BC SealBoston College Magazine Winter 2005
In This Issue
Linden Lane
Works and Days
Letters to the Editor
Special Section
BCM Home
Contact BCM
Coming Events
. Features

Home economics


The Bush administration has proposed to end chronic homelessness within 10 years. Research by the author shows how and why it can be done.

Photo by Bruce Davidson/Magnum photos

Photo by Bruce Davidson/Magnum photos

By Dennis P. Culhane Ph.D.'90

HOMELESS ADULTS IN AMERICA CAN BE DIVIDED INTO THREE CATEGORIES. (I won't be discussing homeless families here, whose need for affordable housing is less distinguishable from the affordable housing problems affecting low-income families nationwide.) About 80 percent of homeless adults are transitionally homeless. According to a study my colleagues and I did in Philadelphia in the late 1990s, they are likely to be homeless only once over a three-year period, and only for an average of three weeks at a time. They leave the shelter system and don't come back—and with good reason.

For most people who will stay in a shelter tonight, the very first thing on their minds is making sure that they won't be there again tomorrow. In most shelters around the country, one has to register at seven o'clock in the evening for a bed, often standing in line for hours. Lights out at nine, and then it's up and out at six or seven in the morning, to spend the rest of the day meting out a meager existence in public spaces. Living in a public shelter is a humiliating experience, and nearly everyone who lands there does everything in his or her power to get out as fast as possible. We don't know where all of these people eventually end up once they leave homelessness, though many do work and most have families. But for a million or so men and women each year, thankfully, a stay in one of our nation's homeless shelters is a onetime and short-term experience.

The remaining two groups of homeless adults—the episodically and chronically homeless—account for roughly 10 percent each of the homeless adults who enter our country's shelters. The episodic homeless move in and out of the shelter system four or five times over the course of a three-year period, averaging a total of 90 days; they stay for 10 or 15 days at a time, then leave for five months at a time, before returning. Their pattern of shelter use is seasonal: Their presence grows in the winter, indicating that some of them are probably living outdoors, on the streets or in makeshift dwellings. They may also have alternative places to stay, which they access when they can (family, friends, partners, even hospitals) or when they are forced to (prisons, jails).

By contrast, when the chronically homeless come into the shelter system, they tend to stay for very long periods of time; some never leave. In Philadelphia, they stay in shelters for close to a year, on average. In New York City, the average is 750 consecutive days. This small subset of the poor amounts to between 150,000 and 200,000 individuals nationwide—and it is for them that the Bush administration and longtime advocates for the homeless have proposed (with surprising accord) a permanent and cost-beneficial solution to homelessness.

Emergency shelters were never designed as permanent housing. Yet half of the adult shelter population on any given night consists of individuals who are chronically homeless, and who effectively live in shelters. In Philadelphia, that is about 800 to 1,000 people. In New York City, it's about 3,000 people.

The chronically homeless are a very expensive group. In shelter costs alone, New York spends nearly $100 million a year on its long-term shelter population. Nationally, I estimate that as much as $2 billion may be spent providing "permanent housing"—in the form of shelter cots—to chronically homeless people, at a cost of about $15,000 per cot per year.

What do we know about the chronically homeless? Using data from Veterans Affairs and Medicaid, we see at least one clear trend: Ascending from transitional to episodic to chronic, the proportion of homeless who suffer from severe mental illness or serious medical conditions increases dramatically. Nearly all of the chronically homeless have some disability, be it a physical handicap, or a mental health or substance abuse problem. The chronic homeless also tend to be older.

DIFFERENT HOMELESS populations require different policies. The transitionally homeless, for instance, possess some individual and social resources that enable them to get out of the shelters and to stay out. Programs and policies aimed at them should facilitate as quickly as possible their relocation and stabilization within the community—where most of them are returning anyway. These programs should provide emergency cash assistance, relocation services, and, when necessary, help in getting a job or job training.

The episodic homeless, many of whom are on the street, need more deliberate and targeted engagement by treatment professionals, and a less intimidating environment than that which exists in most large shelters. Success has been achieved around the country in getting many people living on the streets into so-called "safe havens," which unlike regular shelters don't require sobriety, and provide some privacy in a small and unstructured setting. The programs are kept small to provide a sense of security; rules are kept to a minimum to reduce the fear of being harassed or thrown out by shelter "guards"; and support is provided by treatment professionals, as a trusting relationship is formed. Once individuals are stabilized in safe havens, most are ready and able to move to a permanent housing placement where supportive services are provided—in some cases, to form a community of tenants from among individuals who once struggled with homelessness.

