeting their subsidy, there is no increase in homelessness.
To qualify for D.C. Flex in 2017, applicants had to:
● Be the leaseholder for a rental unit in D.C.
● Have applied for one of D.C. government’s homelessness or rental assistance programs in the past four years.
● Be at or below 30% of Area Median Income (< $33,000 a year for a family of four).
● Be at least 21 years old with custody of a minor.
● Have current or recent employment (within the past six months from the time of their application).
Six hundred sixty-eight households applied and were randomized across five lotteries from May to September 2018, and 229 were selected to receive D.C. Flex. Our study looks at the program’s first-year results. Each participant’s outcome is weighted to account for the fact that people who applied early were in more lotteries, increasing their chance of receiving D.C. Flex.
We relied on six rich data sources. The primary outcome of interest was homelessness, which we define as entering emergency shelter, because unsheltered homelessness is rare among D.C. families. We also measured the use of DHS’s other homelessness prevention services and programs, costs to DHS, and monetary benefits to participants.
Based on self-reported information, study participants were most likely to be single mothers in their late 20s and early 30s with two dependent minors. The majority were employed, had an average self-reported income of approximately $17,500, and spent about 59% of their income on rent.
After one year in the program, participants had an average of $543 of unused funds in their D.C. Flex accounts, less than half the average rent ($1,147). These usage patterns suggest that many households took advantage of the flexibility and budget over the course of the year.
We find that participants in both D.C. Flex and the control group were unlikely to experience homelessness (roughly 2%) in the first year. This finding indicates that the program effectively identified a population further “upstream” from the shelter door and that we would be unlikely to detect a reduction in homelessness. We do, however, see a statistically significant reduction of 12.7 percentage points in the rate people use existing homelessness prevention services, with a complier average causal effect of 28.6 percentage points. This large effect is driven by D.C. Flex enrollees having to give up some forms of housing support, like Rapid Rehousing, to enroll in D.C. Flex. We assessed whether D.C. Flex had any effect on emergency housing assistance or cash welfare receipt and the amount of benefit received by participants. While our estimates suggest decreases in the need for these other programs, none of the impacts are statistically significant or economically relevant.
While D.C. Flex is one of the lower-cost housing programs DHS provides—much less costly than permanent supportive housing or vouchers and slightly more costly than its eviction prevention programs—we find substantial heterogeneity in the program’s impact on costs and benefits. The program can be cost-effective on a yearly basis, but only if high uptake and substantial substitution occur from other, more costly programs. Only 45% of those offered the program enrolled for the first year. The annual net cost to DHS and participant’s benefit were substantially lower than the face value benefit of the program ($7,200) per household per year, suggesting that people have to make tradeoffs in enrolling in the program. We also find that the likelihood of uptake is negatively correlated with measures of housing need—as households’ likelihood of program uptake increases, the cost to DHS also increases. This occurs because services had not previously been provided to many lower-need households (i.e., higher self-reported annual incomes and likelihood of employment, fewer dependents) who now qualified for D.C. Flex. On the other extreme, high-need households tended to opt out of the program when a deeper subsidy with more support services is available to them (e.g., being already enrolled in D.C.’s Rapid Rehousing Program).
If shallow subsidies like D.C. Flex can target low-need families that still experience housing instability, the program may be a good substitute for programs like Rapid Rehousing over the long term. With a lower cost, this represents a win-win for housing agencies and households.
Maria L. Alva is an Assistant Research Professor at the Massive Data Institute at Georgetown University’s McCourt School of Public Policy. Maria works on impact evaluations of public programs and the cost-effectiveness of those investment decisions. Maria earned a DPhil in Public Health and an MPhil in Economics from the University of Oxford.
Natnaell Mammo is a Data Scientist with the Consumer Financial Protection Bureau (CFPB) and a volunteer Data Scientist with The Lab @ DC. Nat graduated with an M.S. in Computational Analysis and Public Policy from the University of Chicago and a B.S. in Economics from the University of Pennsylvania.
Ryan T. Moore is an Associate Professor of Government at American University and a Senior Social Scientist in The Lab @ DC. Ryan’s research interests center around statistical political methodology, with applications in American social policy and in particular economic security, housing, health, and public safety domains. Ryan received his Ph.D. in Government and Social Policy and A.M. in Statistics from Harvard University.
Sam Quinney is the Director of The Lab @ DC. Under his leadership The Lab uses human-centered design, predictive modeling and rigorous evaluation to improve the lives of Washington, DC residents. Sam’s research interests focus on homelessness prevention, public safety, and education. He holds a Masters of Public Policy from the University of Chicago.