1. 1
Economics
or
Personal
Choice:
What
Explains
Homelessness
Along
the
Wasatch
Front?
Jessica
Keomalu
Homelessness
is
a
widely
known
but
seldom
visited
issue
plaguing
the
American
public.
Perhaps
the
greatest
problem
with
homelessness
is
that
the
various
factors
attributing
to
this
social
problem
are
so
expansive.
Highly
reviewed
factors
include
pervasive
substance
abuse,
domestic
violence,
inadequate
skill
or
education,
veterans’
status,
and
incarceration
–
reasons
which
encapsulate
social,
economic,
and
political
reasons.
Still
unaccounted
for
in
this
is
the
idea
of
personal
choice,
for
which
some
individuals
opt
to
adapt,
becoming
voluntarily
homeless
in
the
process.
This
paper
seeks
to
answer
which
factors
explain
homelessness
along
the
Wasatch
Front
of
Utah
with
data
gathered
via
personal
interviews
and
depth
surveys.
To
estimate
for
duration
of
homelessness,
I
use
the
median
number
of
years
each
individual
identified.
Using
one-‐way
ANOVA
and
Ordinary
Least
Squares
Regression,
this
paper
finds
that
economic
factors,
veterans’
status,
and
lifer
status
are
all
significant
factors
in
contributing
to
duration
of
homelessness;
that
is,
the
aforementioned
factors
are
statistically
significant
and
positively
correlated
with
number
of
years
spent
being
homeless.
I. Introduction
The problem of homelessness is one that extends nationwide, arising from reasons related
to the economy and unemployment and to socially contrived factors. While some argue that the
homeless are victims of a discriminatory system or inadequate funding, others hold that the
homeless population can be described as “lazy bums.” Regardless of the case, homelessness
continues to evidence itself as a problem with no easy solution and with characteristics that
require further analysis.
Utah, however, has been able to decrease the rate of homelessness by 91% since 2005,
thanks to the Housing First Program (McEvers 2015). Such a substantial reduction has cut
Utah’s current population of homeless individuals to 3,025 at any given point in time in 2015.
Overall, Utah contributes 2.57% to the total homeless population of the United States
2. 2
(Comprehensive Report on Homelessness 2015, p. 8). If the state was able to reduce
homelessness by 91% over a span of 10 years, is it not reasonable to expect that further
reductions in the rate of homelessness should occur? However, various factors may be significant
in determining why the homelessness rate will not reach zero percent, even if the state produced
significant reductions previously.
The purpose of this paper is to determine which factors, whether economic or personal
choice, contribute to a person being homeless along the Wasatch Front in Utah. This area, which
consists of Ogden, Salt Lake City, West Valley, West Jordan, and Provo, contains the largest
concentration of homeless individuals and families in Utah. I focus specifically on the
individuals identifying as homeless and forego inclusion of homeless families in this study for
simplification purposes.
An Overview of Homelessness in Utah
The U.S. Department of Housing and Urban Development defines homelessness as
individuals and families with no fixed, regular and adequate nightly residence (Comprehensive
Report on Homelessness 2015). Homeless in and of itself represents a social and economic
problem of inefficiency and overall devaluation of human potential. Utah’s Department of
Workforce Services (DWS) asserts that homelessness lessens the likelihood of a homeless person
finding employment and eventual income to purchase rent for a home (Comprehensive Report on
Homelessness 2015). In its Comprehensive Report on Homelessness 2015, the DWS details the
drastic decrease in homelessness: since 2005, homelessness in Utah has decreased by 91%. In
Utah, reasons for homelessness stem from domestic violence, veterans status, mental illness and
physical disabilities, and substance abuse and addiction. Additional reasons for homelessness
arise from economic factors whereby individuals or families find themselves homeless from
3. 3
insufficient skills for a job and lack of education (Comprehensive Report on Homelessness
2015). I include these aforementioned reasons for homelessness in my model.
Economics versus Personal Choice
When discussing homelessness, the issue is commonly marked as one arising from
unemployment, veteran’s status, or substance abuse. Economic factors, however, include lack of
income or lack of education, which may exacerbate the potential for employment, let alone
employment to satisfy standard living conditions and to purchase rent. The Great Recession of
2008-2009, including the Pre-Recession phase, may have contributed to the problem of
homelessness when factoring in the substantial rise in unemployment and the foreclosure of
homes all across the country (Goldman 2009). When thinking of homelessness, we may
associate the problem with lack of affordable housing units or available jobs or income or a
combination. I also include education in that consideration, as (sufficient) education is a
necessary condition for employment at a basic level.
Personal choice, which is exactly how the term sounds, describes a voluntary and
conscious decision to live untethered to an edificial house; rather, proponents engage in
voluntary homelessness as a lifestyle choice. The idea may be incomprehensible to some, for no
average person would willingly subject himself to the elements as a means of lifestyle. Yet, a
subset of the American population chooses voluntary homelessness for reasons including
freedom from restrictive housing codes and landlords, vagrancy travelling, and “life experience.”
