In 2016, 549,000 persons were homeless on a given night in unsheltered locations, emergency shelters, and transitional housing in the United States. Of the half a million homeless persons, 355,000 were unaccompanied adult individuals, including youth between the ages of 18 and 24, and 194,000 were adults and children in families.
To prevent and reduce homelessness, it is important and necessary to determine the factors that lead to homelessness and their degree of relation to homelessness. Limited surveys of 20 cities of various sizes across the country in the U.S. Conference of Mayors Hunger and Homelessness Survey have listed the leading causes of homelessness for unaccompanied individuals and for households with children in those cities in 2015.
The top causes of homelessness fall into several categories: housing, poverty, employment, family disputes and domestic violence, and mental illness and substance abuse. Of the three main causes of homelessness for unaccompanied individuals, affordable housing was cited by 15 of the 19 cities that provided the information, eviction was listed by 1 city, poverty was listed by 9 cities, low-paying jobs were cited by 6 cities, unemployment was listed by 7 cities, family disputes and domestic violence was listed by 3 cities, mental illness was listed by 9 cities, and substance abuse was listed by 6 cities. Twenty cities included information for the three main causes of homelessness for households with children, with affordable housing cited by 16 of the cities, eviction listed by 5 cities, poverty listed by 11 cities, low-paying jobs listed by 7 cities, unemployment listed by 7 cities, family disputes and domestic violence listed by 8 cities, mental illness listed by 2 cities, and substance abuse listed by 2 cities.
To determine the factors that lead to homelessness, their relative importance in influencing homelessness, and ways to prevent homelessness and to reduce homelessness for those who are already homeless, U.S. Census Bureau state data from the 1-Year American Community Survey (ACS) for various housing, employment, and poverty[1] variables were assessed for their correlation and cause-and-effect with the number of homeless persons on a given night in 2016 in each state using Pearson correlation analysis and simple linear regression.
Housing
Various housing factors may comprise access to affordable housing and effect homelessness. First, the number of vacant housing units and vacant rental units is important because without vacant units that are available, persons are left with little option but homelessness. When assessing only the effect of vacancy status on the number of homeless persons, 43% of homelessness was accounted for by the total number of vacant units and 40% of homelessness was accounted for by the number of vacant units for rent. Both the total vacant units and vacant rental units predicted a moderate amount of homelessness with the total number of vacant units predicting a slightly larger percentage of homelessness.
Second, if vacant housing is available, the rent burden or cost of living relative to an individual’s or family’s income is important because if the cost of rent exceeds a certain level of income than an individual or family is less likely to be able to afford the rent and more likely to not be able to rent a unit initially or to eventually be evicted and again left with little option but homelessness, especially households with lower income for whom the remaining percent of income not devote to rent is too small to cover remaining necessary expenses. According to the U.S. Department of Housing and Urban Development, a household is cost-burdened when rent exceeds 30% of household income and a household is severely cost-burdened when rent exceeds 50% of household income.
When assessing only the effect of rent burden on the number of homeless persons, 74% of homelessness was accounted for by the number of low-income households who are 30% rent-burdened (rental households with income between $0 and $19,999 and who paid more than 30% of household income on gross rent in the past 12 months) and 77% of homelessness was accounted for by the number of low-income households who are 50% rent-burdened (rental households with income between $0 and $19,999 who paid more than 50% of household income on gross rent in the past 12 months). Although both the 30% cost-burdened low-income renters and the 50% cost-burdened low-income renters predicted a very large amount of homelessness, as might be expected, the 50% cost-burdened low-income renter numbers predicted a slightly greater percentage of homelessness.
Employment
Several components of employment data comprise employment and persons who are not working and effect homelessness. According to the Bureau of Labor Statistics (BLS), employment data consist of persons in the labor force and persons not in the labor force and the labor force is composes of employed persons and unemployed persons. The BLS defines unemployed persons as the number of persons who do not have a job, have actively looked for work in the past 4 weeks, and are currently available for work and includes persons who are not working and are waiting to be recalled to a job from which they had been temporarily laid off. The BLS defines persons not in the labor force as persons who are neither employed nor unemployed and includes retired persons, students, those taking care of children or other family members, and others who are neither working nor seeking work.
Therefore, the two components that comprise people not working are (1) persons who are in the labor force but are unemployed and (2) persons who are not in the labor force. Persons who are in the labor force but are unemployed are persons who desire to work and therefore may have few resources and may be likely to experience homelessness. Some persons who are not in the labor force may not have other income nor sufficient resources and may be likely to experience homelessness.
When assessing only the effect of employment of prime working age people ages 25-54 on the number of homeless persons, 76% of homelessness was accounted for by the number of unemployed persons, 70% of homelessness was accounted for by the number of persons not in the labor force, and 72% of homelessness was accounted for by the total number of unemployed persons and persons not in the labor force. Both the number of unemployed persons and number of persons not in the labor force predicted a large amount of homelessness. Although the number of unemployed persons predicted a slightly greater percentage of the number of homeless persons, the combined total number of unemployed persons and persons not in the labor force likely encompass an overall larger number of persons who are likely to become homeless.
