Citation
Anti-homeless laws in U.S. cities

Material Information

Title:
Anti-homeless laws in U.S. cities a quantitative study of three theories
Creator:
Whelley, Collin Jaquet ( author )
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (133 pages). : ;

Thesis/Dissertation Information

Degree:
Master's ( Master of Arts)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Political Science, CU Denver
Degree Disciplines:
Political Science
Committee Chair:
Robinson, Tony R.
Committee Members:
Cheever, Kathryn A. L.
Sphen, Thorsten H.

Subjects

Subjects / Keywords:
Homeless persons ( lcsh )
Homelessness -- Government policy ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
The purpose of this study is to better understand why cities in the United States are increasingly employing anti-homeless laws by testing three established qualitative theories. This study uses quantitative methods to investigate the variance in the total number of anti-homeless laws in a city, with a sample size of 102 cities. This study examines 15 independent variables to test the three established theories in question. The first is a theory of homeless and criminal threat, the second is a theory of neoliberal growth, and the third theory considers the impact of cultural shifts and gentrification within a city. Results indicate that all theories need to be revised, that cultural-political indicators have the most effect on anti-homeless law growth and stagnating or rising property crime rates and stagnant or falling housing values also have a significant effect. However, the majority of variance in the number of anti-homeless laws in a given city remains unknown.
Thesis:
Thesis (M.A.)--University of Colorado Denver. Political science
Bibliography:
Includes bibliographical references.
General Note:
Department of Political Science
Statement of Responsibility:
by Collin Jaquet Whelley.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
898484721 ( OCLC )
ocn898484721

Downloads

This item has the following downloads:


Full Text
ANTI-HOMELESS LAWS IN U S. CITIES:
A QUANTITATIVE STUDY OF THREE THEORIES
by
COLLIN JAQUET WHELLEY
B. A.University of Dayton2006
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado Denver in partial fulfillment
of the requirements for the degree of
Master of Arts
Political Science
2014


2014
COLLIN J WHELLEY
ALL RIGHTS RESERVED
li


This thesis for the for the Master of Arts degree by
Collin Jaquet Whelley
has been approved for the
Political Science Program
By
Tony R. Robinson, Chair
Kathryn A. L. Cheever
Thorsten H. Sphen
April 29,2014


Whelley, Collin Jaquet (MA, Political Science)
Anti-Homeless Laws in U.S. Cities: A Quantitative Study of Three Theories
Thesis directed by Associate Professor Tony R. Robinson.
ABSTRACT
The purpose of this study is to better understand why cities in the United States are
increasingly employing anti-homeless laws by testing three established qualitative
theories. This study uses quantitative methods to investigate the variance in the total
number of anti-homeless laws in a city, with a sample size of 102 cities. This study
examines 15 independent variables to test the three established theories in question. The
first is a theory of homeless and criminal threat, the second is a theory of neoliberal
growth, and the third theory considers the impact of cultural shifts and gentrification
within a city. Results indicate that all theories need to be revised, that cultural-political
indicators have the most effect on anti-homeless law growth and stagnating or rising
property crime rates and stagnant or falling housing values also have a significant effect.
However, the majority of variance in the number of anti-homeless laws in a given city
remains unknown.
The form and content of this abstract has been approved. I recommend its publication.
Approved: Tony R. Robinson
IV


DEDICATION
I dedicate this thesis to the men, women and children who will be sleeping on the
street tonight; to those fleeing violencepersecutionand famine and struggling with
illness; to those whom I have worked with and those I have not; to those we have lost.
Your life and memory lives in those who choose to bare witness to your suffering, your
struggleyour resiliencyand your love. Thank You.


ACKNOWLEDGEMENT
It has been a privilege to be apart of the New Directions in Politics and Public
Policy program at the University of Colorado Denver. I have never been pushed to work
so hard. Thank you to the professors who have challenged, engaged, and taught me. I
would like to offer a special thanks to Kathryn A. L. Cheever, Ph.D and Thorsten H.
Sphen, Ph.D for all your work as part of my Thesis Committee and for the hours you
spent as my professors. I want to thank Tony Robinson PhD for inspiring me to continue
working on issues related to homelessness and poverty, for your help in framing my
questions and help in designing my experiment. One of the best decisions of my
academic experience was to take your class in Seoul, South Korea. My time at the
University of Colorado Denver is an experience that has greatly expanded my desire for
life long scholarship.
Thank you to my friends and family. Thank you Scott for your mentorship,
wisdom and friendship. Thank you to my mother Sarah and father Peter, you have
instilled me with roots of support and wings of wonder that I will cherish as long as I
live. Thank you to my brother Patrick and sister Katy, your genuine interest in my life
and your far-reaching accomplishments have only ever pushed me to further actualize my
own potential. The most important thanks go to my wife Merida. Your constant support
and affection has allowed me to accomplish more than I ever thought possible. Thank
you.
vi


TABLE OF CONTENTS
CHAPTER I. INTRODUCTION 1
II. ANTI-HOMELESS LAW AND THEIR PROLIFERATION 14
Anti-Homeless Legislation 14
Theory of Threat 22
Neoliberal Economic Dynamics and Anti- Homeless Laws 29
Cultural Shift Theory 35
III. METHODOLOGY 45
Threat of Homelessness and Crime Test 48
Neoliberal Growth Test 49
Cultural Shift Test 50
IV. RESULTS 52
Descriptive Statistics 52
Regression Model Results 57
Independent Variaole Regression Results 58
Theory of Threat Test Results 60
Neoliberal Growth Test Results 68
Cultural Shift Test Results 72
V. DISCUSSION 77
VI. CONCLUSION 104
REFERENCES no
APPENDIX 117
A. Results Summary 117
B. Regression Summary 118
C. List of Independent Variables 119
D. Standardized Results 123
E. Pearsons Correlation Matrix 124
vii


LIST OF TABLES
TABLE 1. Categories of Anti-homeless laws 14
2. Anti-homeless Law Categorization 46
3. Pearsons Correlation Results 54
4. Overall Regression Model Results 58
5. Multiple-Regression Results 59
6. Threat Summary Statistics 68
7. Neoliberal Summary Statistics 71
8. Culture Shift Summary Statistics 75
9. Summaries of Revised Hypotheses 100
vm


LIST OF FIGURES
FIGURE
1.Residual plot of Ave Homeless Population
and Anti-Homeless Laws
64
IX


CHAPTER I
INTRODUCTION
On May 14th, 2012, Denver Colorado passed a city-wide "Unauthorized
Camping ordinance. This ordinance prohibits the use of anything but clothes on ones
body to protect against the elements for the purpose of shelter, and violations are
punishable by up to a $999 fine. The city of Denver, as well as the surrounding metro
area, continues to have more homeless people than shelter beds and the availability of
shelter beds are not always adequate for those with physical disabilities, chronic mental
illnesses, substance abuse disorders, for transgendered persons, or those without
transportation (NAEH, 2012; Denver City Council, 2012a; Robinson, 2012).
Criminalizing the act of sheltering with a blanket or any other protection against the
elements in a city without universal access to shelter is in fact criminalizing a group of
people who cannot or will not access shelter.
This ordinance was passed twenty-one days after the United Nations commended
the United States Department of Justice and United States Interagency Council on
Homelessness (USICH) for withdrawing support for laws criminalizing behaviors vital to
survival for those living on the street, while encouraging more effective and humane
policy recommendations (UN OHCHR, 2012). Despite the UN recommendation to resist
such laws, and despite the federal government withdrawing support from such laws,
many cities like Charlotte NC, Tampa FL, and Denver CO, continue to pass camping
bans and other laws that criminalize survival behaviors of homeless people.
1