Based on these two definitional concepts, I explore their relationship to the Wasatch
Front while simultaneously including factors traditionally associated with homelessness by the
DWS. If the rate of homelessness has decreased by a substantial percentage in the past 10 years,
some reason must exist to explain why homelessness in Utah has not been completely
4. 4
vanquished. It may be that homelessness cannot be zeroed out; a remainder of the population
may not wish to have a home, continually contributing to an addition to the homeless population.
For the purposes of this study, I will focus on economic reasons, disabilities, social reasons,
domestic reasons, personal choice, veterans’ status, and lifetime of homelessness as factors
influencing an individual’s length of time spent being homeless.
Description of Factors of Homelessness
This paper focuses on seven categories that characterize homelessness: (1) economic
reasons refer to lack of income, job loss, and lack of education; (2) disability includes the span of
mental and physical disabilities and handicaps; (3) social reasons are contrived from substance
abuse and addiction as well as incarceration, which represent the miscellaneous social issues
associated with homelessness; (4) domestic reasons encapsulate family issues or disputes,
domestic violence, and divorce; (5) personal choice refers to an individual’s conscious decision
to live a lifestyle of voluntary homelessness; (6) veterans’ status must be included, as a majority
of the homeless population nationwide is a veteran; and, (7) lifer describes the situation in which
an individual has grown up homeless or has been homeless from an early age or has been
exposed to homeless/housing instability from early onset of life, possibly indicating a
predisposition to homelessness as an adult.
I expect a priori that duration spent homeless increases when the individual has lifer
status: an individual is likely to have spent more time being homeless when he or she has been
consistently exposed to homelessness. I also expect to find that personal choice results in
insignificant results, as this population itself is limited in number and may actually be a rarity
among the homeless population. By conducting this study on homelessness, I hope to discover
reasons explaining homelessness in the current sampled population of the Wasatch Front, as well
5. 5
as forming a connection between the length of time spent being homeless and the reasons why an
individual is homeless.
This paper outlines a review of the literature followed by a theoretical overview and
hypothesis, description of the data, and explanation of the methodology, leading to the results of
the study and concluding with implications and limitations.
II. Homelessness: A Review of the Literature
Economic Factors of Homelessness
A key component of homelessness derives from an inability to procure income for rent:
when a person finds himself unable to secure income, he is subsequently unable to pay for rent.
Therefore, the prices of rent and homes greatly influence the number of homeless people in an
area. Particularly, areas with higher rents face a higher-than-average homeless population (Crane
& Warnes 2000; Park 2000). These areas are usually metropolitan areas with large population
densities. Such areas also have fewer low rent vacancies available, which creates even more
difficulty for low-income individuals to obtain housing. The study of Quigley, Raphael, and
Smolensky (2001) reveals that areas with greater income inequality, such as that of metropolitan
areas and big cities, are subject to higher levels of homelessness. Generally speaking,
homelessness is prevalent in big cities with large inequalities in income distribution.
A high level of income inequality, that is a large distribution of income inequality with a
significant portion of a population earning lower levels of income, increases the incidence of
homelessness. In a similar study conducted by Quigley and Raphael (2001), an increase in
income inequality causes a spike in the demand for low-quality income and a decrease in the
demand for middle-quality housing. Mansu, et. al. (2002) build upon this work and conclude the
existence of a link between income inequality and homelessness. A 20% decrease in the lowest
6. 6
quintile of their sample population caused an average increase in the homelessness population by
23.5%. Their results also indicate that a 20% increase in the incomes of the top quintile resulted
in no effect on homelessness, which is to be expected. A surge in income at the upper levels of
the income distribution does virtually nothing for people at the lower end of the distribution who
are facing a substantially higher likelihood of being homeless.
On the basis on income alone, low wages are linked to an individual being homeless
(Glomm & John 2002). Conversely, high wages result in an individual being housed. If a person
is unable to secure income necessary for payment of rent and wishes to secure a loan, he is
unable. A lack of income coupled with an inability to borrow forces the individual into
homelessness. Ensuing loss of productivity becomes an additional problem, as the individual is
now unable to earn any rent at all and cannot secure the loan, pushing him further into
homelessness. Income itself, in addition to income inequality, plays a substantial role in a
person’s choice, situational or otherwise, to consume zero housing and become homeless.
Educational attainment is often linked to income. The more education a person receives,
the more income he is also likely to earn. Jarvis (2015) reports an inverse relationship between
an individual’s educational attainment, as one might expect. High school graduates spend 49%
fewer nights being street homeless as compared to a non-high school graduate. As the level of
educational attainment increases, time spent homeless decreases significantly.