This has been a very effective strategy for reducing the street population in several cities. In Philadelphia, for instance, police and outreach workers conduct a one-night count of the city's street population each month. At one point in 1998, there were 850 people on the street. By August 2003, after the expansion of safe havens and an aggressive outreach effort, officials estimate that the number of people on the street dropped to below 125.

The solution to chronic homelessness among people in shelters, on the other hand, is direct placement in permanent supportive housing. Most of the chronically homeless are stable, even if they are disabled. They are capable of stable residency, as their long and continuous stays in shelters—some of the least habitable housing anywhere—demonstrates. Policies should promote a view of this population, first and foremost, as potential long-term housing tenants. With access to supportive services, usually in the form of teams that include case managers, nurses, and social workers, they can live independently. Such "supportive housing" programs, which exist already to varying degrees throughout the country, have shown great success in resolving homelessness among people long thought to be the toughest population to serve. In New York City, chronically homeless individuals have been placed in converted hotels and in market-rate apartments scattered around the city; in Phoenix, they have been moved into clustered apartments; and new projects in San Francisco include master leases on apartment buildings by the city's Department of Public Health. In some places, tenants are asked to sign a contract with the housing agency to make sure that they remain engaged in the social services network, or at least to ensure that if things start to unravel, there's a network set up to catch them. Many of the residents like to have this safety net. In such programs around the country, success rates at housing the chronically homeless have been high, with 80 percent retention after two years. Of the 20 percent who do leave, half fall out because of behavioral or legal problems, including a small subset who return to homelessness; half move on to better or alternative permanent housing.

There are sound economic reasons to focus on the chronically homeless. They are the most expensive users not just of emergency shelter, but also of other public systems—in particular, health care and corrections. My colleagues and I recently evaluated a supportive housing initiative in New York City—the New York/New York Agreement begun in 1989—the largest such program in the country. We found the cost of placing the chronically homeless in permanent housing to be offset, or quite nearly so, by the savings that resulted in other social service systems.

ANYONE WHO visited New York City in the late 1980s will remember that homeless people were nearly everywhere. There were 750 people living in one park in lower Manhattan. Virtually every subway station harbored them. Entire new authorities were created around Times Square and Penn Station to deal with the hundreds of people living in and around the train stations. The issue of homelessness consistently polled as the number one concern of the public. And it consistently polled as the number one concern of local businesses, which feared the negative impact on tourism.

In 1989, the mayor and the governor, David Dinkins and Mario Cuomo, struck a deal to build a substantial amount of housing for homeless people with mental illness. Each put up half of $220 million in capital, and each agreed to come up with a certain portion of the $65 million annual operating costs for the housing and the services. Between 1990 and 1995, they built some 3,600 units of housing targeted specifically for people who were both homeless and mentally ill.

A parallel event occurred about that time which doesn't usually get as much attention, but which was just as important to reducing homelessness in New York City. In 1992, advocates for people with AIDS successfully sued the city, arguing that AIDS sufferers should not have to live in a homeless shelter when they become indigent. The case established a right that exists nowhere else in the country: If you lose your home in New York City and have AIDS, the city must find emergency private housing for you within 24 hours—as well as some form of subsidized housing for the long term. This ruling eventually led to nearly 27,000 New Yorkers with AIDS living in subsidized housing, many of whom were homeless or nearly so.

It's become an urban legend that after his election as mayor in 1993, Rudolph Giuliani dramatically reduced the homeless street population in New York City by arresting the squeegee men (the ragtag men who swarmed cars at stoplights to wash windows and collect a fee—and who, as it happens, were mostly not homeless). In fact, it was this huge investment in housing, particularly housing for people with mental illness or AIDS, that made the real difference.

Advocates for the homeless have been claiming for years—taking out full-page ads in the New York Times and elsewhere—that it is cheaper to give the homeless a home than to put them in a shelter or hospital or prison. But they didn't have empirical evidence to support that conviction. My colleagues and I set out to learn if it was true, by studying the economic effects of the New York/New York Agreement.