Anti-homeless laws are laws and ordinances that criminalize the behavior of
people living in public places (such as sleeping, sitting, covering oneself, or even
urinating); which would not otherwise be illegal if a person had a private place to conduct
those activities (NLCHP, 2011). In other words, these laws criminalize life-sustaining
activities of those living on the street. The National Law Center on Homelessness and
Poverty (NLCHP) (2011) surveyed 234 cities in the United States and found that forty
percent of cities ban campingin some areas of the city, sixteen percent ban camping
city-wide, thirty-three percent ban sitting and lying in particular locations, fifty six
percent ban loitering in particular locations, twenty-two percent of cities have a city-wide
ban on loitering, fifty three percent of cities ban begging in particular locations, and
twenty-four percent of cities ban panhandling city-wide. In addition, NLCHP found
from a study of 188 cities between 2009 to 2011 that there was a seven percent increase
in laws prohibiting panhandling, a seven percent increase in laws prohibiting camping in
particular places, and a ten percent increase in loitering bans in particular places (NLCHP
2011).
A question arises as to the reason why laws that criminalize homeless behaviors
are increasing across cities in the United States, even while the United Nations, the U.S.
Federal governmentand countless homeless advocate and aid organizations have spoken
out in opposition specifically to anti-homeless laws. Why do cities pass these laws? The
purpose of this study is to better understand the political, cultural, and economic, factors
that push contemporary cities to adopt anti-homeless legislation. Extensive qualitative
research has been conducted highlighting the growth of anti-homeless laws and
conjecturing as to the reasons for that growth. However, much of the qualitative work has
2


not been supported by quantitative methods. This study will use quantitative methods to
build on and test the theories that already exist in the literature regarding the promotion
of anti-homeless laws.
Denver City Councilwoman Susan Shepard offered testimony and an explanation
as to why she believed the most recent urban camping ban was passed in her city. The
Councilwoman disagreed that the camping ban was a result of compassion for the
homeless and joined the minority in opposition. She stated,
Despite all the talk about compassion this is not really about compassion. How
many times have we heard the words we have to keep our Denver core strong
(or) we cant weaken our downtown business sector or everything will fall
apart? So is this about compassion or is it an economic strategy that we are
talking about here? Personally..! think...this is code. Homeless are bad for
business and they are ugly and dirty and stinky and they make our streets look bad
and we are sick of dealing with them. So lets just sweep them under somebody
else5s mg...(Denver City Council, 2012b).
Councilwoman Shepard5 s assessment, while extremely critical of business
interestshypothesizes that Denvers camping ban is the result of an economic agenda.
This hypothesis is not without merit. Over the past two decades the United States has
experienced a back to the citymovement where wealthy professionals leave suburban
areas for downtown urban centers. During this same period, as wealth and new urban
professionals increasingly move to cities, laws criminalizing behaviors of homeless
populations, referred to as anti-homeless laws, have also spread (Florida, 2012; Mitchell
1997).
Anti-homeless laws are an attempt by city officials to control the development
and shape of the city as it grows. Development and re-development projects are taking
place at great speed across the globe. Many cities in the United States are following the
same trajectory of population and development growth. The new spaces that are created
3


as part of the new urban growth dynamics, however, are not always produced in a way to
meet the needs of entire populations. More often, the needs of the multinational
corporations and power brokers who control investment dominate the creation of new
urban spaces, and as a result, their needs are met before the needs of broader society.
Turner (2002) writes that public property ownership is shifting to private
investors who tend to constrict the democratic use of public space, as the new private
spaces are not created for everyone. Turner (2002) focuses on public-private partnerships,
such as tax increment financing (TIF) and business improvement districts (BID), where
public finances subsidize private development in an effort to speed up development.
These partnerships result in the private control over large swaths of cities that are at times
controlled by a single development firm. One example of explicit transfer of assets from
public control to private control is the sale of public housing to private property owners
and development corporations. Similar examples exist in New York Citys Zuccotti Park
and Denver Colorado's Eddie Maestas Park. Zuccotti Park, the creation of a public
private partnership, is controlled by a private entity that has some control over the
public use of that space to the extent it can define what public means. Denver
Colorados Eddie Maestas Parklocally referred to as Triangle Park is now in the
control of Denver Urban Gardens. Denver Urban Gardens has funding and city support to
completely replace the park, that for years gave rest to homeless individuals waiting for
food and shelterwith an urban garden designed to beautify Denvers skid rowMeyer
2013).
Harvey (2012) argues that, "The actually existing right to the city, as it is now
constituted, is far too narrowly confined, in most cases in the hands of a small political
4


and economic elite who are in a position to shape the city more and more after their own
particular needs and hearts desirep. 24). Meaningthe development firms and private
entities that have gained control of city blocks and parks are now able to regulate the use
of the space that they control. They regulate the use of space to coincide with their
private interests (such as interests in carefully managed spaces attractive to shoppers and
tourists) and are legally able to ignore broader social interests (such as the interest in
providing public spaces where even very low-income people can gather).
The increasing privatization of urban space is part of a growing global and U.S.
trendwhich can be called urban neoliberalism. As part of a pro-market turn in
governing philosophy felt across the globe, urban neoliberalism involves the opening up
of new market forces in urban spaces, the devolvement and dismantling of the welfare
and redistributive systems, such as public housing, and a political and economic
commitment to attracting private capital investment above all else (Mitchell, 2003).
The neoliberal cities of the United States are seeking to attract wealth through
public subsidies of corporate projects, as well as through marketing strategies to brand
cities as clean, safe and exciting places to play (Wyly & Hammel, 2003). The neoliberal
goal is to build a clean, pure, safe environment for tourists, shoppers and conventioneers,
as well as spaces for those with expendable income to spend in new retail, sports,
entertainment and luxury housing venues (Mitchell, 2003; Amster, 2003, Sennatt, 1970;
Ferrell, 2001). As part of the neoliberal turn, U.S. cities are also trying to attract direct
foreign investment in developments and post-industrial industries such as high tech
biotech and financial services (Florida, 2012; Krizmk, 2011).
5


Some scholars argue that to accomplish these goals more and more cities are
actively participating in excluding portions of their own populations from access to urban
spaces, and use the growth of anti-homeless laws as evidence. Those promoting this
perspective argue that cities are choosing to ban behaviors of the homeless, and seeking
to rid themselves of homeless peoples presencein order to create the spaces attractive to
capital investment. Mitchell (1997) writes,
In city after city concerned with 'livability/ in other words, making urban centers
attractive to both footloose capital and to the footloose middle classes, politicians
and managers have turned to...a legal remedy that seeks to cleanse the streets of
those left behind by globalization and other secular changes in the economy by
simply erasing the spaces in which they must live...(p. 305).
The legal remedies Mitchell (1997) refers to are anti-homeless laws.
Today the trend continues; the prevalence of anti-homeless laws continues to
increase across U.S. cities (NLCHP, 2011; USICH, 2012). Mitchell (1997), Smith (1996)
and others argue that as new city development is centered on building world-class
consumption spaces and purified playgrounds for adults with expendable incomes, power
brokers in U.S. cities are pushing to clear the homeless off the streets. However,
Mitchells (1997) and Council Woman Shepards (2012) assessments of the economic
causal factors pushing cities to pass anti-homeless laws are not the only theories existing
in literature and have, for the most part, not been supported with quantitative evidence.
The economic focus of Mitchell (1997) and Shepard (2012) may not properly
consider other possible contributing factors within cities concerning the promotion of
anti-homeless laws. Other factors, which might drive the increase in anti-homeless laws,
include crime rates, political party strength, and cultural changes. Cultural changes refer
to shifting populations of a city that may be driving support and advocacy for a particular
6


policy. The shifting makeup is tracked by observing changing racial groups, education
levels, migratory patterns, wealth, and how groups are integrating or segregating across
the urban landscape. The growth, decline or dispersion patterns of a particular cultural
group may have an effect on the level of support for or acceptance of particular political
policies. Alternative explanations for the contemporary rise of anti-homeless laws are
important to examine because anti-homeless laws are not a new creation in the same way
that neoliberal politics have dominated contemporary urban history.
Anti-homeless laws, like homelessness itself, have a long history in this country.
While Brenner and Theodore (2002) indicate that neoliberal policies and reforms began
to blossom across US Cities in the 1970s and 1980s with the elections of Prime
Minister Margaret Thatcher and President Ronald Ragan, Wasserman and Clair (2011)
indicate that Anti-homeless laws or poor laws can be traced back to the Middle-Ages
and were present in the United States shortly after the founding of the country. Anti-
homeless laws are not a new political phenomenon, but a re-emerging one. A historical
perspective suggests that there may be causal factors besides economic policy that drive
the growth of anti-homeless laws.
Mitchell (2011) argues that, historically, vagrancy laws were used to control a
large population of poor, escaped and freed slaves, after emancipation, when reserve
labor accumulated as people moved to the cities. Similarly, during industrialization,
especially during times or recession and depression, large numbers of single men as well
as families constructed shantytowns and encampments along rivers and in parksand they
were often the subject of past anti-homeless laws (Mitchell, 2011).
7


