(Photo by iStock/joebelanger)
What is not working about the US response to homelessness? Every year, cities spend more on services, but homelessness is on the rise and public frustration with the status quo is growing. Providers are exhausted. Residents are skeptical. Leaders feel stuck.
What if we’re trying to solve too many problems at once with a system built to solve none of them well? This thought experiment draws on our work at the Harvard Kennedy School Government Performance Lab (GPL) with state and city leaders across the US to explore what it might take to radically redesign the way communities serve people with severe mental illness living on the street. There is a helpful and illuminating precedent from communities across the country about what it takes to reimagine a complex system.
The Challenge and the Broken Status Quo
Cities and counties are spending billions of dollars on homelessness services each year, and yet when the GPL speaks with government leaders and community members across the United States, we hear similar concerns:
- As a city leader, I care about people in my community who are experiencing homelessness, but I’m frustrated that we keep being told the only solution is putting more money into the same system. We need to be able to show that this spending works or we will lose residents’ trust.
- My fellow service providers are well-intentioned and care a lot, but funding rigidity means we spend a lot of energy on administrative overhead, tracking activities back to specific funding sources rather than just trying to serve people with the resources they need, when they need them. We’re always told the system is too expensive, but none of us individually has enough resources or flexibility to do what we think needs to be done for our clients.
- My loved one has an illness and is living on the street. We have tried to get them help, but it’s an impenetrable tangle of overlapping services. Providers tell us they don’t have space or can’t do anything because my loved one is hard to work with.
The homelessness response system has two parts—one is large but completely invisible to most community members. The other is much smaller but extremely visible. These two parts of the system represent different underlying problems that require separate solutions but have been cobbled into one Byzantine system.
The large, invisible part of the homelessness response system is for people who are facing homelessness because they cannot afford housing. Often, these people are precariously housed, close to missing payments, or doubled up with friends or family when one of a myriad of stressors—job loss, health care expense, interpersonal dispute, rent raise—pushes them into homelessness. As housing costs skyrocket, more and more people live on the edge where a small stress can tip them over. This part of the system is the biggest but goes largely unseen, since families and others experiencing economically driven homelessness are often not visibly homeless. Many of them cycle between friends’ houses, go to their jobs in the morning after sleeping in their cars, or send their kids to school from shelters. The homelessness response system provides eviction prevention, time-limited housing subsidies, emergency rental assistance, and navigation supports to try to help stop people from falling into homelessness or speed up their reentry into housing. There are many clever policy and operational fixes that jurisdictions can take to improve the efficiency and delivery of these services. However, what residents and policymakers need to understand is that these services, while vital to blunting the harms of rising housing costs, do not address the root cause of the problem, which is housing affordability. Even funding these vital services in full and running them perfectly will not “solve” homelessness. To turn off the pipeline—to reduce the number of people entering homelessness—jurisdictions must figure out how to increase income or drastically reduce the cost of housing, including through significantly increasing housing supply.
The second, visible part of the homelessness system is a group of people who cannot stay stably housed without extra support, even if cheap or free housing was given to them. Many of these folks are experiencing severe mental illness (SMI) or behavioral health challenges like substance use disorders, often exacerbated by the stressors of street homelessness. These individuals are often living on the street, have experienced homelessness for years, and may have cycled in and out of incarceration, emergency psychiatric care, and shelter living. Advocates often point out that this group is a small fraction of the total number of people experiencing homelessness. But for better or worse, visible street homelessness at the intersection of severe mental illness and behavioral health is the salient symbol of homelessness and an outsized driver of public perceptions about public safety, the efficacy of homelessness services, and government’s ability to address disorder. Anise Vance, the Assistant Director of the Community Safety Department in Durham, North Carolina, told me, “At this moment, there are almost 1,400 people who are experiencing homelessness in Durham. A bunch of those are people with jobs, are families with kids who can’t afford their rent. But about 200 of them are neighbors with severe mental illness and behavioral health challenges. I think that if you asked a Durham resident how many people are homeless in Durham, they’d guess 200. Because those are the neighbors all the calls are about, who are in and out of our other institutions the most often, the neighbors that people see, whose behavior concerns them or makes them feel unsafe. So I understand that if we can’t address the visible neighbors with severe mental illness, residents will assume that we aren’t effective, and they’ll lose trust in the whole system.”
