What the research says about the best way to spend money now to solve long-term homelessness


Op-ed: We have lots of data on what works to end homelessness. Now we have to get to work.

The Supreme Court’s decision overturning the Centers for Disease Control and Prevention’s eviction moratorium has put millions at risk of housing instability and homelessness. If households are unable to receive emergency housing assistance in time, many could join the nearly 600,000 people who were left homeless on any given night in the United States before the pandemic. To reverse the growing trend of homelessness, people facing housing instability need effective, evidence-based solutions. Fortunately, decades of rigorous research have shown practitioners and policy makers what can work. And right now, the influx of relief funds presents an opportunity to finally expand these evidence-based programs, but those dollars need to be spent quickly. To truly end homelessness, states and cities must both develop effective short-term solutions and conduct additional research to build long-term, sustainable stability.

“Housing instability” is a term that encompasses both homelessness and factors that increase the risk of homelessness, such as couchsurfing and eviction. Those expelled are moving more frequently, making it harder to maintain steady employment, loosening community ties and, for the 1.3 million students who experienced homelessness during the school year 2018-2019, interrupts education.

Homelessness and housing instability disproportionately affect Black people, members of the LGBTQ community, people with severe mental illness, veterans, survivors of domestic violence, and several other marginalized communities. And that was before the pandemic and the impending eviction cliff.

Fortunately, there is already strong evidence on many effective strategies to end homelessness. At J-PAL North America, a research organization based in the Department of Economics at the Massachusetts Institute of Technology, we work with researchers and policy makers to conduct rigorous randomized evaluations of policies that actually lift people out of poverty.

For example, research has shown that subsidizing rent through the Housing Choice Voucher program, commonly referred to as Section 8, can reduce homelessness. Voucher holders have a lower incidence of homelessness and an increased likelihood of living in stable housing. Vouchers have also been shown to help beyond simply putting a roof over someone’s head. They decrease family separation and psychological distress, and increase school attendance and food security. Additional research by J-PAL researchers has shown that providing comprehensive mobility services and improving access to neighborhood information in conjunction with vouchers can enhance their positive impacts. But only 25% of eligible households can receive vouchers due to insufficient funding, and applicants spend years on waiting lists. (President Biden has offered an additional $5 billion for the Housing Choice Voucher program, which would help another 200,000 eligible families.)

Randomized evaluations have shown the effectiveness of Housing First. For decades, most housing services have required people to meet employment or sobriety requirements in order to be housed. Housing First programs literally do what they say: provide housing first, without preconditions.

Several randomized evaluations have tested Housing First programs and found that people in permanent supportive housing spend about half the time without housing or in hospital than those not in the program. Utah policymakers then used this research to expand the evidence-based program. The state became a national model after adopting a Housing First approach in 2005 and reducing chronic homelessness by 91% by 2015.

There is still much to learn about how to end homelessness, and rigorous random assessments can help. First, further research can provide insights into how to improve programs that we know can work. For example, despite the success of vouchers, they are not just a “golden ticket”. It is difficult to get a voucher and then find landlords who will accept such vouchers, even when discrimination based on “source of income” is technically illegal. The low housing stock and discriminatory practices make finding housing difficult, if not impossible, for voucher holders. We need more research on ways to increase owner acceptance of vouchers.

Second, randomized assessments can be used to identify other evidence-based strategies. While eviction moratoriums have offered short-term solutions, researchers are exploring potential long-term solutions. Many studies are underway on the provision of emergency financial assistance and legal support services to people facing eviction.

The work of ending homelessness takes a lot of money. The US Treasury Department encourages states and localities to use federal relief funds for evidence-based interventions and assessments “designed to be evidence.” However, US bailout funds are time-limited and must be spent by the end of 2024. By spending these funds on both evidence-based programs like Housing First and Generative Assessments evidence, stakeholders can simultaneously address immediate needs and support housing stability for decades to come.

Rohit Naimpally is Senior Director of Research and Policy at J-PAL North America, where he works with governments, nonprofits, foundations, and private companies to develop rigorous evidence on policies and programs that have an impact on poverty.

Laina Sonterblum is Policy Associate at J-PAL North America. She supports the development and implementation of randomized evaluations and synthesizes research findings to promote evidence-based programs and social policies.


Comments are closed.