Wasserman and Clair (2011) argue that prior to the recession in 1890, the
homeless or hobos were understood as a migratory workforce with a lifestyle that
was somewhat glorified. As migratory employment vanished and the recession of 1890
devastated the demand for migratory work, those traveling workers became stuck in one
place or another. Wasserman and Clair (2011) interpret this reality of a growing, semi-
permanent reserve army of labor in many cities as related to the deterioration in the
publics conception of homelessness as well as to a re-emergence of poor laws
While large shantytown encampments are much less common in America today,
anti-homeless laws have again re-emerged. Mitchell (2011) writes that the re-emergence
of anti-homeless legislation in the 1980s and 1990sa re-emergence that continues
today, coincided with neoliberal changes in the way government welfare was
administeredas hard financial times created a cultural atmosphere of compassion
fatigue for the homeless (p. 943).
Conversely, Wasserman and Clair (2011), borrowing from Axelson and Dail
(1988), argue that the conceptions of homelessness as deviant, dangerous and needing
behavioral control began well before the political and economic developments of the
1980s and 1990s. Wasserman and Clair (2011)writeCurrent conceptions of
homelessness are most directly rooted in the negative attitudes developed in the period,
when homelessness transformed from a semi-legitimate nomadic lifestyle to a public
nuisance that offended the sensibilities of wealthier citizensp.9). Wasserman and Clair
(2011) do not explicitly argue that poor laws are a result of cultural changes brought on
by increased homelessness rates and economic stagnation, but the foundation is laid for a
deeper examination of various cultural views of homelessness and how they might drive
8


anti-homeless legislation, in addition to considering the economic motivators of such
laws.
Wasserman and Clair (2011) argue that cultural conceptions of homeless people,
connected to a growing lack of mobility among growing homeless populations at the turn
of the centurymay share a relationship with the explosion of poor laws during that
time. Advancing a similar cultural theory for anti-homeless sentimentMurphy and
Tobin (2011) argues that the public conception of homelessness concerning causes, and
cultural assumptions about who is homelesshave swung from structural causes and
supportive policies to personal responsibility and punitive measures.
Although there has been no quantitative study regarding the connection between
the public conception of homelessness and anti-homeless legislation some qualitative
assessments argue that a view of homelessness as largely voluntary promotes municipal
efforts to criminalize homelessness (Whelley, 2013). While economic factors may play a
large role in shaping anti-homeless legislation, further study is needed to examine
cultural, political and other factors that might also influence the growth of anti-homeless
laws.
Mitchell (2011) offers an effective qualitative explanation of how challenging
urban political and economic environments fostered the growth of anti-homeless laws in
the 1980s and 1990s. Howevereven as the economy recovered in the 1990s and
through the boom and bust cycle of the 2000sthe prevalence of anti-homeless laws
continued to grow. The specific reason(s) why anti-homeless laws are growing in U.S.
cities still remains somewhat of a mystery, since those laws have tended to expand
whether economic times are good or bad, in both flourishing and declining U.S. cities. In
9


addition, Mitchell5 s (2011) understanding of the neoliberal determinants of the recent
burst in anti-homeless laws does not necessarily contradict Wasserman and Clairs (2011)
cultural perspective, but cannot easily explain why poor laws similarly expanded in a
very different timethe turn of the century.
The following literature review will seek to expand upon several unproven
theories of why more and more cities are seeking to criminalize the survival behaviors of
those living on the street. The literature review will be separated into two sections. First,
the nature of anti-homeless laws will be explained in greater detail, and the thesis will
offer a brief introduction of two differing perspectives concerning support for and
opposition to these laws.
The next section will focus on three different theoretical explanations for the
recent increase in anti-homeless laws. The first explanation, which I call the "threat
hypothesis will focus on the possibility that growing homeless populations and crime
rates instigate a perception of threat in the minds of some urban residents concerning the
homeless. This perception of threat leads localities to pass more laws restricting the
behaviors and presence of homeless people in public locations. The second theoretical
explanation will discuss neoliberal economic growth as a driving force behind the growth
of anti-homeless laws, in that increased global competition for capital investment is
alleged by some to lead cities to be more restrictive of homeless behaviors, in an effort to
create downtown spaces more attractive to international investors. The last explanation
will focus on growing racial and political groups, namely white and affluent
professionals, which may be shifting the city culture towards fear of urban disorder. This
shift in culture instigates a culture clash between new white and affluent residents and
10


urban street culture that existed prior. The clash creates cultural panic and pushes the
cities establishment toward the regulation of public spaces (i.e. anti-homeless laws). As
urban areas attract new residence and racial diversity expands across those same urban
landscapes, the cultural makeups of cities are changing. Political tools to control the
public landscape may be supported out of a reaction to cultural disorder. Cultural fear
may be connected with the threat and neoliberal theories; however, little is known about
possible effects of shifting culture on anti-homeless laws.
Most of the existing literature is founded on qualitative analysis, in that scholars
have presented logically compelling theories explaining the growth in anti-homeless
laws, supported by analysis of the discourse of urban officials, etc. But there has been
very little quantitative testing of these theoriesa gap that this thesis will work to
address. The three established theories investigated in this study are tested using a
quantitative method through multiple regression analysis. The number of anti-homeless
laws in one hundred and two American cities will be compared with possible correlating
data associated with the three theories presented above.
The conclusion of the quantitative analysis offers some surprises. In the
qualitative scholarly literature, the neoliberal growth theory is one of the most common
explanations accounting for the rise in anti-homeless laws. For this study, the neoliberal
theory was hypothesized to hold the strongest relationship to anti-homeless laws.
However, contrary to prediction, the evidence shows that, economic growth measures
actually have the least powerful relationship with anti-homeless laws. Furthermore,
slower growth rates seem to have more effect on anti-homeless laws than more robust
11


growth rates. It is hard to attribute the rise in anti-homeless legislation to a surge in
neoliberal economic growth patterns, therefore.
Another common theory is that rising rates of homelessness and rising crime rates
create a threat in peoples mindshelping to explain a surge in anti-homeless laws.
Again, the hypothesis of this thesis, that the threat theory was expected to hold strong
relationships with anti-homeless law, is largely challenged. The data do not support this
hypothesis in its entirety. Evidence suggests that the number of homeless persons in a
given city and average crime rates have little relevance in a discussion related to the
growth in homeless laws. At the same time, anti-homeless laws may be positively
correlated with falling homeless rates and falling crime rates (however, they are
also correlated with rising property crime rates).
In the end, this thesis will argue that the quantitative data show that the cultural
shift theory proves to have the most statistically significant relationship with anti-
homeless laws and not in the way first predicted. Rather than complimenting the
neoliberal theory with corresponding demographic trends indicating gentrification and
segregation of wealth and race, the results indicate a culture shift of a different kind.
According to the results of this study, the number of anti-homeless laws in a city holds
the strongest relationship with a positive shift in Democratic political voting growth (i.e.,
the stronger the rate of growth in the Democratic vote, the more anti-homeless laws
result). This finding is supported by smaller, but significant, cultural associations between
a decrease in urban segregation and high percentages of non-white residents, with higher
numbers of anti-homeless laws.
12


Changing cultural landscapes, particularly in the ways listed in the preceding
paragraph, may be instigating fear and insecurity that pushes cities to enact policies to
control the public space. Public discourse may remain centered on the homeless and some
may insist that this is a homeless issue, however the trends that associate with anti-
homeless laws are not a reaction to absolute or rising homeless populations. The reasons
why cities are implementing anti-homeless laws are disconnected from homelessness
because anti-homeless laws are reactionary policies to cultural shifts, slow economic
growth and stagnant or rising property crime rates. This conclusion contradicts much of
the literature and complicates the interpretation of discourse centered on the economy and
homelessness.
13