To begin with, we identified every individual who had been placed in housing through the city/state agreement, from 1989 to 1997, which totaled 4,679 people. In addition, we collected data on shelter users whose demographic characteristics and homeless histories matched the housing recipients (they formed our control group). We also collected data from Veterans Affairs, state and municipal hospitals, Medicaid, and the state and city departments of corrections on the people who stayed in their systems. We spent four-and-a-half years gaining access to these records. Our final study population—cases and controls—consisted of nearly 10,000 homeless individuals.

We found that the annual cost of being chronically homeless in New York City, in the absence of housing, was nearly $40,500 per year. In a typical two-year period, the average homeless person with mental illness stayed four-and-a-half months in a shelter, two months in a state psychiatric hospital, two weeks in a public hospital, a month in a private hospital paid by Medicaid (in addition to two months' worth of outpatient visits), a week in a VA hospital, 10 days in prison, and 10 days in jail.

About 85 percent of the actual dollar costs incurred per year—that is, nearly $35,000—were for days spent in hospitals. Prison and jail costs amounted to about $1,000 (only 12 percent of the homeless spent any time at all in prison or jail).

And what happened when the chronically homeless were placed in housing? They reduced their use of those other publicly funded systems by just over $16,200 a year, on average. They were hospitalized less frequently, and they got arrested less frequently. When they were hospitalized or incarcerated, they didn't stay as long, because they had somewhere to be discharged to.

The one area of service where costs went up was outpatient Medicaid, and that was expected. People weren't going to the emergency room or checking into the hospital for routine medical care when they could obtain outpatient care and recuperate in their own homes. In all, about 85 percent of the savings that accompanied housing placements came in health care, from reduced time in hospitals.

And what was the actual cost of the housing? Taking into consideration the debt service, operating expenses, and the cost of supportive services, my colleagues and I estimated an annual per unit cost of $17,200. Subtract from that the roughly $16,200 in savings that resulted from placement in such a unit, and the net cost of providing housing turned out to be roughly $1,000 per housing unit.

Put another way, of the $65 million that the city and state of New York spend each year on this program, they get $59 million back in terms of reduced use of public services. And these estimates of savings are in fact low. My colleagues and I did not include every publicly funded service that the chronically homeless would have otherwise used—including some proportion of the federally funded homeless services that are provided in shelters but are not tracked by the government agencies we surveyed. We did not include, as economists would have, estimates of the costs to crime victims that might be avoided, or the reduced costs incurred by the courts or police who must deal with the effects of street homelessness. And we didn't put a dollar value on the benefits associated with the fact that some chronically homeless people who receive supportive housing eventually end up getting a job, paying taxes, and reconnecting with their families and their children.

THERE IS ample reason to be optimistic about the New York/New York model as a way of solving chronic homelessness. If we house the 150,000 to 200,000 people who are chronically homeless in the country today, we can be fairly sure they will not be replaced by large numbers of new homeless in a few years' time. Why? First of all, as noted above, supportive housing vacates at a rate of 20 percent every two years. With 150,000 total units, that would translate into 30,000 vacancies available for new people each year.

Furthermore, the homeless population is aging. Looking at data from New York City, for instance, we find that the average age of the adult homeless population in 1987 was 28—a figure that came as a shock at the time, because most people's image of the homeless was based on the elderly men on "Skid Row" during the 1950s and 1960s. Since the 1980s, however, the homeless population has continued to age—at a rate of more than 10 months per year. The average age now is about 42. Close to 70 percent of the adult homeless are persons born in the years between 1950 and 1964—otherwise known as the Baby Boom era.

In short, a large component of the homelessness problem among single adults is in fact a "cohort" problem, in the demographic sense of the word. Twenty-five years before this generation would bear down on the Social Security and Medicare systems, poor young-adult Baby Boomers, including some who suffered from severe mental illness, busted through the country's social safety nets, which were ill-prepared for them.

We now face a stark choice. In 15 years, we can have a chronically homeless population consisting mainly of elderly people, just as several decades ago we had the skid row homeless (another birth cohort, whose roots were in the Great Depression). Or we can build a housing capacity for these people now and maintain that capacity for the generations of dependent adults who will follow.


Dennis P. Culhane earned his Ph.D. in psychology at Boston College in 1990 with his dissertation "On Becoming Homeless." He is a professor, and senior fellow at the Leonard Davis Institute of Health Economics, at the University of Pennsylvania. This article is drawn from a talk he delivered at BC, which was part of the Graduate School of Arts & Sciences' celebration of its 75th anniversary.


. . .
Alumni Home
BC Home