By Lucius Couloute
It’s hard to imagine building a successful life without a place to call home, but this basic necessity is often out of reach for formerly incarcerated people. Barriers to employment, combined with explicit discrimination, have created a little-discussed housing crisis.
In this report, we provide the first estimate of homelessness among the 5 million formerly incarcerated people living in the United States, finding that formerly incarcerated people are almost 10 times more likely to be homeless than the general public. We break down this data by race, gender, age and other demographics; we also show how many formerly incarcerated people are forced to live in places like hotels or motels,⤵ just one step from homelessness itself.
The transition from prison to the community is rife with challenges. But before formerly incarcerated people can address health problems, find stable jobs, or learn new skills, they need a place to live.
This report provides the first national snapshot of homelessness among formerly incarcerated people, using data from a little-known Bureau of Justice Statistics survey. Our analysis builds on existing research showing that past incarceration and homelessness are linked. National research suggests that up to 15% of incarcerated people experience homelessness in the year before admission to prison.1 And city- and state-level studies of homeless shelters find that many formerly incarcerated people rely on shelters, both immediately after their release and over the long term.2
We find that rates of homelessness are especially high among specific demographics:
In the following sections, we take a closer look at these populations. We also break down how many formerly incarcerated people are living in marginal housing⤵ - a step away from homelessness.
We find that people experiencing cycles of incarceration and release - otherwise known as the “revolving door” of incarceration - are also more likely to be homeless.3
People who have been to prison just once experience homelessness at a rate nearly 7 times higher than the general public. But people who have been incarcerated more than once have rates 13 times higher than the general public. In other words, people who have been incarcerated multiple times are twice as likely to be homeless as those who are returning from their first prison term.
Unfortunately, being homeless makes formerly incarcerated people more likely to be arrested and incarcerated again, thanks to policies that criminalize homelessness.4 As law enforcement agencies aggressively enforce “offenses” such as sleeping in public spaces, panhandling, and public urination - not to mention other low-level offenses that are more visible when committed in public - formerly incarcerated people are unnecessarily funneled back through the “revolving door.”
Previous research has shown that formerly incarcerated people are most likely to be homeless in the period shortly after their release.5 Our data supports this research: We find that people who spent two years or less in the community were more than twice as likely to be homeless as those who had been out of prison for four years or longer.
Homelessness among recently released individuals is a fixable problem. States can - and should - develop more efficient interagency systems to help formerly incarcerated people find homes. But longer-term support is also needed: Our analysis found that even people who had spent several years in the community were 4 times more likely to be homeless than the general public.
Within the broad category of homelessness, there are two distinct populations: people who are sheltered (in a homeless shelter) and those who are unsheltered (without a fixed residence).
We find - in keeping with previous research on homelessness in the general public - that the sheltered and unsheltered formerly incarcerated populations have significant demographic differences.6
For example, we find important differences by gender. Overall, formerly incarcerated women are more likely to be homeless than formerly incarcerated men. But among homeless formerly incarcerated people, men are less likely to be sheltered than women, whether for reasons of availability or personal choice.
|Homeless(Rate per 10,000)||Sheltered(Rate per 10,000)||Unsheltered(Rate per 10,000)|
We find that formerly incarcerated Black men have much higher rates of unsheltered homelessness than white or Hispanic men.
The data also suggests that women of color experience unsheltered homelessness at higher rates than white women. (Though there were too few unsheltered formerly incarcerated Black and Hispanic women in our dataset to analyze,7 the rate of unsheltered homelessness among white women was substantially lower than the rate for women generally. Therefore, it is clear that formerly incarcerated Black and/or Hispanic women experience unsheltered homelessness at significantly higher rates than white women.)
|Black(Rate per 10,000)||Hispanic(Rate per 10,000)||White(Rate per 10,000)||Total(Rate per 10,000)|
Black women experienced the highest rate of sheltered homelessness - nearly four times the rate of white men, and twice as high as the rate of Black men. Combined with our breakdowns of race and gender separately (see Figure 1), this analysis shows that Black women face severe barriers to housing after release.
The high rates of homelessness among Black women are especially striking in light of our similar finding, last month, that unemployment rates among formerly incarcerated Black women were higher than any other demographic group.8 Our findings illustrate that Black women, in particular, have been excluded from the social resources necessary to succeed after incarceration.
Measuring homelessness among formerly incarcerated people is a critical step forward, but it doesn’t fully capture the exclusion of formerly incarcerated people from stable housing - the kind of housing most people need to thrive and contribute to their communities.