Leaving aside the pressing and important question of how to create housing affordability nationally that has garnered so much attention recently, let’s focus for this piece on this second part of the system.
Radical Redesigning
What would it look like to radically redesign homelessness services for individuals with severe mental illness (SMI) experiencing chronic or street homelessness? What would it take in terms of executive leadership, community and provider innovation, contracting methods, and funding? What are the risks, major considerations, or necessary conditions of these changes? How might a city test this out?
Let’s do a blue-sky thought exercise together. Imagine that we lead a mid-sized American city. Here is the status quo we’re likely facing:
A focus on process, not people: Someone experiencing severe mental illness or behavioral health crises may need intensive, individualized support that addresses both their housing and mental health needs simultaneously, often requiring coordination across multiple systems that weren't designed to work together. One client might need someone to help her navigate the complexities of the Social Security disability system while also managing her bipolar disorder and securing housing that will accommodate her emotional support dog. Another might have schizophrenia and requires intensive case management to coordinate his treatment, conduct a housing search, and replace identification documents that were stolen. A key insight from practitioners working with these populations is that individuals facing such challenges need to be known personally, with someone dedicated to guiding them through a complex, daunting process. This means having a single, dedicated point of contact who wakes up each day focused on what supports the client needs to get into stable housing and treatment. That might involve helping them obtain a birth certificate, accompanying them to psychiatric appointments, ensuring they’re connected to clinical care, following up on medication adherence, assisting in apartment searches that allow pets, advocating with landlords about their housing history, coordinating with treatment teams, or simply providing a steady presence amid chaos. Under the current system, these folks often get bounced between multiple providers, referred into services that are full, and told to wait months until resources become available, falling through the cracks due to a lack of consistent support.
No clear purpose: Is the purpose of services to get people housed? Or to give them services on the street that help address immediate concern, such as behavioral health needs? Or to respond to emergency calls? To reduce crime? Without a clear purpose, it’s hard to make progress. This lack of clear vision is embodied in internally contradictory contracts. Because different contracts and services serve different purposes, they do not add up to coherent progress in any specific direction. In 2018, the GPL helped Chicago’s Department of Family and Support Services redesign 80 percent of its $330M annual budget with an outcomes-based focus. One of the things the GPL heard from providers during the diagnostic phase was a confused and contradictory understanding of the purpose of the contracts. Contractors didn’t see their services as contributing to any specific shared aim.
Everyone is involved but no one is accountable: Without a clear shared purpose, spending pays a wide set of providers for specific inputs or activities, not for results. No provider is responsible for following someone all the way through and doing everything they can to get them housed. Instead, providers are paid for conducting activities (street outreach, navigation, support services). The Streets to Stability services map of homelessness response services in San Francisco illustrates one example of this fragmentation, which increases the burden on both staff and clients to navigate. Beth Sandor, who led the national initiative Built for Zero as a cofounder of Community Solutions and is now leading the nonprofit Field Impact Partners, told me, “We have cities paying for navigators to navigate the other navigators. When you are paying someone $4,000 per client to just try to make sense of your system, maybe the solution is to change the system.” A provider in a large city said, “The overlapping boundaries create the ‘everything is my responsibility, so nothing is my responsibility’ mentality. There is always someone else to blame. Clear lines of responsibility are necessary.”