Quigley, Raphael, and Smolensky (2001) use a simple model of housing choice via utility
functions subject to the constraint of price of housing and income. Poverty levels stem from lack
of income (from any form such as a job or as social security benefits). To view the situation of
poverty and income is to see a person facing budget constraints. A person receiving little to no
income has the choice of consuming rent or foregoing it and becoming homeless. Similarly, even
7. 7
the richest homeless person is purely indifferent between consuming extremely dilapidated,
“abandonment quality” housing at predetermined rent or being homeless with no rent
consumption. Decisions made by people encountering homelessness are ones that rest on the
premise of consumption and preferences (Quigley & Raphael 2001). O’Flaherty (1995) finds that
people with more income will always spend more for a means of shelter than be homeless.
However, people with insufficient or nonexistent incomes may have no other choice than to be
homeless or seek alternative means of temporary shelter.
Additional economic factors that increase likelihood of homelessness are factors which
relate to eviction rates. Cranes and Warnes (2000) reveal six “risk factors” for a greater
likelihood of default on rent or home payments, eventual eviction, and inevitable entry into
homelessness: (1) changes from stable to instable income, (2) obtaining a mortgage in middle or
old age, (3) living in a home with “disturbed behavior” such as drug use, (4) complete stoppage
of housing benefits, (5) living alone with no other supporter, especially in cases of a recent co-
resident supporter, and (6) prior stint of homelessness. These factors may be further compounded
by additional social factors, which creates a hole from which escaping homelessness becomes all
the more difficult. Jarvis’ (2015) findings somewhat contradict that of Cranes and Warnes: the
researcher concludes that homelessness caused by evictions is “less intense.” That is, individuals
homeless as a result of eviction spend 27% fewer nights being street homeless as compared to
individuals who are homeless from other factors. From this, we see that some of these risk
factors overlap with those provided by the DWS report.
Social Descriptors of Homelessness
Social factors often go hand-in-hand with economic factors of homelessness. That is,
individuals who are homeless are usually in such a situation due in part to reasons that coincide
8. 8
with economics factors. Having a mental disability or post-traumatic stress disorder as a side
effect of military combat significantly increases the likelihood that an individual will not receive
adequate income (Jarvis 2015). Here, what can be called “social factors,” those which describe
all factors not attributed to economics or personal choice, encompass the rest of the factors
attributed to the homeless population.
Veterans are much more intensely homeless than non-veterans in that they experience
longer durations of homelessness (Jarvis 2015). Early (2004) reports that a Vietnam veteran is
39.61 times more likely to become homeless as compared with a non-veteran. They are also
associated with more episodes of post-traumatic stress disorder (PTSD) as a mental illness and
alcohol abuse as compared to non-vets. Therefore, veterans who suffer from PTSD encounter an
increased tendency to becoming homeless.
Severe mental illness itself is a strong contributing factor to homelessness (Piat, et. al.
2015). In the same study, Piat et. al. (2015) report that homeless individuals with mental health
issues experience “significant disadvantage, as they are likely to be impacted by poverty, family
instability, and domestic violence.” Moreover, mental illness is also associated with substance
abuse problems. A large proportion of the homeless population experiences some type of abuse
with alcohol or illicit drugs in conjunction with mental health problems. Early (2005) also finds
that problems with substance and alcohol abuse were statistically significant factors to explain
homelessness. These factors were not found to be significant determinants of shelter use.
Childhood abuse and trauma are significant risk factors for homelessness. Homeless
people have “experienced an average of 8.8 adverse life events, including incarceration,
suicidality, parental abandonment and the death of a mother” (Piat, et. al. 2015). Other sudden
life events are likely to cause homelessness, but these individuals who are homeless as a result of
9. 9
a life shock experience homelessness in shorter spans than those facing other factors (Jarvis
2015).
Homelessness as a Personal Choice
Homelessness is not solely a factor of economic or social conditions. Often discounted,
personal choice still constitutes a viable reason for being homeless. Much research on personal
choice as a reason for being homeless is sparse. Parsell and Parsell (2012) state that
“homelessness is depicted as the conscious choice of an individual who has abandoned the
superficiality of long-term work,” thus foregoing socially constructed norms. A person may
desire to be homeless on the basis of several personal reasons, but above all, “homelessness is a
calculated and rational choice of a free agent” (Parsell & Parsell, 2012). Free agency here
describes a state of personal choice – a person is not constrained or under duress to select the
homeless lifestyle. As temperatures drop, the number of homeless people living on the street also
declined (Quigley & Raphael 2001). This indicates that homeless people also choose when and
where they wish to exercise their choice to be homeless. The findings of Quigley and Raphael
(2001) support rational choices among the extremely poor. The articulated choice in selecting
homelessness as opposed to alternative shelter options presents the finding of homelessness as a
means of expressing one’s self in situations where those who are homeless did not feel personal
connection to mainstream society. Further, personal choice may explain a subset of the homeless
population which does not see homelessness as an economic or social condition.
