The Washington PostDemocracy Dies in Darkness

The surprising holes in our knowledge of America’s homeless population

The White House begins a homeless roadshow. But the streets tell a complex story.

A man asks for money along Wall Street in New York in September. (Spencer Platt/Getty Images)

President Trump seems close to declaring a war on homelessness or, depending on whom you ask, a war on the homeless. On the way to California on Tuesday, he griped that people living on the streets are devaluing prestigious addresses, and he vowed to do something about it.

But what do we actually know about the population that, according to the Post’s Philip Rucker and Jeff Stein, Trump says are living on the “best highways, our best streets, our best entrances to buildings?"

Homeless America defies easy characterization. It has more residents than some rural states — more than 550,000 Americans experience homelessness on a typical night, and 1.4 million will spend some time in a shelter in a given year.

Government surveys tabulate everything from annual population estimates to the unemployment rate. But they focus on people in households and, in some limited cases, homeless shelters. They miss the 200,000 or more people who live outdoors.

To measure people whom officials often call the “unsheltered homeless,” we turn to the Department of Housing and Urban Development. HUD’s “Point in Time” counts target the nation’s homeless population when they’re most likely to seek shelter and easier to track down — at night, toward the end of January. At that time, volunteers in hundreds of cities fan out to locate their peers and neighbors living outside, or in temporary spots such as cars and tents, and ask survey questions about who they are and how they are doing. HUD requires the outdoor counts at least once every two years. Many organizations do them annually.

The volunteers helping with those counts often have deep experience with a community’s local homeless population, but in just one night they can’t be expected to canvass every street, peer into every possible living place, or persuade everyone to answer. Studies estimate that New York City counts may miss up to half of the unsheltered population, for example. This year, it stood at about 3,700 — which would already make it one of the highest levels outside of California.

According to HUD’s headline number: Homelessness in America fell steadily after 2010, but it ticked upward from 2016 to 2018. That’s a broad statement — it would be more accurate to say, New York and Los Angeles each rose by enough to offset a decline in the rest of the country in the past few years.

Those two cities (or the counties that contain or comprise them), are home to 6 percent of the U.S. population, but they held about a quarter of the country’s sheltered homeless and a third of those who lived on the streets in 2018. Trends in either city can swamp movement in the rest of the nation.

And neither trend is as simple as it looks. New York’s homeless population increased, but that could be because the city stepped up efforts to house the homeless and thus converted hard-to-count people living on the streets into easier-to-count people in shelter programs. It appeared as though the city’s homeless population climbed, but it may just be counting better.

In Los Angeles, on the other hand, the rise in the homeless population came largely in the streets. It reflects a visible crisis that, Rucker and Stein report, has prompted Trump officials to consider drastic measures, such as razing tent camps and moving homeless people to temporary or refurbished government facilities.

As Trump prepares big push on homelessness, White House floats new role for police

Daniel Flaming of the Los Angeles-based Economic Roundtable said the margin of error in these counts would be large — if it were published at all. The counts are conservative, and they provide a low-end estimate.

In many other cities, including Washington, homelessness has fallen in recent years, HUD data show. In the case of cities such as Washington, Chicago, Denver and Philadelphia, that mostly reflected a decline in the population in shelters and transitional housing. The number of people on the streets rose in all four cities from 2016 to 2018.

The HUD records don’t include people who are staying with someone else on a temporary basis, among other groups. Data show that men, black Americans, the mentally ill, domestic-violence survivors, substance abusers and veterans all experience homelessness at higher rates.

Alternative sources include the once-every-decade census and student-related data from the National Center for Education Statistics, but they aren’t comprehensive or frequent enough for our purposes. Instead, much of what we know about Americans experiencing homelessness comes from a vast body of published research.

“While we can’t reasonably expect to count every homeless person in the country, sociologists and economists can often follow smaller populations very closely within the confines of a carefully designed experiment,” said Krista Ruffini of the University of California at Berkeley. “From these studies, we can learn important lessons about the characteristics of homeless populations and how best to serve them, lessons which can then be applied in similar communities nationwide.”

Ruffini, along with Notre Dame’s William Evans and David Phillips, recently reviewed hundreds of articles about the costs, causes and solutions to homelessness, and circulated their results recently as a working paper from the National Bureau of Economic Research. Among many other things, they explored the link between rent costs and homelessness.

“Higher rents are correlated with higher rates of homelessness, which we might expect,” Ruffini said. “But we found little correlation between rising rents and rising homelessness, making it hard to attribute changes in homelessness — including the increases in NYC and L.A. — to changes in housing prices alone.”

In one California county, the public cost of homelessness worked out to about $83,000 for each “persistently homeless” person, according to a 2015 report by Flaming and his collaborators. His team linked 25 million records produced between 2007 and 2012 by the justice and health-care systems, social services, nonprofit groups and housing agencies for almost 105,000 people who were, at some point during that period, homeless in California’s Santa Clara County.

HUD Secretary Ben Carson, other top Trump officials to visit California as White House ramps up homeless push

“The episodic, snapshot nature of the point-in-time count can lead people to think of homelessness as a problem that reinvents itself every year,” Flaming said. “But the dynamics, and too often the same individuals, persist over time.”

At the southern end of California’s San Francisco Bay, Santa Clara ranks one of the nation’s most prosperous counties, because of Silicon Valley cities such as Palo Alto, Mountain View and San Jose. It’s also one of its most homeless, slotting in behind only Los Angeles and Seattle, with 5,448 people living on the streets in 2018, HUD data show.

Most (53 percent) of the money spent on the homeless went to health care, the team found. About a third (24 percent) went to criminal justice, particularly jails, and 13 percent was spent on social services. Almost half of medical costs (47 percent) were spent on just 5 percent of the homeless population. On average, a person experiencing homelessness incurred annual medical fees of $5,148; those in the top 5 percent crossed the $100,000-a-year barrier.

The team found that, if a program targeted the right people, it could prevent or reduce so many expenses that it would offset the cost of housing.

Trump officials tour unused FAA facility in California in search for place to relocate homeless people

HUD’s huge Family Options experiment tested several such programs. Researchers assigned 2,282 families in emergency shelters some type of assistance, ranging from the usual array of waiting lists to accelerated housing options including long-term subsidized housing.

The families getting those long-term subsidies, which empowered people to find their own places on the private market, fared best. They put 30 percent of their income toward rent and the government made up the difference. Such families were half as likely to be homeless or doubled up in someone else’s home. They reported less psychological stress. Their children missed fewer days of class, switched schools less often and got in less trouble in school.

“Homelessness is a symptom of a systemwide problem,” said Claudia Solari, an author of the report who is now with the Urban Institute in Washington. “Society often blames the individual and sees homelessness as a personal problem, but this ignores the role that our social institutions play, such as not providing enough affordable housing.”