The Family Homelessness Systems Initiative, developed by the Bill and Melinda Gates Foundation, provided a decade of funding and support to three counties in the Pacific Northwest, beginning in 2009, to reform their homeless housing and service delivery systems for families. The overall goal of the Initiative was to reduce family homelessness in King, Pierce, and Snohomish Counties by improving the efficiency and effectiveness of the systems. The key targeted outcomes were reducing the length of time families experience homelessness, improving their housing stability, and decreasing returns to homelessness.

Westat, a national research organization, conducted a rigorous longitudinal outcome evaluation to examine the effects of the systems changes on the experiences and outcomes of families served. This report provides 18-month outcome findings based on an analysis involving a cohort of families served after systems reform compared to a cohort of families served prior to reform.

The evaluation provides strong evidence that system changes had a number of impacts on the families served. Key findings include:

• The Housing First orientation of the reformed systems, reducing the reliance on shelter and transitional housing with a broader array of assistance that prioritizes housing such as diversion and rapid re-housing, led to greater and quicker access to permanent housing and more nights in that housing, despite a tightening housing market.
• Families served after reform were less likely to experience sheltered homelessness in the 18 months following system entry than families served prior to reform, but were more likely to experience unsheltered homelessness, especially while waiting for assistance or if they were unable to access housing.
• After entering housing, however, families served after systems reform returned to homelessness (sheltered and unsheltered) at a rate comparable to families served prior to systems reform.
• Families also experienced greater improvements in employment and income than families prior to reform, even controlling for the fact that they came in with higher employment and income. The same economic conditions that likely made it difficult for families to find housing also likely enabled them to increase their employment and income.
• Parent-child intactness, child absenteeism, and the rate of school transitions during the 18 months following receipt of initial assistance in the system did not appear to be affected by the systems reforms.

Several individual family characteristics, in addition to systems changes, had effects on the outcomes. These include:

• Although families of color, especially Black/African American families, are disproportionately represented among those experiencing homelessness, none of the outcome findings varied significantly between families with African American and White heads of household (HOHs), controlling for other characteristics. Families with HOHs that were multiracial or other races, in fact, had some improved housing outcomes over comparable families with White HOHs. Families with Hispanic HOHs, however, were less likely to access permanent housing and had shorter stays than those with non-Hispanic HOHs.
• Having employment, education, and income all increased the probability of having positive housing, employment, and income outcomes for families before and after reform.
• Within each cohort as well, family size and a recent history of homelessness or less time in one’s housing prior to entering the system decreased both the probability of accessing permanent housing and the number of nights in that housing, and increased the number of nights in shelter. Recent evictions likewise decreased the probability of accessing permanent housing and the number of nights in that housing.
• Having a permanent housing subsidy at baseline was also a strong factor in influencing housing access and stability. Families who accessed housing after reform, especially those who accessed it in the first 180 days, were significantly more likely to have a subsidy than families who were unable to access housing during the entire 18-month follow-up.

The study findings have a number of implications for communities nationally. Many of the implications reinforce the work that is underway through Federal and state efforts and funding. The main implications for communities include:

• Prioritize getting families into housing as quickly as possible, as quicker access relates to longer stability;
• Maintain access to shelter separately from coordinated entry to avoid unsheltered homelessness while families wait for other assistance;
• Strengthen ties with employment agencies and work to improve families’ human capital given the strong association of human capital to housing outcomes;
• Bridge the homeless service system with the public housing authorities, recognizing the critical role subsidies play for some families in maintaining their own housing;
• Consider providing additional supports to families who enter coordinated entry with larger numbers of children, histories of homelessness, and recent evictions; and
• Reduce any remaining stock of transitional housing and/or consider repurposing it or targeting it to those who might have repeated difficulty accessing housing and return to homelessness.

Follow-on studies can build on the findings from the current study in a variety of ways. For example, as systems have now increasingly implemented dynamic prioritization procedures and are targeting assistance to families with different histories and needs, future studies should examine how the targeting of different types of assistance relates to families’ access to and stability in housing. More controlled studies of new system interventions, such as diversion assistance, may also be beneficial in understanding the role each plays in reducing homelessness. Finally, future research should adopt a stronger racial equity lens. Such research should ensure that sufficient numbers of families across racial groups are represented in quantitative studies to conduct a sensitive assessment of families’ experiences and outcomes. Additionally, it should ensure that the studies are co-designed with families from the relevant groups to ensure that the studies are sensitive to the biases and inequities they experience.