The number of unemployed persons and persons not in the labor force who are age 25 to 54 excludes the persons ages 16 to 24 and who are minors living with parents or guardians and/or in school and the persons ages 55 to 64 who have income from retirement and/or disability. Yet, some persons who are ages 16 to 24 and 55 to 64 are likely to not have any income and in need of income to avoid homelessness. To account for the relative impact of these persons in homelessness, the number of persons age 16 and older who had income below the poverty threshold were assessed.
When assessing the number of persons who are age 16 and older with income below the poverty threshold who are unemployed or not in the labor force, the number of persons who are not in the labor force accounted for 73% of homelessness and the number of persons who are unemployed or not in the labor force accounted for 73% of homelessness. The number of persons not in the labor force and the total number of persons who are unemployed or not in the labor force with income less than the poverty threshold who are age 16 and older predicted a slightly greater percentage of the number of homeless persons than the number persons not in the labor force and the total number of persons who are unemployed or not in the labor force who are ages 25 to 54 (73% and 73% for ages 16 and older and 70% and 72% for ages 25 to 54, respectively).
Persons who are ages 55 to 64 and are unemployed — and more connected to the workforce — are more likely to have sufficient other income or resources, including income or resources not accounted for in the poverty calculations, to avoid homelessness than persons ages 55 to 64 who are not in the labor force. When assessing the number of persons who are age 16 and older with income below the poverty threshold who are unemployed, the number of persons who are unemployed accounted for 71% of homelessness. This might explain why the number of unemployed persons with income less than the poverty threshold who are age 16 and older predicted a slightly less percentage of the number of homeless persons than the number of unemployed persons ages 25 to 54 (71% and 76%).
Poverty
Poverty is defined as income that is below the level necessary to support a minimally-sufficient standard of living. The U.S. Census Bureau measures poverty using a set of money income thresholds that vary by family size and composition and if a family’s total income is less than the family’s threshold, then the family and every individual in it is considered in poverty. As total employment was composes of persons who are in the labor force and unemployed and persons not in the labor force, poverty is a composite variable that may encompass unemployment and lack of labor force participation, along with low-income employment, seasonal employment, and part-time employment, and other sources of income such as Social Security, Veterans’ payments, retirement income, educational assistance, alimony, child support, and other miscellaneous sources.
In addition to categorizing a person as in poverty, the depth of poverty shows the degree of income deficit from the level needed for a minimally-sufficient standard of living. In general, individuals and families with income only half (or 50%) of the poverty level are described as in deep poverty and may be more likely to have difficulty meeting a minimally-sufficient standard of living and to experience homelessness.[2] Of course, in geographic areas where the cost of living is higher, persons with income near poverty may experience great difficulty in maintaining a minimally-sufficient standard of living as well. Likewise, persons with additional expenses for child care, medical expenses, transportation, or other necessities may experience greater difficulty in maintaining a minimally-sufficient standard of living.
When assessing only the effect of poverty on the number of homeless persons, 68% of homelessness was accounted for by the number of persons in poverty with income below the poverty threshold and 69% of homelessness was accounted for by the number of persons in deep poverty with income below 50% of the poverty threshold. Although both the number of persons in poverty and the number of persons in deep poverty predicted a large amount of homelessness, as might be expected, the number of persons in deep poverty predicted a slightly greater percentage of homelessness.
Conclusion
The analysis of the ACS state data on housing, employment, and poverty support the indications by 20 cities that affordable housing, low-paying jobs, unemployment, and poverty are related to homelessness and were factors that heavily influenced homelessness. These analyses indicate that decreasing poverty, increasing employment, reducing the proportion of household income used for rent and utilities, and expanding the number of vacant housing units and vacant rental units is likely to prevent and reduce homelessness.
[1] ACS data for poverty were used instead of data from the Current Population Survey (CPS) because the ACS is a larger survey with a sample size of 3 million as opposed to the smaller survey of 100,000 households for the CPS, data collection is more rigorous with fewer non-responses for the ACS as a larger number and easier survey methods are used and it is a mandatory survey instead of a voluntary survey like the CPS< the ACS population universe is more expansive and includes persons in institutions that are not included in the CPS, and the data collection is continuous throughout the year for the ACS with the reference period constituting the 12-month period prior to the response month and the survey estimates centering around December 2015 – proximate in time to the collection period for the homeless data of January 2016 – instead of 12-month interviews for the CPS conducted only in February through April 2016 that center around July 1, 2015.
[2] According to the U.S. Census Bureau, although the official poverty definition does not include capital gains, noncash benefits (such as public housing, Medicaid, and food assistance), and tax credits, it uses money income before taxes. Therefore, the U.S. Census Bureau poverty data may include some income that is not available to provide a minimally-sufficient standard of living and exclude some money or benefits that are available to provide for a minimally-sufficient standard of living.