CHAPTER II
ANTI-HOMELESS LAW AND THEIR PROLIFERATION
Anti-Homeless Legislation
Anti-homeless laws are city ordinances that criminalize the behavior of
individuals living in public spaces (NLCHP, 2011). The National Law Center on
Homelessness and Poverty (NLCHP) and the United States Interagency Council on
Homelessness (USICH) group these laws into six main categories, as can be viewed in
Table 1.
Table 1:Categories of Anti-Homeless Laws
Categories of Anti-homeless laws as defined by NLCHP (2011) and USICH (2012)_____________
(1) Enactment and enforcement of laws that make it illegal to sleep, sit, or store personal
belongings in the public spaces oi cities without sufficient shelter or affordable housing; _
(2) Selective enforcement against homeless persons of seemingly neutral laws, such as loitering,
jaywalking, or onen container ordinances; ________________________________________________
(3) Sweeps of city areas in which homeless persons live in order to drive them out of those areas,
frequently resulting in the destruction of individuals5 personal property, including important
^personal documents and medication;_______________________________________________________
(4) Enactment and enforcement of laws that punish people for begging or panhandling in order to
move poor or homeless persons out of a city or downtown area;
(5) Enactment and enforcement of laws that restrict groups sharing food with homeless persons in
jmblic spaces ___________________________________________________________________________
(6) Enforcement of quality of life ordinances related to public activities and hygiene (e.g.
public urination) when no public facilities are available to people without housing. (NLCHP,
2011; USICH, 2012) __________________________________________________________________________
The most prevalent anti-homeless laws are panhandling, loitering, and camping
ordinances banning these kinds of behavior in particular places (NLCHP, 2011).
Citywide ordinances that ban the same behaviors across the entire city are less common,
but much more restrictive. The NLCHP (2011) found that out of 234 cities, sixteen
percent banned camping citywidewhile forty percent of the cities banned camping
14


in particular places.1 2 3
NLCHP (2011) also found that the prevalence of anti-homeless laws continues to
increase. As mentioned before, from 2009 to 2011 there was a seven percent increase in
anti-panhandling laws nationwide, a seven percent increase in camping bans for
particular places and a ten percent increase of loitering bans in parts of a city. The
NLCHP (2011) also surveyed people experiencing homelessness as well as homeless
service providers across 26 states in an attempt to better understand if these laws were
being enforced. Fifty-five percent of homeless respondents reported being cited or
arrested for camping or sleeping in a public place.
As anti-homeless laws grow in U.S. cities, the theoretical understandings of these
laws, as well as the understanding of the causes of homelessness, remain divided.
Discourse surrounding the support of and opposition to anti-homeless laws is generally
divided by opposing theoretical frameworks of structure and agency as they apply
to the causes and solutions of homelessness. Neale (1997) writesA structural
1 Camping" refers to the act of sheltering ones self with items such as blankets,
cardboard, newspaper, etc. This is different than Sleeping. Sleeping refers to the -
act of sleeping in public and not necessarily to the use of items to protect from the
elements.
2 The Creative Class refers highly educated, highly mobile, typically young
professionals working in creative or knowledge based industries (Florida, 2012).
3 In an attempt to identify geographic trend in anti-homeless laws I included a
measure of temperature. Intuitively, if locations are cold and homeless people are
outside during the harsh weather, a theory of agency, a theory that supports anti-
homeless laws, would be undermined by the fact that surviving in the cold is not
something one would likely choose to do. The average temperature for all cities in the
sample was 58.36 degrees Fahrenheit with an average January temperature of 38.93
degrees Fahrenheit. Temperature was an important variable to consider in order to rule it
out as a possible confounding variable for the rest of the study. In the end, colder cities
varied in the same way that warmer cities varied in terms of the number of anti-homeless
laws. For example, in 2011 Corpus Christi had 8 anti-homeless laws (below average)
with an average yearly temperature of 72 degrees Fahrenheit and an average January
temperature of 5b degrees Fahrenheit. Jacksonville FL, with comparable temperatures
had 15 anti-homeless laws in 2011.With respect to colder cities such as Detroit,
15


explanation of homelessness locates the reasons for homelessness beyond the individual,
in wider social and economic factors (p.49). A structural approach holds that the
structure of society is responsible for making people homeless through such mechanisms
as a poorly functioning economy, inadequate affordable housing stock, and soaring
medical expenses that drive people into bankruptcy. Under this structural perspective
because society is responsible for creating the environment causing people to be
homelessit is therefore societys obligation to change the environment in such a way as
to allow homeless people to improve their living conditions and become self-sustaining.
And, because the homeless are largely not to blame for their situation, the behaviors that
they must partake in on a day-to-day basis, are unavoidable behaviors of survival. Any
attempt to criminalize their behaviors of survival is effectively criminalizing the person
who has no other choice but to be homeless. Mitchell (1998) writes in support of the
structural viewpoint: if homeless people can only live in publicand if the things one
must do to live are not allowed in public space, then homelessness is not just
criminalized; life for the homeless is made impossiblep.10).
Converselyan agency understanding of homelessness allows one to argue that
anti-homeless laws do not criminalize people, but rather the behavior in which these
people choose to partake. Wright (1997) argues that authoritarian responses to
homelessness, like anti-homeless laws, emerge from an understanding that homelessness
is a result of poor and illegitimate choices. Therefore, society has less of an obligation to
ensure access to housing or shelter. Most importantly is the fact that agency arguments
support legislative tools that attempt to push people toward culturally appropriate
16


public behavior because the inappropriate behavior of living on the streets is viewed as
voluntary.
Wasserman and Clair (2011) expand on this understanding of agency by
including arguments drawn from human service, medical and substance treatment
perspectives. These perspectives may initially contradict discourses of agency, but
inevitably provide cover and ammunition for agency and anti-homeless law supporters
alike. For example, individual pathological explanations of homelessness, such as
substance abuse or mental illness, place the cause of homelessness in the deficiency of
the individual. This deficiency is often understood as a wrong choice (substance abuse) or
due to a deficient mental health or capacity (for which individuals are also often blamed
forat least to the extent they choose to live on the streets and not get adequate personal
help). Those who choose not to seek treatment allegedly choose their substances or
their illness over shelter and help. In other words everyone, even the sick, is responsible
for his or her own situationfrom an agency perspective.
The view of human service workers and mental health professionals may often be
structuralin that many of these people do not seek to blame homeless people for their
situation, but the argument used to support a medicalized viewpoint is often adopted and
used by governing officials in support an argument of agency that allows for new laws
restricting homeless behavior in public places (Wasserman and Clair, 2011, Mitchell
2011;1997). Even the common substance treatment slogan take ownership promotes
this agency understanding (Wasserman and Clair, 2011).
While taking ownership of ones health an wellbeing is vital to ending
homelessness on an individual basis, the barriers to becoming self sufficient and
17


empowered to make healthy choices are complexities that are overlooked by those who
mostly understand homelessness as a choice.
Lyon-Callo (2000) explains the medicalization conundrum stating
On the one hand, recent efforts may have facilitated increased services to reform,
treat, and retrain individualized homeless people, and such efforts to improve the
lives of some individuals who are homeless. On the other hand, however, the
continuum of caremedicalized and individualized) approach also does not
fundamentally address questions of access to and distribution of resources in the
community., disease within the discourses of helping actually obliterates
discussion of alternative explanations and thus hinders developments aimed at
resolving homelessness through altering class, race, or gender dynamics. When
homelessness is individualized and medicalized, those concerns remain peripheral
to the central work of normalizing perceived shortcomings or deviancy within
homeless people (p. 330).
Agency perspectives stem from an understanding that homeless people are
homeless because they are deficient and make bad choices. In this way the agency
perspective logically supports the implementation of anti-homeless laws to help reform
homeless behavior. The normal perception of shortcomings and deviancy also supports a
view of homelessness as criminal. If homeless are considered deviant and criminal, as
Lyon-Callo, (2000), Wright (1997), Amster (2003) chronicle, than there emerges an
argument concerning the degree of responsibility for deviant behavior and the right of the
public to restrict it. Mitchell (2003; 2011) names this dichotomous understanding of
homelessness a discourse of deservingancTundeserving poor Mitchell (2011)
connects the undeserving perspective with anti-homeless laws by stating
The response was a criminalization of homeless people in many cities. Laws were
passed that outlawed everything from sleeping outdoors, to sitting in sidewalks, to
free food giveaways. A traditional division between deserving poorwomen
children, and those who behaved themselves in ways dominant society deemed
sufficiently grateful to charity) and undeserving poormen of working age
those who lived on the streets or in encampments and refused to enter shelters or
rehabilitation programs) was reassertedwith a vengeance (934)
18


Once the behavior of the individual homeless person is deemed deviant, than the
discourse of personal responsibility and undeservednessforms and from that
understanding a logical progression toward anti-homeless legislation may develop. The
most important conclusion to take from these arguments is that, according to the agency
perspective, homeless people are predominantly the cause of their own situation.
Homelessness is the result of poor and illegitimate choices. From this perspective,
policies regulating the deviant behavior of the homeless people are justified.
Believing that homeless people are broadly responsible and are principle agents of
their own homelessness supports the viewpoint that people are voluntarily homeless and
their behaviors may be viewed as criminal because those behaviors (such as sleeping or
sheltering in public places) are inconsistent with the goal of social order or the goal of
city development. By prohibiting bad behaviorthe city forces people to make better
choices. Sawhill (2003) writes "unless the poor adopt more mainstream behaviors, and
public policies are designed to move them in this direction, economic divisions are likely
to growp.1).SimilarlyMatthew Arnold stateswithout order there can be no society;
and without society there can be no human perfectionas quoted in Mitchell2003, p.
14).
Many homeless advocates, mental health and service providers, the Obama
administration and the United Nationswho all offer a structural perspective on
homelessness, criticize this agencyapproach, and call for the end to anti-homeless laws
(NLCHP, 2011; USICH, 2012; UN 2012). In addition, recent scientific evidence suggests
that authoritarian responses to homelessness are ineffective and counter-productive to the
goal of decreasing the presence of homelessness in a city (USICH 2012, Cooter, Meanor
19