III. Theory and Hypothesis
Utility Maximization Theory
The idea behind homelessness relates to utility maximization, a theory which examines
how each individual attempts to make himself as well off as possible, given the circumstances in
10. 10
which he finds himself (Lipsey, Steiner, and Purvis, 1987, p.131). An individual “demands each
good up to the point at which the marginal utility per dollar spent on it is the same as the
marginal utility of a dollar spent on another good” (Lipsey, Steiner, and Purvis, 1987, p.133).
Thus, a functional relationship exists between utility and consumption of various goods and
services (Cochrane & Bell, 1956, p. 30).
Quigley and Raphael (2001) attempt to study the comprehensive set of alternative
hypotheses that changes in the rate of homelessness are the result of circumstances in the
housing market and income distribution. The researchers assert that a homeless individual has
made a choice based on income, being “indifferent between consuming ‘abandonment-quality’
housing…on one hand, and homelessness at zero rent on the other hand.” To supplement their
assertions, the researchers use a simple model of housing choice, which takes the form of a
utility maximizing function. The idea of choice and resource allocation forms the cornerstone of
all subsequent research, such that an individual’s choice to allot resources to one preference over
another dictates utility maximization. Based on the work of the authors, this study uses utility
maximization as a theoretical foundation to examine the homeless population along the Wasatch
Front. Regardless of context of location, utility maximization among homeless individuals or
individuals prone to homelessness does not change.
Troutman, Jackson, and Ekelund (1999) study the data from the Housing and Urban
Development’s 1984 survey to gauge the impact of various policies in combatting homelessness.
To get to the deeper implications of their research question, the researchers analyze the
maximization of behavior by the homeless, operating under the assumption that an individual,
homeless or prone to becoming homeless, weighs the costs and benefits of alternatives and
selects the option that maximizes benefit. In another set of chain analysis, Troutman, Jackson,
11. 11
and Ekelund examine an individual’s choice among unsubsidized private housing, government
subsidized housing, and homeless shelters. Unemployment and poverty rates, which restrict
income to tight levels and reflect reduced earnings, should increase the number of individual
selecting homeless shelters. However, it should also be noted that the researchers do not consider
is a person’s option to select being homeless over being housed – a situation which must also be
considered, even among rational people. We may assert that even rational people may select
homelessness as a response to their budget constraint.
Allgood and Warren (2003) examine the determinants that influence length of time being
homeless. They assume that an individual receives utility from consuming housing h, leisure l,
and other goods g, also factoring for income and time to enjoy leisure. Individuals will consume
a good (set to housing, leisure, and other) subject to a budget constraint.
For the purposes of this study, homelessness is identified as a function of explanatory
factors, ceteris paribus, as given by the following function:
(1)
Homelessness = f(economic, disability, social, domestic, personal choice, veteran, lifer)
Equation 1 describes homelessness as a function of economic problems, mental or
physical disability, social problems, domestic problems, veterans’ status, and lifer status. For the
purposes of our study, we also assume an individual maximizes his or her utility based on the
following function, ceteris paribus:
(2)
U=f(Housing, All other goods)
Equation 2 represents the basic utility function for a homeless person, where utility is a
function of housing/rent costs per month and income spent on all other goods (Figure 1). We
derive the condition of the marginal rate of substitution, given by equation 3:
12. 12
(3)
MRS = slope of the budget constraint, or
MRS =
!!"#$%&'
!!"" !"!!" !""#$
Economic Theory Summary and Hypothesis
Every individual is subject to choices regarding income allocation. He can spend it on
good X, such as video games, or on another good Y, such as a new computer. In a more realistic
example, an individual must choose between good R, the price of rent, or good D, the price of
drugs. In the surveys conducted with the homeless population of the Wasatch Front, drugs seem
to have a connection to being homeless so much so that the individual receives more utility for
drugs than for housing. In any case, a person who is homeless demonstrates his utility
maximization by allocating income, however scarce, to drugs over rent or payments for housing.
If a person receives more utility for one good over another, he will forego the less desirable good
and use income for the good he desires.
Based on the review of literature, I expect to find that lifer status is an overwhelmingly
significant factor that explains the number of years a person is homeless. If a person has grown
up exposed to homelessness, I hypothesize that he will experience more years being homeless.
Given a priori expectations, I also hypothesize that economic factors are significant in duration
of homelessness.
IV. Data
Data was obtained via a convenience sample of homeless individuals along the Wasatch
Front in a sample area that spans from Ogden into Salt Lake City, West Valley City, and West
Jordan, down to Provo. In more isolated areas, a snowball sample was obtained via referrals from
one survey participant to another. Data collection dates are as follows: March 6, 16, 18, 20, 22,
13. 13
23, 25, 28, 29, and 30, spanning over 28 man hours in personal interviews, resulting in a total of
74 usable surveys (see Appendix 1 for breakdown of true sample size). Survey administration
was done on the street, as many homeless shelters did not agree to administration in the facility
due to legal liability; in these instances, data were collected from individuals identifying as
homeless who were directly outside of or around the shelter. The shelters included Rescue
Mission of Salt Lake, the Road Home, Family Promise of Salt Lake City, the Road Home in
Midvale, and Prove Homeless Shelter. For privacy reasons, Rescue Mission’s Women’s Center
did not allow surveys to be administered near the facility.