To better measure the scope of the problem, we created a second metric - housing insecurity - that includes formerly incarcerated people who are homeless (both sheltered and unsheltered) as well as those living in marginal housing like rooming houses, hotels, or motels.9
Housing insecurity provides a more realistic measurement of the number of formerly incarcerated people denied access to permanent housing. While we found that 203 out of every 10,000 formerly incarcerated people were homeless, nearly three times as many - 570 out of every 10,000 - were housing insecure.
We also uncovered notable demographic differences by expanding our view to the housing insecure population: Hispanics, for example, were more likely than people of any other race to live in marginal housing. Men had much higher rates of marginal housing than women, resulting in high rates of housing insecurity. And older formerly incarcerated people experienced the highest rates of housing insecurity.
Ideally, this report would directly compare the prevalence of housing insecurity among formerly incarcerated people to that of the general public. Unfortunately, the equivalent national statistics on housing insecurity do not yet exist. Even without that comparison, however, it’s clear that having been to prison is a major risk factor for housing insecurity.
Stable housing is the foundation of successful reentry from prison. Unfortunately, as our data show, many formerly incarcerated people struggle to find stable places to live. Discrimination by public housing authorities and private property owners,10 combined with affordable housing shortages,11 continues to drive the exclusion of formerly incarcerated people from the housing market.
Part of the problem is that property owners and public housing authorities have the ability to implement their own screening criteria to determine if an applicant merits housing12 - a process that often relies upon criminal record checks as the primary source of information. In practice, this means local authorities and landlords have wide discretion to punish people with criminal records even after their sentences are over.
The use of credit checks, exorbitant security deposits, and other housing application requirements - such as professional references - can also act as systemic barriers for people who have spent extended periods of time away from the community and out of the labor market.13
Excluding formerly incarcerated people from safe and stable housing has devastating side effects: It can reduce access to healthcare services (including addiction and mental health treatment),14 make it harder to secure a job,15 and prevent formerly incarcerated people from accessing educational programs.16 Severe homelessness and housing insecurity destabilizes the entire reentry process.
Fortunately, on-the-ground advocates across the country have made important progress in reducing overall homelessness.17 But an estimated 550,000 people are still homeless on any given night in the United States,18 many of them individuals with a history of criminal justice system contact. It’s critical that policymakers develop comprehensive responses to this problem, rather than continuing to punish those without homes.
All people - and particularly those carrying the stigma of criminalization - need these solutions. In such a wealthy country, it’s time we eliminate homelessness for good.
This report provides the first national estimates of homelessness among formerly incarcerated people, but these estimates likely understate the problem. Because the effects of intermittent homelessness last longer than your last night on the street, the best measures of homeless include those who have experienced homelessness in the last year. However, there is not yet a way to calculate this fuller picture of homelessness among formerly incarcerated people.19
Nevertheless, our findings make it clear that the 600,000 people released from prisons each year face a housing crisis in urgent need of solutions. State and local reentry organizations must make housing a priority, and provide additional services thereafter - a strategy known as “Housing First.”20 If formerly incarcerated people are legally and financially excluded from safe, stable, and affordable housing, they cannot be expected to successfully reintegrate into their communities.
Excluding formerly incarcerated people from stable housing harms not only individuals, but public safety and the economy at large. State- and city-level policymakers have the power to solve this housing crisis:
To the extent possible, this report uses terms commonly found in the literature on homelessness in the United States. However, given the limitations of the data set we used, the terms and definitions used in this report are not always consistent with those used by the U.S. Department of Housing and Urban Development (HUD), which is the data source we use for comparisons with the general public. Appendix Table 1, below, summarizes the differences between the terms used in this report and terms used by HUD.
|Term||Definitions used in this report||Equivalent Department of Housing and Urban Development (HUD) definitions|
|Homelessness||Includes people who reported their current, usual residence as:
||“Literally homeless” includes any “Individual or family who lacks a fixed, regular and adequate nighttime residence, meaning:
|Sheltered Homelessness||Includes people who reported that they currently live in a shelter most of the time. The type of shelter was not specified.||Includes individuals and families “who are staying in emergency shelters, transitional housing programs, or safe havens.”|
|Unsheltered Homelessness||Includes people who reported that they are currently homeless or have no fixed residence most of the time.||Includes people “whose primary nighttime location is a public or private place not designated for, or ordinarily used as, a regular sleeping accommodation for people (for example, the streets, vehicles, or parks.)”|
|Marginal housing||Includes people who reported currently living in a rooming house, hotel, or motel most of the time. Unlike the HUD definition, this is not exclusive to those whose housing is being paid for by charitable organizations or government programs.||HUD does not use this term, but includes people living in hotels and motels paid for by charitable or government programs in its definition of “literally homeless.”|
|Housing Insecurity||A combined measure that includes people experiencing sheltered and unsheltered homelessness, and people living in marginal housing.||HUD does not use this measure, but includes children living in a hotel or motel due to lack of alternative adequate accommodations in its description of “Additional Forms of Homelessness and Housing Instability,” using data from the U.S. Department of Education. Other living situations included in HUD’s analysis of additional forms of homelessness and housing instability include:
|Source||National Former Prisoner Survey (2008)||U.S. Department of Housing and Urban Development (HUD) Resources:
Appendix Table 2, below, summarizes all of our findings on housing from the National Former Prisoners Survey. Note that the “general public” rates come from our calculation of HUD homeless counts and Census Bureau population estimates for 2008, and that all data is reported as rates per 10,000 population.
