How System Modeling Can Help Build a Stronger Response to Homelessness

September 13, 2018

What Is System Modeling?

Local homelessness services systems comprise a series of interconnected interventions. The programs that people access, the way they move between programs, and the length of time they stay in each program, all have repercussions for other parts of the system. When the relationships between these interventions are well understood, a community can “right-size” each component to address peoples’ actual needs and preferences and achieve an efficient and effective allocation of resources.

System modeling makes this type of analysis possible. Through this approach, Continuums of Care (CoCs) determine the amount of each intervention to provide, based on assumptions about the annual inflow of people seeking assistance (overall and for priority populations, including those disproportionately represented in the homelessness services system), the share of those people that will need different types of interventions, and the time spent in each intervention.

System modeling generally involves the following steps:

  1. Map out the components of the desired homelessness services system (e.g., diversion, short-term crisis housing beds, rapid re-housing, permanent supportive housing).
  2. Estimate the annual number of people projected to be served, including any populations that have been prioritized for assistance (e.g., youth or people with chronic patterns of homelessness).
  3. Review actual system utilization data to determine the types of programs people access, how long they stay, and whether they exit to permanent housing, another program, or elsewhere.
  4. Identify “ideal” pathways, meaning combinations of interventions that would improve permanent housing outcomes for people experiencing homelessness.
  5. Develop assumptions about the number and percentage of people that would need each pathway to exit homelessness to permanent housing and the average length of assistance expected to be needed from each system component.
  6. Use these assumptions to develop a model for an ideal homelessness services system that efficiently and effectively allocates resources based on varying levels of need (e.g., minimal or limited intervention to intensive or longer-term assistance) to maximize the number of individuals that can be served and successfully access permanent housing. System models can be designed to generate inventory targets for each system component and an estimate of the cost to meet these targets. Or given inventory projections, system models can help CoCs assess the extent to which they can meet need with projected system resources.
  7. On an ongoing basis, review actual utilization data and projections and adjust the allocation of resources within the system as needed to ensure continued maximize efficiency and effectiveness.

How Did System Modeling Help One Community Meet Its Goals?

Since 2002, the City of Indianapolis has regularly published an updated Blueprint to End Homelessness that lays out a roadmap for the city’s planned activities to address homelessness over a five-year period. Those involved in preparing the most recent report for 2018-2023, wanted to take a closer look at:

  • Who experiences homelessness in Indianapolis?
  • What types of housing and services are currently available for these individuals and families, and what types of housing and services most effectively serve various populations?
  • For each type of intervention (e.g., rapid re-housing, permanent supportive housing, etc.) how many beds or housing units does Indianapolis currently provide? How well does the current distribution address the need, and how many beds or units are needed to effectively end homelessness?

To answer these questions, the Coalition for Homelessness Intervention and Prevention (CHIP), a local non-profit organization, engaged Abt Associates to conduct a system modeling exercise. The team analyzed data on current system usage and developed inventory recommendations for various household types and subpopulations, including Veterans, people experiencing chronic homelessness, and families with children.

What Did the System Modeling Process Look Like for Youth and Young Adults?

After the initial system modeling exercise was complete, Abt worked closely with CHIP, the Indianapolis CoC, and other community leaders that were part of a Youth Strategic Planning Committee, to engage in additional planning and modeling for: 1) unaccompanied youth ages 13 to 17; 2) single young adults ages 18 to 24; and 3) young adults ages 18 to 24 with children of their own. These groups are collectively referred to as youth and young adults (YYA). This effort included consultation with providers as well as several YYA who had experienced homelessness.

One of the key decisions the group made was to adopt HUD’s definition of homelessness, which  encompasses four categories , for the system modeling. Stakeholders in Indianapolis then estimated the number of YYA who currently experience homelessness each year.

During the youth system modeling, stakeholders also discussed the location of existing providers relative to the location of youth experiencing a housing crisis. Through this conversation, they identified a need to have more geographic dispersion of new projects serving youth experiencing homelessness.  This recommendation was, in part, driven by the race of youth accessing services and the racial makeup of providers and the neighborhoods in which they were located.

The committee created a map of their desired system, starting with entry points where YYA would be identified (e.g., schools, hospitals, social service providers). Additional categories on the map include points of engagement, such as drop-in centers or street outreach, crisis response assistance, and permanent housing options. In each of these categories, stakeholders identified existing resources (such as transitional housing and shelters), aspirational resources (host homes and new affordable housing to be developed), and resources currently in development (including rapid re-housing).

With this vision of their ideal system, the committee combined programs into pathways that would facilitate a rapid exit to permanent housing. This model included estimates of the percentage of YYA expected to access each pathway, and how long they would stay at each program along the way. These assumptions were “road tested” with youth who had lived experience with homelessness and organized into an “ideal” system to end homelessness among youth in Indianapolis.

What Was the Impact of This Exercise?

The system modeling exercise has yielded several benefits for stakeholders in Indianapolis, including:

  • Strong applications for competitive federal funding programs to address youth homelessness.
  • A framework to guide the allocation of state funds and engage public and private funders at the local and state levels.
  • A roadmap for local staff who are working with YYA, providing direction for their work.
  • Reallocated resources to better meet the needs of YYA in their community or to redesign programs to align with the vision of the youth system.
  • A framework in place to strengthen work with partners, like the university and community college systems, to increase the number of providers entering data in the HMIS, and to generate more accurate estimates of YYA homelessness from which to refine the system map and inventory needs.

This article is adapted from a series of blogs published by the Center for Evidence-Based Solutions to Homelessness, a resource center dedicated to ending and preventing homelessness by connecting research to practice. To learn more about the Center, visit .

Back to News