Providers are constrained by rigid funding and fractured contracts: The homelessness response system distributes funding through a complex web of interwoven federal, state, local, and philanthropic funding. Each line of funding has its own restrictions and reporting requirements, and the funding is granted out through intermediary governance structures like HUD-mandated Continuums of Care or city agencies. Even though huge amounts of money flow through the system, the dispersal of the spending across many agencies and non-profit providers and the rigidity of the funding sources means that providers don’t have the ability to meet individual needs in creative and flexible ways. As one example, the permanent supportive housing in a city may be funded directly by HUD through contracts with local housing providers, but the street outreach teams that contact the clients who need the permanent housing may be funded by Emergency Solutions Grants flowing from HUD to the city. And although the Continuum of Care is nominally coordinating local action around homelessness, it typically doesn’t hold the contracts themselves and therefore can’t impose meaningful consolidation or performance management strategies, including making funding more flexible or reallocating funding from underperforming providers to stronger ones. The most effective providers in a city are often ones who cobbled together funding in a creative way that allows them more autonomy and flexibility in their program design.
* * *
In our imaginary city, we want to wipe the slate clean on our approach to individuals with SMI living on the street and start from scratch. We have the perfect hand—a committed mayor willing to pay the political cost for upsetting the status quo, innovative and strong providers eager to partner with the city on something new, and local philanthropists who want to fund innovative experiments with flexible dollars. How might we design our new system?
Define success: First, we need to replace multiple competing goals with a clear vision for success. Because of the many entrenched stakeholders, this takes strong leadership. Here’s one example of a definition:
In our city, individuals who were living on the street with severe mental illness are stably housed. We will measure this in three ways:
- Reduction in visible street homelessness.
- Placement stability: Percentage of homeless individuals with SMI who successfully maintain stable housing placements (including being re-housed quickly if they become unhoused).
- Inflow reduction: People with SMI who have been housed do not return to street homelessness and new unhoused people with SMI in our city do not enter street homelessness.
Find flexible resources: Eventually we’ll want to argue for significant changes to the overly rigid ways that federal and state funding are structured and make data-driven changes to existing spending. But to get a demonstration of our new system off the ground, let’s ask a group of local philanthropists in our city to back our pilot, perhaps through a public-private partnership. Our ask to them is for large-scale, multi-year funding with almost complete flexibility. Our commitment is that we’re going to measure outcomes, and if the new system works, we will shift money from things that don’t work to things that do. These flexible funds will let us test an ambitious system redesign while uncovering which requirements still need to be reworked going forward.
Attempting to cobble together Medicaid reimbursements, assisted living waivers, housing vouchers, disability benefits, cash aid, and other funding sources often stymies local programs from creating responsive services, leaving a patchwork with many gaps. When the GPL helped Massachusetts design a permanent supportive housing project focused on housing individuals with long-term homelessness and complex medical conditions, one of the most pressing challenges the state faced was how to combine siloed funding sources into a set of flexible, adaptable services for people with complex needs. Ultimately housing over 1,055 individuals, the program used new philanthropic funding, Medicaid funds with a waiver to test new ways to pay for services, and other innovations to weave together new private dollars with existing resources to create a pool of flexible capital. These funds allowed providers to plug service gaps or supplement the cost of services, from things like building maintenance costs to on-site mental health support.
Know people by name: Create a list of individuals with SMI who are unsheltered. These are folks who likely have acute, highly individualized needs. Know who they are by name. The nonprofit Community Solutions has been working with jurisdictions across the country to help them build quality By Name Lists (BNL) as the foundational data needed to effectively address homelessness, so a pilot of this idea could start with a Community Solutions partner that already has a quality BNL. In a mid-sized city, this might be ~150-300 people. A large city like San Francisco where the Point in Time Count showed just under 3,000 people living in tents or on the street may want to start with an initial pilot of the 300 hardest-to-serve individuals or a specific geographic area.