The survey consisted of 27 total questions broken down into the following categories (see
Appendix 2 for actual survey administered): general and demographic questions, 1-9;
employment, income, and welfare questions, 10-15; housing questions 16-22; personal choice
questions, 23-27. A total of 74 respondents took the survey. Originally, 81 homeless individuals
were interviewed but seven individuals did not provide answers for every question;
subsequently, those seven interviews were discarded, the partial results of which will not
included in this study.
Question five in the general section asks, “Why are you homeless,” to which respondents
may select from 10 options with the last option being “j. Personal Choice.” Any respondent who
selected this option was then asked to continue with questions 23-27 whereas individuals who
did not select option j. Personal Choice were not asked to answer. Participants were approached
and asked to complete a survey as part of an economics research class at Weber State University.
They were told that as compensation, they would receive $1 for their time. Participants were
interviewed, and the researcher marked the responses as they answered. Verbal comments were
14. 14
also noted where appropriate. Interviews varied in time, but averaged 7-10 minutes. All data
collected via paper survey were then transposed into an Excel spreadsheet for coding.
Of interest and worth mentioning in this section of the paper is the duration of
homelessness: of the 74 respondents, 17 have been homeless for 10 or more years; 22 have been
homeless for seven to 10 years; 18 have been homeless for four to six years; 6 have been
homeless for two to three years; 5 have been homeless for six months to one year; 6 have been
homeless for five months or fewer. Figure 2 provides a breakdown of these durations. This
population also expressed their homelessness in terms of their characterization of their situation:
50 individuals identified as being homeless for the first time in a prolonged, consistent period of
time; 13 individuals identified as lifers, or being consistently homeless or growing up exposed
and living from temporary place to temporary place; 10 individuals identified as homeless but
experiencing intermittent periods of rent payment for a dwelling; 1 person identified as being
homeless for a period but living with family or friends and then returning to homelessness. For
68% of the sample population, homelessness is characterized by the first option listed. In
answering the question of why the individual is homeless, most said that they were homeless due
to lack of income, job loss/lack of education, and lack of affordable housing options. Three
individuals answered that they were homeless as a matter of personal choice. Figure 3 shows a
complete chart of reasons the individual attributed to their homelessness situation.
V. Methodology
I use categorical data in the form of factors of homelessness, summarized by the term
REASON, which are labeled economic, disability, social, domestic, personal choice, veteran, and
lifer. These data are independent, unrelated groups, which is a hallmark of a one-way ANOVA
test. Continuous data take the form of median years spent homeless, denoted by MEDYEARS,
15. 15
which was taken by arriving at the midpoint of each length of homelessness duration category.
For example, zero to one year resulted in 0.5, two to three years resulted in 1.5, and so on. For
durations grater than 10 years, I use a conservative estimate of 10. One assumption of ANOVA
calls for the endogenous variable to be measured on a continuous level.
Because I wish to test the correlation between the various reasons to which homeless
individuals attributed their entrance into homelessness and the number of years a person has been
homeless, I need to run an ordinary least squares (OLS) regression, shown by equation 4:
(4)
Median Years Homelessi = β0 + β1ECONOMIC + β2DISABILITY+ β3SOCIAL+ β4DOMESTIC+
β5CHOICE+ β6VET+ β7LIFER + εi
By converting years to the median, we can estimate that the length of time a person has
been homeless is at least a conservative estimate of his entire duration spent homeless. Doing so
allows us to regress the number of years with the explanatory variables in an attempt to view the
relationship between the two.
Early (2004) uses a binary logistic regression model to analyze his data, which could be
replicated in this study. This will be explored further as an area for improvement in the
conclusion section.
VI. Results
To begin, I test the explanatory variables for their descriptive statistics (from figure 2). In
analyzing these statistics, I look to find any outliers that may skew my ANOVA and further
regression. At first glance, we see that CHOICE, possibly DOMESTIC and VET, shows outlier
tendency. For this, I take to basic count data to confirm outliers. Figure 3 shows that there are
only 3 instances of personal choice out of 74. I then look to DOMESTIC and VET, with means
16. 16
of 0.0811 and 0.0946, respectively, to conclude that there are six individuals identifying
domestic problems and seven individuals identifying veterans’ status as being the contributory
factor for their entrance into homelessness. Even with these considerations, I decide to run
ANOVA with these variables. I run a one-way ANOVA to determine if there is a statistical
difference in the mean median years among the seven different groups of the exogenous variable
REASON. The resulting p-value is 0.003, which we accept as statistically significant (Table 1).