& Soli, 2012;). The USICH (2012) cites NLCHP (2009) and Roman and Travis (2004)
indicating that,
Rather than helping people to regain housing, obtain employment, or access
needed treatment and services, criminalization creates a costly revolving door that
circulates individuals experiencing homelessness from the street to the criminal
justice system and back. Sweeps can also result in the destruction of the personal
property of people experiencing homelessness, including identification documents
and medication. It can be much more difficult to secure employment, benefits and
housing with a criminal record. Many of these measures include criminal
penalties for their violation; therefore, they actually exacerbate the problem by
adding additional obstacles to overcoming homelessness (p 8).
In this same tradition, a Denver Colorado community group, Denver Homeless
Out Loud (DHOL), recently produced an evaluative report on the effects of the urban
camping ban on the homeless population. DHOL surveyed over 500 homeless individuals
across the city. The conclusions of the report indicated that since the city passed a
restrictive ban on public camping most respondents have not been able to access
dependable shelter...(and) the majority of respondents say their life has become more
challengingmore stressfuland less safe since the ban was enactedRobinson2012
pp.8-9). The report also indicates that 73% of respondents are turned away from shelter
with some frequency... 66% of respondents who used to live downtown say they now
usually sleep in more hidden and unsafe location(s)...(and) 37% say they have sometimes
chosen not to cover themselves from the elements (such as using a blanket) due to the
camping banRobinson2012, p.8).
Regardless of concerns of counter-productivity, cities across the country continue
to implement anti-homeless laws. The literature indicates that the discourse of agency
and medicalization continues to dominate public debate. Whether or not the discourse of
20


agency has grown is a question that has not been quantitatively examined, but is assumed
and supported by qualitative assessments of the rise of anti-homeless legislation.
Trying to understand why the discourse of agency has dominated public debate is
similar to an attempt to understand why cities are increasingly using anti-homeless laws
because assumptions of agency are fundamental in the promotion of anti-homeless laws.
However, trying to chart the ascension of the agency discourse is a different question that
will not be examined in the present study.
In addition, there is a possibility that some come to the conclusion that anti-
homeless laws are appropriate through an individualized and medicalized understanding
of homelessness. This understanding is not necessarily formed from an agency
understanding of homeless peoples behavior. Ratherit may be argued that causes
beyond a homeless persons control have led him or her into homelessness (such as
mental illness or alcoholism), and yet it may still be wise to have a lattice of local laws
that force that homeless person to seek the kind of indoors help (social workersdrug
and detox counseling) that might improve their lives, rather than making it easy for
troubled people to live on the streets.
For example, behavior modification, a strategy used to set boundaries between
care givers and patientsuses the construction and destruction of boundaries to improve a
patient5 s focus on the most important task of care and wellbeing. Some argue that anti-
homeless laws are these kinds of artificial 'boundaries5 that push people toward help.
Anti-homeless laws may be considered boundaries in the sense that they protect
homeless people from the deviant behavior that created their homelessness or keep
them homeless. Even in this approach, however, the underlying assumptions are that
21


people have the agency to access care or need to be forced into accepting care, as well as
the availability of adequate care and services.
The structural viewpoint opposes anti-homeless laws because they criminalize a
class of people by criminalizing survival behaviors because supporters of this viewpoint
argue homeless individuals are not wholly responsible for the situation in which they find
themselves. The agency perspective supports anti-homeless laws because people either
need to be pushed into making better decisions or their choice of homelessness is
illegitimate and should be criminalized. Beyond the different understandings of
homelessness that support or object to criminalizing homeless behavior, there are
economic, political, and cultural explanations of why the prevalence of anti-homeless
laws has grown.
This next section will focus on three largely qualitative theories that were
developed to explain the reasons why cities implement anti-homeless laws. The three
theories explored here are, first, the homeless threat, which suggests that the growing
presence of homeless people and crime push cities to implement anti-homeless laws.
Second, I explore the possibility that the political and economic effects of the neo-
liberalization of city space might explain anti-homeless laws, as concerns for economic
competitiveness may drive political decisions to criminalize homelessness. Lastly, I will
examine whether the growth of affluent professionals and a growing proportion and
segregation of white residents within a city may have any effect on a citys level of
support for anti-homeless laws.
Theory of Threat
22


The theory of threat argues that anti-homeless laws, tools used to establish and
maintain order in a city, are a function of disorder. Historically, vagrancy laws were used
to control poor populations. Mitchell (2011) argues, speaking about the creation of anti-
homeless laws prior to the 19605s, that rising homeless populations at the turn of the
century were met with cultural panic in the emerging middle class. Mitchell (2011)
writes,
Three crucial results of the rise of this massive reserve army of migratory and
often homeless men (and some women) are particularly important for
understanding contemporary homelessness in America. First, it generated a
serious moral panic among the emerging bourgeoisie, especially in times of
economic crisis when the bourgeoisies own precarious social stability was at risk.
A result, vagrancy laws were reinforced and new policing methodssometimes
quite violentwere introduced to contain and corral the wandering poor while
separating the worthy from the unworthy (p. 936)
These Anti-homeless laws remained in place in many cities until the 1960swhen
a series of federal, state and U.S. Supreme court cases stmck down vagrancy, loitering
and other anti-homeless laws (Mitchell, 2011). Between the 1960s and 1980s, most cities
saw a relaxation of public space anti-homeless lawsincluding groundbreaking cases
allowing even mentally ill homeless people to refuse to go to shelters, as long as they
were not an immediate threat to themselves or others.
Siegel (1997) arguessometime in the 1980s we came to the end of our national
romance with cities and the public spaces that define themp.169). Siegel (1997) also
argues that those romantic and nostalgic notions concerning free and open and creative
public spaces were replaced with fear and avoidance as his conception of culture and
safety within cities deteriorated (Siegel, 1997).
Mitchell (2011) touches on two key variables that are possibly pushing
contemporary cities to implement new anti-homeless laws. The quote above suggests,
23


that prior to the 19805s, there was a relationship between a rise in homelessness, fear
being instilled in the middle-class, and a rise in anti-homeless laws. This theory argues,
echoed by Siegal(1997), that a rise in the homeless population creates a cultural reaction
across the middle-class, which is self-conscious of losing new social status, and is
seeking to implement political policies that will solidify their way of life (Ryan (1976);
Gans 1995; Barak 1991; Kawash 1998). Anti-homeless laws are then a reaction to a real
or perceived threat in that the presence of homeless people threatens the middle and
upper class way of life. Mitchell (2011) also indicates that the threat instigated by the
presence of homeless men and women is amplified during times of national economic
hardship.
Although Mitchell (2011) moves on to hypothesize about other factors that have
pushed cities to implement anti-homeless laws after the 19805s, the theory of threat
warrants further discussion. The perception of homelessness has been and continues to be
associated in the minds of many with criminality. It is this association that is largely
responsible for fostering perceptions of threat.
Amster (2003) writes that for the past six centuries, homelessness has been
associated with criminality. Homelessness may not be seen as a crime itself, but is
considered to be related to crime (Amster, 2003). This understanding of homelessness
remains present in contemporary discourse advocating for laws regulating public space,
like anti-homeless laws. Underneath this discourse, there exists a language of fear of
losing ones city to chaos and disorder.
In this veinWilson and Kellings (1982) famous Broken Windows theory has
significantly influenced public discourse and policy concerning homelessness and anti-
24


homeless law over the past three decades. Wilson and Kelling (1982) developed the
Broken Windows theory during a period of growing crime rates in urban areas in New
York City in the 70s and 80s. As an increase in crime occurred within the same period
as an increase of homelessness in urban areas, Wilson and Kelling (1982) hypothesized
that if a city wanted to significantly decrease major crime, prevent the further
deterioration of the city environment and stop white flight cities needed to focus on
policing minor offences such as the presence of the homeless in public areas. It is minor
offences such as panhandling and public sleeping that Wilson and Kelling (1982) believe
would escalate into larger crimes.
Wilson and Kelling (1982) argued that there is a causal relationship between the
visibility of disorder, a lack of attention to small-scale disorderly behaviors, the
deterioration of community intervention and outlook, and rising crime. Broken Windows
theory hypothesizes that allowing disorder to exist on a small scale creates an
environment whereby disorder becomes more widespread, possibly violent, and
destructive to the community that deteriorates in the midst of disorder. According to this
theory, the presence of disorder does not force people to commit crimes, but allows crime
to flourish because disorder has undermined the populations ability to care for the
environment within which they live. People who pass by care less and less about the
environmentand thereforethe destruction of that environment becomes less offensive.
In addition, those who do care about the environment become discouraged by the
presence of disorder, presume a lack of governability, and lose faith that positive civic
action will create positive change.
25