|Sheltered Homeless (per 10,000)||Unsheltered Homeless |
|Sheltered & Unsheltered Homeless |
|Living in rooming house, hotel, or motel |
|Total housing insecure
|General public||All general public||13||8||21||n/a||n/a|
|Formerly incarcerated||All formerly incarcerated||98||105||203||367||570|
|Race or ethnicity||Black||117||123||240||358||598|
|Race and gender||Black men||108||124||233||369||602|
|Age||24 and under||52||91||143||132||274|
|45 or older||124||134||258||607||865|
|Time in prison||Less than 12 months||127||151||278||368||646|
|120 months or longer||69||n/a||n/a||409||n/a|
|Year released (Years since release)||2007-2008 |
(less than 2 years)
|2004 or before |
(4 or more years)
|Prior history||Incarcerated more than once||136||143||279||434||713|
|Incarcerated only once||67||74||141||312||453|
Figure 4, below, explains our method of calculating “housing insecurity.” Housing insecurity captures the full extent to which formerly incarcerated people lack stable housing, even if they are not literally homeless. We define this term in more detail below:
This report’s analyses of homelessness and housing insecurity are primarily based on our analysis of an underutilized government survey, the National Former Prisoner Survey, conducted in 2008. The survey was a product of the Prison Rape Elimination Act, and mainly asks about sexual assault and rape behind bars, but it also contains some very useful data on housing.
Because this survey contains such sensitive and personal data, the raw data was not available publicly online. Instead, it is kept in a secure data enclave in the basement of the University of Michigan Institute for Social Research. Access to the data required the approval of an independent Institutional Review Board, the approval of the Bureau of Justice Statistics, and required us to access the data under close supervision.
The practicalities of having to travel across the country in order to query a computer database limited the amount of time that we could spend with the data, and other rules restricted how much data we could bring with us. Additionally, if the number of respondents falling within any one group was too small, we were not allowed to export the data for that group due to privacy concerns.
Using this survey data, we were able to produce the first national estimates of homelessness among formerly incarcerated Americans. We also uncovered many other questions, which we do not yet have the necessary data to answer on a national level, but which suggest avenues for further research:
Even so, we believe that the analyses presented in this report begin to illuminate the severe housing-related inequalities experienced by criminalized people.
We used the National Former Prisoner Survey (NFPS) as our main data source for measuring homelessness and housing insecurity among formerly incarcerated people. This survey began in January 2008 and concluded in October 2008, and was derived from the Prison Rape Elimination Act of 2003, which mandated that the Bureau of Justice Statistics investigate sexual victimization among formerly incarcerated people.
The NFPS dataset includes 17,738 adult respondents who were previously incarcerated in state prisons and under parole supervision at the time of the survey. Individual respondents were randomly selected from a random sample of over 250 parole offices across the United States.
It is important to note that because this survey was given to people on parole, it is not a perfect tool to measure homelessness and housing insecurity among all formerly incarcerated people. Some incarcerated people are released without supervision, and their ability to attain stable housing may be different than those on parole. Previous research suggests, however, that parole officers have a minimal (or at best, inconsistent) effect on post-release housing stability. A national survey of state parole agencies in 2006 found that most - 60% - had no housing assistance program. Two regional studies of post-release shelter use, meanwhile, had conflicting findings: In New York, parole increased the likelihood of shelter use, but it appeared to reduce shelter use in Philadelphia.21 These mixed results are unsurprising: A synthesis of the literature explains that there is “little collaboration among [corrections and social service] systems and little consistency over time. What results is a prisoner reentry system that is disconnected from the housing and homeless assistance services system and from the neighborhoods where released prisoners live.” Future research should more closely examine the effect of supervision on homelessness and housing stability.
We drew upon specific NFPS survey questions for this report:
To measure homelessness in the general public, we used the Department of Housing and Urban Development’s Point-in-Time counts of sheltered and unsheltered homeless people, along with Census Bureau population estimates. This data is from 2008, the most recent year in which comparable data for formerly incarcerated people exists.