From Equation 5, I derive the following equation:
(5)
Median Years Homelessi = -0.12 + 6.40ECONOMIC+ 4.78DISABLE + 6.13SOCIAL +
2.22DOMESTIC + 1.29CHOICE + 3.02VET + 2.46LIFER + εi
Three of the seven variables tested produced significant results (Table 2). The
ECONOMIC variable shows a p-value of 0.050, indicating significance. VET also produced
statistical significance with a p-value of 0.024. LIFER shows a p-value of 0.014, the smallest p-
value of all the explanatory variables. If we look at SOCIAL, we see a p-value of 0.075, which is
statistically significant with a 90% level of confidence. In keeping with traditional literature, we
may choose to include addiction as a statistically significant factor. In essence, four of the seven
categories attributed to entry into homelessness are significant in some way, staying true to the
nature of the literature. It is worth mentioning that CHOICE produces the p-value with the least
amount of significance, which is to be expected with a small number of respondents who
selected personal choice as the reason to which they attribute their entry into homelessness.
The R-squared statistic is reported at 33.7%, which is relatively low, but based on the
significance in three explanatory variables, we may still be able to draw important conclusions.
We may also attribute this low R-squared statistic to fallacies in predicting human behavior in
17. 17
that we cannot be certain what predictors may influence a person’s likelihood of becoming
homeless.
VII. Conclusion
Earlier in this paper, I hypothesize that individuals who identify with LIFER status will
spend a longer time, as measured by years, being homeless. From the results, we see that this
holds true with a confidence level of 95%. We can draw from the implications that being
exposed to homelessness may be associated with predisposing factors, fostering a perpetual and
vicious cycle. While further data collection and analysis are needed for this, Kelly (2007) finds
that exposure to homelessness as a child greatly increases the likelihood of the same child
becoming homeless at a younger age and for a longer duration of time. Based on these findings
and the work of Kelly (2007), programs should be put into place that target at-risk,
unaccompanied youth or homeless families with young children, as to prevent a cyclical pattern
of homelessness. At least a portion of homeless individuals in Utah can attribute their entry and
subsequent pacification of their situation to their childhood and the exposure they gained while
their families experienced homelessness.
As expected, economic factors are a significant factor in duration of homelessness. This
category includes lack of income, lack of education, and job loss. We expect lack of income to
be part of this category, as lack education or skill may result in few job prospects. Job loss
equates to zero income. When thinking about the implication of what economic factors mean as a
whole, we must consider why it is that such factors are positively correlated with duration of
homelessness. To pique at such considerations, we must examine motivations in searching for
employment and motivations for remaining homeless, the topic of which may be examined for
further research.
18. 18
Further, those who identify as veterans also spend more time being homeless.
Characterizations of veterans include post-traumatic stress disorder (PTSD) or other mental
disabilities resulting from their line of duty. Left untreated, Elbogen et. al. (2008) find that PTSD
is a significant factor in leading to homelessness as a result of interference with everyday
function, leading to job loss and unemployment. An undiagnosed or untreated mental illness
associated with veterans may manifest soon after tour of duty, disrupting everyday activity and
causing premature unemployment, the likes of which may then lead to homelessness without
other sources of income or assistance. Elbogen et. al. (2008) also report that homelessness may
be most difficult to reverse for this population, as mental health deterioration is difficult to
restore.
However, interestingly enough, disability is not a statistically significant factor in our
model. Although it has a role in veterans’ status, it is not prominent enough on its own to call for
statistically significant result. Undeniably, the expanse of the homeless population nationwide
falls prey to some type of disability, be it physical or mental, often undiagnosed or untreated as is
above. While we cannot conclude that disability is a factor affecting length of time spent
homeless, we also cannot ignore the disabilities plaguing the homeless population as a whole.
Personal choice and domestic problems are statistically insignificant, though positively
correlated with length of time spent being homeless. The title of this paper asks which factors
explain homelessness along the Wasatch Front. We can certainly say that personal choice is not
significant in explaining homelessness along the Wasatch Front, though it holds true for three
members of our sampled population. This study, in fact, points to economic factors, veterans’
status, and lifer status as explaining homelessness along the Wasatch Front of Utah.
Limitations and Deficiencies
19. 19
The limitations of this study are numerous. Specifically, data analysis was not as robust
as it could have been, thus presenting lapses in data presentation. The dependent variable is
categorical in nature: there are many categories of years in which a person may describe his
length of homelessness. It is worth mentioning that ordinarily, my dependent variable is
categorical in nature, as it represents many categories of year segmentation. My independent
variables are also categorical in nature, with the reasons of homelessness representing possible
selection categories. However, due to lack of understanding multinomial logistic regression, I
convert the categories of years spent homeless into median years, thus allowing the length of
time in years to become continuous data.
Early (2004) runs a logit regression to estimate the probability of being homeless as a
function of household characteristics and the characteristics of the city where the household is
located. From his results, which are presented in three different models, further testing using my
results may explain more than what has been presented. To regress for length of time spent
homeless categories (0-1 years, 2-3 years, etc.), my research would stand to gain more accuracy
by using binary logistic regression wherein the dependent variable (length of time spent
homeless) was scored either 0 or 1.
