Breaking the Homelessness-Jail Cycle with Housing First: Results from the Denver Supportive Housing Social Impact Bond Initiative

Research Report

Breaking the Homelessness-Jail Cycle with Housing First: Results from the Denver Supportive Housing Social Impact Bond Initiative

Abstract

The Denver Supportive Housing Social Impact Bond Initiative (Denver SIB), launched in 2016 by the City and County of Denver, aimed to increase housing stability and decrease jail stays among people who were experiencing chronic homelessness and had frequent interactions with the criminal justice and emergency health systems. The initiative provided supportive housing (a permanent housing subsidy plus intensive services) through a Housing First approach, which aimed to quickly get people out of homelessness and into housing, without requiring that participants meet preconditions or requirements.

The Urban Institute, along with our partners at The Evaluation Center at the University of Colorado Denver, tracked the implementation of the initiative and evaluated the efficacy of supportive housing over a five-year period. The evaluation implemented a randomized controlled trial of 724 individuals. People eligible for the supportive housing program were randomly assigned to one of two groups; individuals in the treatment group were offered supportive housing services, while individuals in the control group received usual care services in the community. A total of 363 people were randomized into treatment, while 361 were placed in the control group. Of those in the treatment group, 79 percent (285 people) were located, engaged, and housed. We found the following:

  • Denver SIB supportive housing program participants spent significantly more time in housing than those in the control group, as measured by housing assistance. Those referred to SIB supportive housing received 560 more days of housing assistance over three years, compared with those who received services as usual in the community.
  • After accessing supportive housing, most participants stayed housed over the long term. Excluding those who died during the observation period, 86 percent of participants remained in stable housing one year after entering housing. At two years, the housing retention rate for living participants was 81 percent. At year 3, the rate was 77 percent.
  • Shelter stays for Denver SIB supportive housing program participants decreased dramatically. When counting all instances of shelter use—including during the day and at night—over a three-year period, those referred to supportive housing had 127 fewer shelter visits compared with their peers in the control group. This represents a 40 percent reduction in shelter stays because of supportive housing.
  • Police interactions went down. People referred to supportive housing experienced eight fewer police contacts and four fewer arrests than those who received usual services in the community. This represents a 34 percent reduction in police contacts and a 40 percent reduction in arrests.
  • The reductions in jail stays and jail days were notable: In the three years after randomization, participants referred for supportive housing had almost two fewer jail stays and spent an average of 38 fewer days in jail than those who received usual care in the community. This represents a 30 percent reduction in unique jail stays and a 27 percent reduction in total jail days.
  • Denver SIB supportive housing program participants used short-term or city-funded detoxification services less often than those in the control group. In the three years after randomization, people referred for supportive housing had four fewer visits to a short-term or city-funded detoxification facility than those who received usual services in the community. This represents a 65 percent reduction in detoxification services. The differences between the two groups’ uses of emergency medical services were not statistically significant.
To reuse content from Urban Institute, visit copyright.com, search for the publications, choose from a list of licenses, and complete the transaction.