Hinkle (2014) creates a four-staged flow chart explaining the escalation of the
Broken Windows hypothesis. Stage one is disorder goes untreated and is noticed by
residents stage two is citizens become fearful and withdraw from the community
stage three is informal social control decreases and/or is perceived to be low by
criminals and stage four is disorder and crime increase as criminals increase their
activity in the areap.214). Those who are destructive flourish while those who seek to
build and create must find other locations and markets hospitable to their desires. When
those productive people leave, the city will fail. In short, this theory suggests that minor
crime (or even non-criminal disordercan destroy cities by escalating into major crime
and by dissuading productive populations from staying in the community. Included in
minor offences are the behaviors that homeless people participate in (such as gathering
on street corners, sleeping in public, being in parks after curfews, etc.) as they live in
public spaces.
Wilson & Kelling (1982) support regulating homeless behavior because the
homeless person is an example of disorder. By regulating the homeless person from being
seenthe city is saving itself from destruction. Wilson and Kelling (1982) write:
The citizen who fears the ill-smelling drunkthe rowdy teenageror the
importuning beggar is not merely expressing his distaste for unseemly behavior;
he is also giving voice to a bit of folk wisdom that happens to be a correct
generalizationnamely, that serious street crime flourishes in areas in which
disorderly behavior goes unchecked. The unchecked panhandler is, in effect, the
first broken window (p. 4).
Wilson & Kellings (1982) hypothesis uses a language based on fear and threat.
This same fear and threat, of a city descending into chaos, manifests in discourse that
supports anti-homeless law. In the last 30 years many elected ofncials have shared the
same perception as Wilson and Kelling regarding a homeless threat. City officials, similar
26


to Wilson & Kelling, fear that the homeless presence threatens the future economic and
cultural possibilities of cities (Smith, 1996). Wright (1997), for example, documents how
the officials of Chicago and San Jose mobilized against the homeless around the
discourse of fear.
As an example of city ofricials utilizing the discourse of homeless threat in their
support of anti-homeless lawsDenverColorados newly passed camping ban is a case in
point. Denver adopted a city wide camping ban ordinance on May 14, 2012an
ordinance that restricts homeless people anywhere in the city from sheltering themselves
in any way from the weather with anything other than their basic clothing. Denvers
Mayor, Michael Hancock, and a newly elected uouncilman Albus Brooks, started
working on some form of a camping ban ordinance soon after taking office (Meyer,
2011) based on tneir perception of a rising homeless threat. Councilman Albus Brooks
stated, uWe believe that numbers (the rise of homeless populations in Denver) have been
increasingly alarming... so this is about public safety, sanitation, and the protection of
our right of wayDenver City CouncilApril 24, 2011).The Mayor, voicing a similar
perspectivestatedWe only have one Downtown. We cant afford to lose our city
core(Meyer2011).In support of the camping ban, Denver City Councilman Charlie
Brown also argued that the camping ban is a fight for sanityit is time we fight to
change this culture of chaos in our cityDenver City CouncilMay 14, 2012). These
three elected officials in Denver Colorado spoke about the fear of homeless people
eroding the very fabric of the city environment by their presence.
Similarly, former New York City Mayor Rudolf Giuliani not only explicitly used
Broken Windows theory in his argument in support of the regulation of space and
27


cracking down on minor crimes, but also utilized a language of fear and threat. Giuliani
stated:
We have made the Broken Windows theory an integral part of our law
enforcement strategy. This theory says that the little things matter. As James Q.
Wilson describes itIf a factory or office window is brokenpassersby observing
it will conclude that no one cares or no one is in charge. In time, a few will begin
throwing rocks to break more windows. Soon all the windows will be broken, and
now passersby will think that, not only no one is in charge of the building, no one
is in charge of the street on which it faces... so more and more citizens will
abandon the street to those they assume prowl it. Small disorders lead to larger
onesand perhaps even to crime.... Theres a continuum of disorder. Obviously
murder and graffiti are two vastly different crimes. But they are part of the same
continuum, and a climate that tolerates one is more likely to tolerate the other, (in
Coman & Mocan2005, p. 237).
In addition, scholar Fred Siegel (1997) uses this theory to justify criminalizing the
behaviors of the homeless by equating homelessness with criminality in contemporary
political theory. Siegel(1997), writing about the threat felt by citizens of New York City,
states:
What unnerved most city dwellers, however, was not crime per se but, rather, the
sense of menace and disorder that pervaded day-to-day life. It was the gang of
tough exacting their daily tribute in the coin of humiliation. It was the street 'tax5
paid to drunk and drug-ridden panhandlers... It was the threats and hostile
gestures of the mentally ill making their homes in the parks (Siegel, 1997, p.169).
Siegel(1997), and others argue that the sight of disorder destroys the moral fabric
and the core of cities. Disorder is defined in such a way that equates the sight of poverty
and homeless people with the sight of disorder. Siegel attempts to separate the behavior
itself from the nature of the person, in order to criminalize the behavior without admitting
he is criminalizing the existence of a homeless person; however, it is a half-hearted
attempt. As Wright (1997) suggests, this kind of discourse inevitably proposes that the
homeless people are dangerous to society and must be pushed out of sight (p.16).
28


The language of fear and threat is connected to the presence or the perception of
public disorder. The language of fear and threat concerning homeless persons persists
throughout public discourse and political argument. To explore to what extent such fears
and such discourse actually drive the expansion of anti-homeless laws, this thesis will
quantitatively test whether a growth in local homelessness and crime rates have a
relationship with the implementation of anti-homeless laws. If the threat theory offers
a good explanation for the rise in anti-homeless laws, we should see that as rates of
homelessness and crime actually increase in a city (thus increasing perceptions of
threats), a growth in anti-homeless laws would follow. For this reason, theory suggests
that anti-homeless laws will show a positive correlation with homelessness and crime
rates in cities.
Similar theories have been tested in other political science subject areas. Boushey
and Luedtke (2011) tested a threat hypothesis to explain anti-immigrant laws across state
legislatures. They found a significant relationship between the percent rise in immigrant
populations with a rise in the number of state immigration bills that were directed at
controlling the influx of immigrants in that state (Boushey & Luedtke, 2011).
The threat hypothesis entails an assumption that the threat of rising disorder in the
form of increasing homelessness (or a growth in actual crime rates) will inspire a
municipalitys government to enact anti-homeless laws. As opposed to this approachI
will explore in the next section how anti-homeless laws perhaps are less a function of
growing homelessness and crime, but rather are a function of the political and economic
effects of the neo-liberalization of cities. If no significant relationship is found between
homelessness, crime rates, and anti-homeless laws, then perhaps the language of threat is
29


merely a tool used by political actors and power brokers to drum up support among the
populace for anti-homeless laws that offer perceived financial returns to powerful city
players. However, due to the richness of the literature and records of public discourse
that supports the theory of threat, I hypothesized that the measures of threat used would
hold strong positive relationships with anti-homeless laws. Below I will discuss such
economic factor that may be correlated with a growth in anti-homeless laws.
Neoliberal Economic Dynamics and Anti-Homeless Laws
According to an economic approach, cultural variables, such as the psychological
reaction to the rise in homelessness and poverty, are not sufficient to explain the re-
emerging growth of anti-homeless laws. Exploring underlying stmctural conditions in
the current urban political-economy might supplement a focus on public disorder, in an of
itself.
Cities today are very different places than in the 1960s. Globalization, neoliberal
economic policies, and revanchist regimes have significantly shifted political, economic,
cultural, and institutional relationships of the contemporary U.S. city, and this changing
economic order is argued by some to be the true catalyst of rising anti-homeless
legislation. Revanchism refers to a reactionary and cruel policy of dispossession and
control, enforced by the wealthy and well placed against the social rabble of their day
(Smith, 1996). Revanchist policies are political reactions to perceived economic threats,
more than a reaction to a fear of disorder and crime, per se.
Smith (1996) understands revanchist policies to be less about fear and more about
economic control. Smith (1996) argues that revanchism is a vengeful policy
implemented and supported by the wealthy in order to regain economic control of the city
30