Lastly, as was mentioned earlier, sample size obtained in this study did not meet the true
sample size that was derived statistically. A true sample size of 385 individuals was needed to
ensure, at a higher level, a more robust study.
20. 20
References
Allgood, S., & Warren, R. S. (2003). The duration of homelessness: evidence from a national
survey. Journal of Housing economics, 12(4), 273-290.
Cochrane, W. W., & Bell, C. S. (1956). The economics of consumption; economics of decision
making in the household. New York: McGraw-Hill.
Crane, M., & Warnes, A. M. (2000). Evictions and prolonged homelessness. Housing
Studies, 15(5), 757-773.
Early, D. W. (2004). The determinants of homelessness and the targeting of housing
assistance. Journal of Urban Economics, 55(1), 195-214.
Early, D. W. (2005). An empirical investigation of the determinants of street
homelessness. Journal of Housing Economics, 14(1), 27-47.
Elbogen, E. B., Beckham, J. C., Butterfield, M. I., Swartz, M., & Swanson, J. (2008). Assessing
risk of violent behavior among veterans with severe mental illness. Journal of traumatic
stress, 21(1), 113-117.
Glomm, Gerhard, and Andrew John. Homelessness and labor markets. Regional Science and
Urban Economics, 32(5), 591-606.
Goldman, D. (2009, January 09). Worst year for jobs since '45. Retrieved April 19, 2016, from
http://money.cnn.com/2009/01/09/news/economy/jobs_december/
Jarvis, J. (2015). Individual determinants of homelessness: A descriptive approach. Journal of
Housing Economics, 30, 23-32.
Kelly, E. (2007). The Long-Term Effects of Homelessness on Children.Council Report for
Vermont Senate.
Lipsey, R. G., Steiner, P. O., & Purvis, D. D. (1987). Economics. New York: Harper & Row.
McEvers, K. (2015, December 10). Utah Reduced Chronic Homelessness By 91 Percent; Here's
How. Retrieved April 19, 2016, from http://www.npr.org/2015/12/10/459100751/utah-
reduced-chronic-homelessness-by-91-percent-heres-how
Parsell, C., & Parsell, M. (2012). Homelessness as a choice. Housing, Theory and Society, 29(4),
420-434.
Park, J. Y. S. H. (2000). Increased homelessness and low rent housing vacancy rates. Journal of
Housing Economics, 9(1), 76-103.
Quigley, J. M., & Raphael, S. (2001). The economics of homelessness: The evidence from North
America. European Journal of Housing Policy, 1(3), 323-336.
21. 21
Quigley, J. M., Raphael, S., & Smolensky, E. (2001). Homeless in America, homeless in
California. Review of Economics and Statistics, 83(1), 37-51.
Troutman, W. H., Jackson, J. D., & Ekelund, R. J. (1999). Public Policy, Perverse Incentives,
and the Homeless Problem. Public Choice, 98(1-2), 195-212.
23. 23
Figure 2. Breakdown of Duration Spent Being Homeless, N=74
Duration
Number of
individuals
0-5 months 6
6months-1year 5
2-3 years 6
4-6 years 18
7-10 years 22
longer than 10 years 17
0-‐5
months
8%
6months-‐1year
7%
2-‐3
years
8%
4-‐6
years
24%
7-‐10
years
30%
longer
than
10
years
23%
Time
spent
homeless,
in
percentages
24. 24
Figure 3. Breakdown of reasons why individuals attributed their entry into homelessness
REASON Number of individuals, where N=74
Economic 29
Disability 10
Social 8
Domestic 6
Choice 3
Vet 7
Lifer 11
Economic
39%
Disability
14%
Social
11%
Domestic
8%
Choice
4%
Vet
9%
Lifer
15%
Breakdown
of
reasons
individuals
attributed
their
entry
into
homelessness
25. 25
Appendix 1. Computation of ideal, true sample size
Based on Utah’s Comprehensive Report on Homelessness 2015, there are 3,025 homeless
individuals at any point in time (based on 2015 point in time count).
n=[(z-score)^2 * St.Dev(1-St.Dev)]/(ME^2),
where:
n is the sample size, to be determined
z is the z-score of 1.96, correlated with a confidence level of 95%
ME is the Margin of Error, taken to be +/-5%
St.Dev is the Standard Deviation, taken to be 50%
To calculate sample size based on 95% confidence,
n = [(1.96^2) * (.5)(.5)]/(.5^2)
n = 384.16
*However, we can calculate the true sample based on the size of the population.
True sample = (n*N)/(n+N-1)
where:
n is the sample size, as calculated above
N is the total population size
To calculate true sample size,
True sample = (384.16*3025)/(384.16+3025-1)
True sample = 340.97
Based on a 95% confidence level with a +/-5% margin of error, 341 people would need to
be surveyed.
However, I expanded my confidence level to 99% to create an upper bound limit in the
number of people I need to survey. Because we know that a larger sample size may
indicate more significance in the study, I performed the previous calculations, this time
subbing for 99% confidence with an associated zscore of 2.576.