There is one minor, but notable, difference between HUD’s Point-in-Time counts (which we used to calculate homelessness in the general public) and our NFPS data (which we used to calculate homelessness among formerly incarcerated people). HUD’s Point-in-Time counts relied upon special local groups, called Continuums of Care, to record and report the total number of sheltered and unsheltered homeless people during the last 10 days in January 2008. The National Former Prisoner Survey, conversely, asked subjects about their housing status directly.
The non-profit, non-partisan Prison Policy Initiative was founded in 2001 to expose the broader harm of mass criminalization and spark advocacy campaigns to create a more just society. The organization is known for its visual breakdown of mass incarceration in the U.S., as well as its data-rich analyses of how states vary in their use of punishment. The Prison Policy Initiative’s research is designed to reshape debates around mass incarceration by offering the “big picture” view of critical policy issues, such as probation and parole, women’s incarceration, and youth confinement.
The Prison Policy Initiative also works to shed light on the economic hardships faced by justice-involved people and their families, often exacerbated by correctional policies and practice. Past reports have shown that people in prison and people held pretrial in jail start out with lower incomes even before arrest, earn very low wages working in prison, and face unparalleled obstacles to finding work after they get out.
Lucius Couloute is a Policy Analyst with the Prison Policy Initiative and a PhD candidate in Sociology at the University of Massachusetts Amherst, his dissertation examines both the structural and cultural dynamics of reentry systems. Most recently he co-authored Out of Prison & Out of Work, which provided the first-ever unemployment rate among formerly incarcerated people.
This report benefitted from the expertise and input of many individuals. The author is particularly indebted to Dan Kopf for retrieving this data from the ICPSR Physical Enclave, Amy Sawyer for her valuable insight into the state of homelessness today, Alma Castro for IRB assistance, Allen Beck for his insight into the NFPS, the ICPSR staff for their data retrieval support, Elydah Joyce for the illustrations, Maddy Troilo for background research, my Prison Policy Initiative colleagues, and the Connecticut Coalition to End Homelessness for their helpful guidance on “Housing First” strategies.
This report was supported by a generous grant from the Public Welfare Foundation and by our individual donors, who give us the resources and the flexibility to quickly turn our insights into new movement resources.
Numerous studies show that up to 15% of currently incarcerated people experienced homelessness in the year leading up to their incarceration. For more on this line of research see: Profile of Jail Inmates, 2002 and Education and Correctional Populations and Jail Incarceration, Homelessness, and Mental Health: A National Study. ↩
See Brianna Remster’s (2017) work on homelessness among formerly incarcerated people in Philadelphia. Half of the formerly incarcerated people in Remster’s study did not stay in a shelter until two years after release. ↩
According to the Bureau of Justice Statistics, 44% of those who were released from state prisons in 2005 were rearrested within one year; 68% within three years; and 83% in 9 years. High rates of rearrest and subsequent re-incarceration after release comprise what is frequently referred to as the “revolving door”. ↩
Because our data source (the National Former Prisoner Survey) contains restricted information, there were limits on what we could and could not export and analyze. Per ICPSR policy, if any query produced a result that included less than 200 respondents, we were not able to export that data. See the appendix for more detail. ↩
Couloute, Lucius and Dan Kopf. 2018. Out of Prison & Out of Work: Unemployment among formerly incarcerated people. Prison Policy Initiative. ↩
There is no widely accepted definition of housing insecurity. Instead, researchers have created different definitions using available data. As such, some researchers have defined and measured housing insecurity using residential moves, ability to pay rent, or rates of “doubling up” and living with others. Our measure represents a broad category of people who self-reported that they are either homeless or living in less-permanent spaces such as rooming houses, hotels, and motels. ↩
Couloute, Lucius and Dan Kopf. 2018. Out of Prison & Out of Work: Unemployment among formerly incarcerated people. ↩
Reid KW, Vittinghoff E, Kushel MB. 2008. Association between the level of housing instability, economic standing and health care access: a meta-regression. ↩
Ferguson, Kristy, Kimberly Bender, Sanna Thompson. 2011. Employment Status and Income Generation Among Homeless Young Adults. ↩
Stephen Lurie. 2013. The Astonishing Decline of Homelessness in America. The Atlantic. Release from Prison — A High Risk of Death for Former Inmates.The New England Journal of Medicine. ↩
For example, numerous studies show that up to 15% of currently incarcerated people experienced homelessness in the year leading up to their incarceration. For more on this line of research see: Profile of Jail Inmates, 2002 and Education and Correctional Populations and Jail Incarceration, Homelessness, and Mental Health: A National Study. ↩
We could not analyze the trans population because the number of trans-identified people in the survey was too small to use, and doing so posed a risk of respondent identification. ↩