that was taken from them in the 1960s and 1970s and in an effort to make large
profits from controlling the reshaping of urban space.
Taking back inner city spaces from the unrest of the 1960s is not only a cultural
strategy, therefore, but can be a very profitable strategy as rent-gaps are found and
exploited through the upward development of once degraded spaces, previously inhabited
by low-income and homeless people. Peck and Tickell (2002) write, that in the 1980s,
...the nation-state became the principal anchoring point for institutions of
(gendered and racialized) social integration and (limited) macro economic
management, neo-liberalization was inducing localities to compete by cutting
social and environmental regulatory standards and eroding the political and
institutional collectivities upon which more progressive settlements have been
constructed in the past (p. 385).
This shift dispersed power from the federal government to the state and then the
state continued to devolve regulatory power to the municipalities. U.S. cities, therefore,
are now the key locales of competition that compete for capital, whereas the broader
nation-state used to be in previous decades. This period of political and economic change
coincides with Mitchells (2011)observation of the re-emergence of anti-homeless laws
coinciding with the deconstruction of progressive social welfare programs at the
municipal level, amid a poorly functioning economy.
Before I continue to unpack these political and economic arguments, it is
necessary to further define three pertinent terms: globalization, neo-liberalism, and
revanchism. Globalization refers to the growing global interconnectedness of local,
national and international economies (Kahler2004). Neoliberalism refers to political
policies of free market capitalism that allow globalization to enter every part of the local
economy. The trend of neo-liberalization in U.S. cities includes the globalization of local
31


markets including housing and development markets, a trend that leads city planning to
become obsessed with global marketability (Harvey, 2012).
Harvey (2012) explains that the shift toward neoliberalism has turned
municipalities focus away from social welfare policies of the New Deal/Great Society
era and toward attracting wealth and producing spaces desirable to global investors.
Harvey (2012) writes, "the politics of capitalism are affected by the perpetual need to
find profitable terrains for capital surplus production and absorptionp. 5).
Revanchism connects globalization, neo-liberalization and anti-homeless law
expansion together by explaining how the arguably cmel neoliberal policies of economic
elites (such as laws banning public sleeping) are adopted in an effort to compete globally,
and as a specific reaction to a previously more progressive era. Revanchists are
supporters of the global neo-liberalization of cities and control the power of city
governments, turning it toward authoritarian responses to homelessness. City Officials
adopt anti-homeless laws because those laws allow powerful actors in a city to more
effectively to pursue global investment. Revanchists seek a clean and safe environment to
attract global capital under the neoliberal global economy in order to maintain their
quality of life and grow local profit opportunities.
According to this revanchism frameworkmunicipalities controlled by either
party today have joined the neoliberal pursuit of creating clean and purified space, which
has necessarily involved them in a harsh attack on the lives of the very poor. There has
been little study of political parties and whether different parties have differing levels of
support for anti-homeless laws; however, revanchist theory suggests that all major
political parties will support the implementation of anti-homeless laws, since such laws
32


are a reflection of the global era we live in and are driven by structural factors of capital
expansion that affect all political parties.
Evidence of economic neoliberalism (including its harsh revanchist aspects)
around the globe dominates contemporary theory regarding the explanation for the
growing prevalence of anti-homeless laws. Beckett & Herbert (2008) write, cities that
depend upon capital investment, tourism, retail, and suburban shoppers for their
economic well-being, the environment on commercial streets has become the subject of
much official attentionp.17). Furthermoreas cities focus on economic growth
strategies that attract capital and depart from Keynesian policies of the past, they must
develop marketing strategies that support this goal (Beckett & Herbert, 2008, Mitchell
1997; Mitchell 2011).
Similarly, Kriznik (2011) observes that contemporary cities are primarily focused
on a global competition for capital and, as a result, must resort to marketing strategies
that are primarily focused on attracting wealth. The re-creation of spaces, used as
marketing toolsinvolves restructuring the meaning and purpose of space, kriznik (2011)
writesYet by reconstructing the meaning of a placecity marketing not only promotes
its qualities but also legitimates the interests of dominant economic or political groups
(p. 295). These marketing strategies have been expanded to the legal regulation of space,
including a growth in anti-homeless laws.
Beckett & Herbert (2008) write "from a political-economic perspective, the
intensifications of urban social control measures stems from the ascendance of neoliberal
global capitalism and the related transformation of urban economies (p.16). Again
social control measures include anti-homeless laws. The economic perspective
33


understands the rise of anti-homeless laws as a function of the degree to which a city has
neo-liberalized, meaning the degree to which the city is engaged in the pursuit of global
capital.
In this vein, Quigley, Raphael and Smolensky (2001) and Raphael (2010) found a
positive relationship between homeless populations and the median cost of rent and a
median rent to income ratio (meaning that as cities become economically more successful
and experience rising income and rent levels, anti-homeless laws expand). Quigley,
Raphael and Smolensky (2001) found that rental vacancy had a negative correlation with
homeless populations and a positive relationship with a competitive housing market.
Quigley, Raphael and Smolensky (2001) make a strong argument for economic
contributing factors and solutions to homelessness. Quigley, Raphael and Smolensky
(2001) write that their study indicates,
...relatively small changes in housing market conditions can have substantial
effects upon rates of homelessness. Consider, for example, a reduction in the rate
of homelessness by one-fourth. The quantitative results suggest that this could be
achieved in the national sample of housing markets by a one percentage point
increase in the vacancy rate (from an average of 8.4) combined with a decrease in
average monthly rent-to-income ratios from 17.5 to 16.8. (p. 51)
QuigleyRaphael and Smolenskys (2001) findings suggest there may be a
connection between the revanchist policies and metropolitan growth that unintentionally
promote homelessness because there is a connection between the housing market and
homelessness. Testing this theory can be achieved through a national sampling of
measures associated with neoliberal growth and revanchist policies present in Quigley,
Raphael and Smolenskys (2001) articleas well as by considering a range of other
growth measures associated with neoliberal economic growth that may correlate with the
growth in anti-homeless laws. If the level of homelessness is associated with neoliberal
34


growth dynamics, as predicted, there should be a covariance with these kind of economic
indicators. Therefore there should be a relationship between levels of development, the
rise of median rent, other economic factors, and the expansion of anti-homeless laws.
Economic growth indicators were expected to hold the strongest relationships with anti-
homeless laws.
An argument that economic and political factors influence the adoption of anti-
homeless ordinances does not mean that there is no role for cultural explanations. Smith
(1996) writes in this regard that the neoliberal discourse in urban areas often becomes
focused on the ugliness of society, the disorder, the filth, and the violence that takes place
on the street. Therefore, revanchism includes culture in its definition while holding that
the economic factors ultimately drive social control policy such as anti-homeless law. To
better separate the neoliberal hypothesis from the cultural shift hypothesis, however,
further discussion of the neoliberal growth hypothesis will focus purely on economic
indicators.
One significant problem for testing neoliberalism is that it is an elusive variable in
that there is no real test for it. Neoliberalism is a broad term, without a clear set of
operational variables that define it. Although later in this thesis I do attempt to isolate
economic variables that are associated with neoliberal priorities, the limitations of
quantitatively measuring the broad concept of economic neoliberalism cannot be
wholly overcome. Before discussing the design of the experiment developed to
empirically test these theories, I will turn to the final theory explaining the growth of anti-
homeless laws to be explored in this thesis.
Cultural Shift Theory
35