When the GPL helped Denver design a Pay for Success project addressing chronic homelessness, we worked with the city to create a list of people who most needed pathways to permanent supportive housing—individuals who were experiencing chronic homelessness and had been cycling in and out of the criminal justice system and emergency care or detox. The city created a way to automatically match program eligibility with information from the police on contacts or arrests of chronically homeless individuals. This real time data then allowed the city to quickly share names of specific individuals who had a high need for care and were at high risk for criminal justice involvement directly with the two providers—the Colorado Coalition for the Homeless (CCH) and the Mental Health Center of Denver (MHCD). CCH and MHDC would proactively reach out to engage the client in a pathway to permanent housing. Tyler Jaeckel, who helped design the city’s program while at the GPL and now serves as the Chief Economic Recovery Officer for the State of Colorado, describes this close coordination as “a partnership between all parties to make sure there was the least amount of time between someone being contacted by police and being offered housing. This is different from what usually happens, where someone can languish on a referral list for months. Where had they been last? Who last had contact with them? The providers, the city, and the other partners learned how to work together to figure out who needed housing and how to get to them as quickly as possible.”
Ask providers to be accountable for people, not process: Radically redesign contracts to create integrated services that will give individuals a consistent touchpoint through their entire journey into housing and treatment. Instead of paying dozens of providers for fragmented street outreach, navigation, and referrals to other providers, offer a small number of large, flexible contracts to organizations willing to take full responsibility for individuals with SMI living on the street, from initial contact through stable housing placement and support. Grant each provider base operational funding that can cover the cost of access to housing units and wraparound services—you can’t ask providers to be accountable for supporting folks in staying housed if you aren’t going to provide enough money to house and support them.
Randomly assign clients from the list you created to providers and hold providers accountable for outcomes for those specific individuals, not for clients that they self-recruit. Providers might understandably argue that they have specialized in a particular population or geographic location of the city. However, if you want to be able to compare provider performance, you need to randomize clients to them rather than letting them self-select, and you need to hold them accountable for the outcomes for every person who has been assigned to them, not just the ones they found easy to work with. This is very different from standard practices and will raise lots of legitimate logistical concerns and questions but is core to running an outcomes-based program.
Consider generous performance bonuses rewarding stable housing placements at key intervals—3, 6, 12, 18 and 24 months—so there is a strong incentive not just to make initial placements, but to keep people housed.
As with any time you focus on and reward outcomes, there is a risk of distortion. Look for ways to independently validate client housing placement and wellbeing to ensure separate lines of feedback outside of data supplied by providers.
In the GPL’s Denver permanent supportive housing project, because there weren’t enough permanent supportive housing slots for all clients, clients were randomized onto one of two tracks of treatment. The first was the control group, which received the usual care services available in the community. The second pathway was the direct handoff between the city and CCH and MHCD. Once an eligible client was referred to one of the providers, that provider was responsible for immediately reaching out to find the client, engaging them, and supporting them throughout the full journey into supportive housing. CCH and MHCD not only provided access to permanent housing but also supportive services through a modified Assertive Community Treatment (ACT) model of intensive case management. The ACT teams provided bridge housing for clients who weren’t ready to enter permanent housing yet, helped clients track down vital documents, connected clients to behavioral health services, and worked to establish the trust to work through a daunting process. Critically, the providers acted as a consistent throughline from street outreach all the way to housing placement and support rather than bouncing clients between disconnected services.
Enable innovation and rapid learning: Your providers are going to be doing something new and hard. Remove all the barriers you can for them. Find creative ways to braid funding and remove specific funding requirements that block them from being able to serve someone from initial touchpoint through placement. To address the needs of high-acuity clients, they need to be able to test innovative, personalized solutions. They’re closer to the ground than you are—free them up to test peer supports, fast-tracked access to clinical care, or other approaches that may not have been possible under more restricted contracts. Put some basic guardrails in place against coercive or unethical actions and then allow providers to draw on their experience and proximity to try innovative solutions. Providers should be able to flexibly subcontract with other providers to bring together the full suite of resources they’ll need.
The Housing and Disability Advocacy Program (HDAP) administered by the California Department of Social Services exemplifies how this combination of flexible funding and evidence-based practices can work. The program serves people who have disabilities and are experiencing or at risk of homelessness through outreach, case management, housing-related financial assistance, and disability benefits advocacy. HDAP grantees are given highly flexible funding that can be used to help program participants address a wide variety of barriers to housing. Grantees are encouraged to engage people with lived experience on local program design and to adopt evidence-based practices, and are provided active technical assistance to continuously improve their program’s reach and impact, including lessons shared across the 74 county and tribal grantees operating HDAP programs. Despite serving a population with particularly complex needs, HDAP participants have shown more positive outcomes across a variety of measures compared to those receiving services through the broader local homeless response systems in the state.
