Many communities will forgo a snapshot count of people experiencing homelessness this year. Advocates see a chance to collect better data.
Every January, dozens of volunteers come together to count the number of people experiencing homelessness in Sonoma County, California.
Led by paid guides who are themselves unhoused, they pile into cars before sunrise, then scatter across a landscape of subdivisions and vineyards that spans a region slightly larger than the state of Rhode Island. The effort, mandated by the U.S. Department of Housing and Urban Development, is meant to provide a rough sketch of the state of homelessness at a particular moment. Similar efforts play out nationwide at least every other year.
But this year is different.
Sonoma County is calling off its point-in-time (PIT) count because of the risks posed by the novel coronavirus, and it’s not alone. So far, 175 Continuums of Care (CoCs) — local bodies that coordinate services for homeless families — have asked for permission to postpone or cancel the PIT count this month, according to HUD. That means about 45% of jurisdictions designated as CoCs won’t be providing an estimate, including major metropolitan areas that have struggled with homelessness, such as King County (home to Seattle) and Los Angeles.
This year’s incomplete count underlines one of the PIT assessment’s biggest weaknesses: its infrequency. While some places like Sonoma have a yearly count, the government only mandates a tally for the total unhoused population every two years. Advocates see an opportunity for a broad pivot toward collecting data on homelessness much more frequently, which would allow communities to gain more insight into who needs shelter and services, and where they’re located. The PIT count can miss people who may be sleeping in an abandoned building, for example, and its timing in January may not give a picture of homelessness in warmer months.
Community Solutions, a New York City-based nonprofit aiming to end homelessness, has put data at the center of its “Built for Zero” campaign. It’s a sort of Vision Zero strategy for housing instability: a movement to ensure that government agencies, shelters, caseworkers, clinics and nonprofits all have access to the same data as they assess and manage housing instability. There’s evidence that it’s working: Since 2015, more than 80 communities have taken up the Built For Zero pledge, an administrative commitment to reduce homelessness to a standard called “functional zero.” So far, 14 of them have essentially eliminated chronic homelessness, veteran homelessness or both.
“If we’re really going to drive down to functional zero where homelessness is rare and brief, a community’s functional response system needs to have data to do that,” John Vu, vice president of strategy for community health at Kaiser Permanente, which is partnering with Community Solutions on the Built for Zero model in 28 of its communities. A central database puts names and faces to individual cases that are abstract in point-in-time counts.
A disciplined approach to data collection and assessment helped Kern County, home to Bakersfield, California, functionally end chronic homelessness over the last year, despite the economic ravages of the pandemic. The Bakersfield Kern Regional Homeless Collaborative, which serves as the CoC there, relied on data initially updated once a month to set goals, determine whether its strategies were working, and create a feedback loop to streamline its process, according to Heather Kimmel, the assistant executive director for the local housing authority. They eventually determined those updates were more useful when performed weekly.
“Without the data, we would not have known if our ideas were effective, when to scale up, and when to try something else,” Kimmel said in an email. From there, the collaborative and its member agencies were able to manage cases, and a public-private partnership with landlords helped to provide those experiencing chronic homelessness with the housing they needed.
Even the PIT count, for all its flaws, provides an idea of what homelessness looks like at both the local and national levels — something that’s especially critical this year as the pandemic continues to strain the U.S. economy. A report published this month by the nonprofit Economic Roundtable estimates that an additional 603,000 individuals could be unhoused nationwide by 2023.
“The problem of unsheltered homelessness appears to be getting worse in a lot of places due to the economic downturns and the health downturns of the coronavirus, so we’re very concerned about that,” says Steve Berg, vice president for programs and policy at the National Alliance to End Homelessness.
The last nationwide PIT count in January 2019 identified nearly 568,000 homeless people during the snapshot tally, about two-thirds of whom were sheltered. But the methodology and the type of data can vary widely from place to place. The National Coalition for the Homeless has observed that PIT counts mostly miss people staying in supportive housing, a population that in 2017 numbered more than 503,000 — suggesting that any meaningful figure is off by approximately half. A recent op-ed in the Chicago Tribune laid out reservations regarding both the PIT count’s accuracy and ethical concerns about how it’s conducted.
The Built for Zero model adopted by places including Bakersfield further shows the limitations of the HUD model, which has served as a backbone of federal efforts to address homelessness. This year, the Bakersfield Kern Regional Homeless Collaborative received an exception from HUD to use the real-time data they gather instead of a January count. Guidance from the department allowing for the alternative numbers sends a signal that there are ongoing conversations around transitioning to more timely data nationwide, Built for Zero co-director Beth Sandor says.
“It’s a huge opportunity in this moment to say the point-in-time count is an outdated approach and it is time for us to have real time data in every Continuum of Care, in every community across the country,” Sandor says.