To calculate sample size based on 99% confidence,
n = [(2.576^2) * (.5)(.5)]/(.5^2)
n = 663.5776
To calculate true sample size,
True sample = (663.5776*3025)/(663.5776+3025-1)
True sample = 544.347
Based on a 99% confidence level with a +/-5% margin of error, 545 people would need to
be surveyed.
26. 26
Appendix 2. Homeless survey
Homelessness survey
GENERAL:
1. Male / Female / Prefer not to answer / Other: _________
2. Age: ________
3. How long have you been homeless?
a. 0-5 months
b. 6 months-1 year
c. 2-3 years
d. 4-6 years
e. 7-10 years
f. Longer than 10 years: __________
Specifically (if know for sure): ____________
4. How would you describe your situation as being homeless?
a. First time being homeless
b. Consistently homeless, living from temporary place to temporary place
c. Homeless, paying rent but then returning to homelessness
d. Homeless, living with family/friends but then returning to homelessness
5. Why are you homeless?
a. Lack of income
b. Family dispute/other family issue
c. Domestic violence
d. Divorce
e. Job loss, lack of education
f. Lack of affordable housing options
g. Incarceration
h. Mental disability or physical disability
i. Drug or alcohol abuse
j. Personal choice
Notes/comments (verbatim):
6. Are you receiving medical care?
a. Yes
b. No
27. 27
7. Are you a veteran?
a. Yes
b. No
8. Are you currently using illicit drugs?
a. Yes
b. No
9. Have you had problems with illicit drugs or alcohol in the past?
a. Yes
b. No
EMPLOYMENT:
10. Are you currently employed?
a. Yes
b. No
11. Do you wish to seek employment?
a. Yes
b. No
12. Are you aware of resources that can help you seek employment?
a. Yes
b. No
13. How long have you been unemployed?
a. 0-5 months
b. 6 months-1 year
c. 2-3 years
d. 4-6 years
e. 7-10 years
f. Longer than 10 years: __________
Specifically (if know for sure): ____________
14. Are you on any type of welfare assistance (Medicare, Medicaid, WIC, etc.)?
a. Yes
b. No
If YES, please name: _____________
15. How much income do you receive each month?
a. $0-$50
b. $51-$100
c. $101-$250
d. $251-$500
28. 28
e. $501-$750
f. More than $751
HOUSING:
16. Have you paid rent or owned a home at any point in the past?
a. Yes
b. No
If YES, what happened that made you become homeless?
17. In the past 60 days, did you spend any nights:
a. in an apartment or home that you owned or paid rent on in the past 60 days? Yes /
No
b. in a hotel or motel room that you paid for? Yes / No
c. in a boarding house, halfway house, or board and care facility? Yes / No
d. in the home or room of a family member? Yes / No
e. in the home or room of a friend or acquaintance within the past 60 days? Yes / No
f. in a hospital? Yes / No
g. in a state institution for mental health services? Yes / No
h. in jail or prison in the past 60 days? Yes / No
i. on the streets, abandoned building, park, etc.? Yes / No
If YES to any items in a-i, ask 17a. If NO, go to item 18.
9a. You said you spend nights in (ITEMS CODED YES IN 16) in the past 30
days. Now I would like to ask you about specific nights spend in these places.
How many nights did you spend at (______) in the past 30 days?
How many nights did you spend at (______) in the past 30 days?
How many nights did you spend at (______) in the past 30 days?
How many nights did you spend at (______) in the past 30 days?
18. Are you aware of the Housing First Program?
a. Yes
b. No
19. In the past 12 months, have you tried looking for a home/permanent place to live?
29. 29
a. Yes
b. No
20. In the past 6 months, has a staff member from a shelter, drop-in center, or other service
talked to you or given you advice about how to get a permanent place to live?
a. Yes
b. No
21. In the past 6 months, has a staff member helped you make contact with someone who
managed an apartment, housing complex, or house to see if you could get into permanent
housing?
a. Yes
b. No
22. Would you like a permanent place to live?
a. Yes
b. No
If YES, what barriers stand in your way? (record verbatim):
If NO, why not? (record verbatim):
PERSONAL CHOICE:
23. So you indicated you are homeless as a testament to free will. Do you enjoy being
homeless?
a. Yes
b. No
24. What are the advantages you find of being homeless? (record verbatim):
25. Have you ever wanted to have your own home?
a. Yes
b. No
26. When did you first adopt this approach to living?
27. How do you support yourself financially?
30. 30
Table 1. ANOVA Results
ANOVA
Results
Independent
Variable
Mean
SE
Mean
ECONOMIC
0.289
0.0577
DISABILITY
0.104
0.0400
SOCIAL
0.081
0.0445
DOMESTIC
0.059
0.0319
CHOICE
0.030
0.0231
VET
0.067
0.0343
LIFER
0.111
0.0431