Be a partner, not just a compliance enforcer: It will be a sign of things going well if the pressure to actually get people off the streets surfaces bottlenecks with more urgency. Providers will likely tell the city that the current supportive housing stock that's available is in the wrong sites or is so poorly maintained that they can't get people to stay there. They may discover that they need new pathways to outpatient treatment slots. The best performance contracts involve the government leaning in as a partner and taking responsibility for fixing the broader system. If your providers tell you they can’t get outpatient spots in time for their clients or can’t get an appointment with a doctor to help verify their disability, work with your hospital system or local clinicians to create a new referral pathway. If they think that housing sites are mismatched from the places clients are willing to live, work with them to find creative solutions to different sites. Their success is your success.
One of the major risks of a radical shift in the status quo is pushback from politically influential providers. And it’s a fool’s errand to pressure recalcitrant providers to take on a new approach that they strongly oppose. Don’t use contracting to try to mechanically force culture change. Begin with listening to and engaging providers. Success in shifting towards an outcomes focus depends on a handful of innovative, mission-driven nonprofit leaders who are willing to put their operational models and reputations on the line for an ambitious new vision.
Invest in ongoing improvement and shift resources going forward: One drawback of a fragmented system that measures on inputs instead of outcomes is that it can be very hard to tell the difference between providers who are effective and providers who talk a good game. Shifting to results-based contracts with rigorous measurement has a way of revealing who is good at their job. Over time, use measurement to identify who your high-performing providers are. One of the best ways to cut through explanations about why something can’t be done is by finding the providers who are proving it can be done. If designed in a thoughtful way, an outcomes-oriented structure can help incentivize fast knowledge sharing across providers as they learn from each other what practices are working. Over time, shift funding away from low-performing providers either toward higher-performing providers or to new entrants, creating more “pull” for new, innovative providers in your ecosystem. Like much of this advice, this seems so obvious as to barely be worth saying. Unfortunately, the reality is that this rarely happens. In many cities, service contracts are awarded to legacy providers who are well-known and politically powerful. It takes real leadership and willingness to pay a political price to shift resources away from low-performing providers. However, publicizing the performance of the providers over time may help build some of the political will needed to do this.
Tyler Jaeckel told me, “A rigorous evaluation of outcomes let us ask the question, ‘Is something really working?’ During this time when everyone is claiming different approaches work, a focus on defendable outcomes gives us a shared language for figuring out what actually delivers.”
Looking Forward
This exercise assumes a combination of visionary local leadership willing to upset the status quo, innovative providers who are excited to try something new, and committed philanthropists who can help bridge from the old funding structure to the new with flexible capital.
Evidence shows that this type of outcomes-focused, person-centered approach works better. In Denver, the city’s permanent supportive housing project was rigorously evaluated through a randomized controlled trial by the Urban Institute. Intuitively, the evaluation found that clients who were given permanent housing access and supportive services were more likely to be housed and less likely to be in shelter than the control group. But compared to the control group who received normal homelessness response services, clients in the program also had a 40 percent reduction in shelter visits, a 34 percent reduction in police contact, a 40 percent reduction in arrests, and a 65 percent reduction in detox facility usage. Participants in the program used emergency care less, yet were more likely to attend a psychiatric care office visit and to receive psychiatric medication.
What we call fragmentation is the logical result of funding mechanisms, incentives, and organizational habits that reify the status quo. But we’ve seen time and again that if we’re willing to shift from propping up a set of processes to instead focusing on a discrete group of people, empowering innovative service providers with the flexibility to design personalized solutions, and pairing tough accountability with real resources, we can deliver outcomes that otherwise seemed impossible.
Read more stories by Gloria Gong.