The Cultural Shift theory is unique because it does not usually appear explicitly in
literature discussing anti-homeless laws. Theories of threat and neoliberalism emerge as
the most prominent defenses of anti-homeless laws. However, cultural changes may
occur as demographics shift as cities grow and prosper or decline. These changes may
influence the growth of anti-homeless laws. The separation of the cultural shift theory
from theories focusing on crime rates and economic imperatives is important for this
study because culture changes resulting from neoliberal growth or as the result of factors
not discussed, may be more responsible for the enacting of anti-homeless laws than
variables discussed in the threat and neoliberal sections.
Essentially, the cultural shift hypothesis proposes that affluent professionals and
new white residents moving into cities are culturally afraid of or culturally adverse to
homeless people. This section will discuss three unexamined cultural and political factors
that may be associated with the rise in anti-homeless laws. The first is an examination of
cultural changes associated with the growth in cities of a middle class and a young,
professional, and upwardly mobile demographic that may be grasping for power and
supporting policies that support their private interests and cultural viewpoints. Similarly,
cultural fear of homelessness may grow as a function of new city residents, as opposed to
growth in homeless populations. Ferrell (2001) chronicles a clash of culture between
residents that move to redeveloping areas and the residents of abandoned buildings,
riverbanksand train yards.
The second possibility is that new urban professionals may experience fear when
being faced with poverty and diversity on a larger scale then they are used too and are
grasping to take control of their new environment. This may simply be a fear of urban life
36


and associated disorder, separate from the fear generated by the actual presence of
homeless persons. Sennett (1970) argued that urban residents were retreating from a
respect and admiration for the creative potential of public spaces to a place of fear and
avoidance of difference and change.
The third angle of the cultural shift theory is the possibility that party politics
has a role to play in anti-homeless legislation. Even though political dynamics are usually
kept separate from cultural theories, changing political compositions are a method by
which one can examine shifts in culture. Mitchell (1997) notes that many of the most
restrictive cities in terms of homeless behavior are thought of as Liberal cities. Smith
(1996) charts a much more conservative-led movement to criminalize homeless behavior
and restrict public space. Neither author presents quantitative evidence, however both
present effective qualitative assessments.
There is reason to believe that anti-homeless law has a political component
because their re-emergence was formed from a neoconservative movement charted by
both Smith and Mitchell. Yet Mitchells (1997) work is in some ways contradictory to
Smith (1996), indicating that party politics might have less to do with the passing of such
ordinances as anti-homeless laws. However, both agree that Broken Windows policing
originated from a Conservative-Republican political viewpoint, and therefore, some
influence of party is still expected to remain.
Symptoms of neoliberal economic growth cannot be sufficiently examined
without discussing the cultural shifts that may occur as a city grows. Change that is
associated with growthnamely gentrification, which entails demographic and cultural
changemay help explain the growth of anti-homeless law.
37


Wyly and Hammel (2005) hypothesize that a change in demography that reshapes
cultural and political dynamics within cities may be responsible for the increase of anti-
homeless laws. "Local authorities in any city usually move quickly against street people
doing any of these things; but our reasoning is that the policies are formalized only under
certain circumstances, and that gentrification is one of the processes that helps to broaden
support for explicitcity-wide quality of life ordinancespp. 9-10). If gentrification is
one of the processes that broaden support for anti-homeless laws then there is an
implication of cultural change occurring as urban upper classes grow in a city. As a city
gentrifies, the new influx of residents and reorganization of where they live, may create a
culture supportive of anti-homeless laws.
Wyly and Hammel (2005) define gentrification by stating that
Gentrification is fundamentally about the reconstruction of the inner city to serve
middle- and upper-class interests. When it avoids direct displacement, the process
usually involves middle-class or developer subsidies that cannot be seen in
isolation from cutbacks in housing assistance to the poor and other attacks on the
remnants of the welfare state (p. 5).
Gentrification involves economic change and demographic change. Some of these
changes may include changes in income levels, racial makeup and segregation patterns in
a city. But as these factors change, there are likely to be powerful cultural changes in a
city as well. Ferrell (2001) chronicles the shift and clash of culture between residents of
abandoned public spaces, homeless, graffiti artists, and other street dwellers and the new
residents of redeveloped lofts and visitors of sports complexes that replaced them.
Ferrell (2001) also indicates how the new residents of renovated spaces, who
moved to the new location to find an authentic and cultural urban experience
immediately seek to cleanse the building and surrounding area of the culture that existed
before their arrival, namely the street culture. New residents inevitably support more
38


development projects to fill the blighted areas surrounding them and fill the abandoned
public spaces with new private cultural spaces. The new spacesonce home to the
homeless and runaway youthbecame intolerant of street culture.
Ferrell (2001) also suggests that the clash was not simply one sided. Graffiti
artists and persistent squatters fought to maintain some hold on their old stomping
grounds by moving to unseen parts and painting graffiti on buildings and passing trains to
remind new and old residents of the culture that was supplanted.
If anti-homeless laws are a function of a citys changing culture (associated with
neoliberal growth and revanchist policies), they would also share a relationship with
gentrification. If gentrification is associated with anti-homeless legislationas is the case
with Wyly and Hammers (2005) study, then the results would promote further
investigation in order to uncover whether the shift in cultural values, or the economic
pursuit of neoliberal growth opportunities was really the most responsible for the surge in
anti-homeless laws. Though I recognize how intertwined these variables are in reality,
this study seeks to investigate the factors of neoliberal growth separate from the cultural
and demographic changes that may be associated with such growth, in order to form a
better understanding of the factors that may be most responsible for growing anti-
homeless laws.
Wyly and Hammel (2005) agree with Mitchells (1997) assessment of
gentrification as a key aspect in explaining anti-homeless laws. Wyly and Hammel
(2005) found that gentrification is generally correlated with one strandexplicit anti-
homeless lawsbut most of the variation among cities comes from the broader urban
context in with reinvestment and revanchism have emergedp.11-12). Some of the
39


variables playing out in a broader urban context are racial segregationwealth
segregationand inequality (Wyly and Hammel2005).
UnfortunatelyWyly and Hammels (2005) sample size of cities studied was too
small to make larger generalizations concerning culture, economics, and gentrification. In
addition, they measured segregation by observing racial segregation of one race at a time.
There was no overall measure of segregation that might give a general understanding of
how white and non-whites are segregated within a city. In short, no firm generalizations
about segregation could be made from their work because different races segregate
differently and make up different percentages of the populations for different cities. For
example, the Asian population of San Francisco is larger and integrated differently than
the Asian population of Boston or Denver, and thus, an overall segregation of minority
groups proved hard to quantify in a standardized way from city to city.
Revanchism is an important term to discuss in greater detail because it is both
connected to the culture of the city as well as to a broader discourse on anti-homeless
laws. Smith (1996) and Mitchell (1997, 2011) understand economic factors as primary
drivers of revanchist policies. Revanchism is a politic of revenge on the lower classes by
the upper classes. The upper class is seeking to take back a city stolen from them and
seize economic potential. The lower classes include the homeless, but also include other
populations that dont fit into a traditional picture of a revanchists vision of a healthy
city. Smith (1996) writes,
More than anything the revanchist city expresses a race/class/gender terror felt by
middle- and ruling-class whites who are suddenly stuck in place by a ravaged
property market, the threat and reality of unemployment, the decimation of social
services, and the emergence of minority and immigrant groups, as well as women,
as powerful urban actors. It portends a vicious reaction against minorities, the
40


working class, homeless people, the unemployed, women, gays and lesbians,
immigrants (p. 207).
The cultural undertones and non-economic factors that Smith (1996) touches upon
are commonly considered as a euphemism for class-motivated warfare on poverty
deviancePapayanis2000: as cited in Van Eijk2010). In other wordsnon-economic
theories explaining the emergence of revanchist policies, such as anti-homeless laws, are
sometimes dismissed as secondary arguments, to the real motivating factor of economic
class warfare. However, Pmijt (2012), Johnsen and Fitzpatrick (2010), and Van Eijk
(2010) refuse to dismiss cultural factors so quickly.
Pmijt (2012) evaluated revanchist theory by following the Dutch anti-squatting
movement. While Pmijt (2012) ultimately affirms the thesis of revanchism, Pmijt lends a
voice to the cultural dynamics of revanchism. Pmijt (2012) indicates that policy makers
use a discourse of exaggerated moral panic that draws out a discussion of culture
war. The culture war dynamic highlights the fact that more variables than economic
variables may be in play for any given city. Pmijt (2012) writes, 'The case of anti-
squatting legislation, as in Van Eijk5s (2010) analysis of urban policy in Rotterdam,
shows that revanchism can have a strong cultural component and that it is not exclusively
bound to the economic logic of gentrification and competition between cities as
suggested by Smith (1996,1998)p.1118).
The culture war complements Ferrells (2001) observations and is in turn
supported by Richard Sennett5s (1970) seminal work, The Uses of Disorder. Sennett
(1970) hypothesized that urban culture is turning away from a respect for the cultural
creativity born in urban public spaces of disorder for a pursuit of unattainable purity or
41


structured space fueled by a juvenile fear of disorderchangeand otherness; the city
itself. Smith (1970) writes,
By defining the outsider and the otherness the we become solidified with
other wes to separate themselves from the reality of pain and the toughness of
life... it is bred out of the way that human beings learn at a certain point in their
growth how to lie to themselves, in order to avoid the pain of perceiving the
unexpectedthe newthe otherness around them (p. 39).
Sennetts assessment of the origin of avoidancea cultural fear of experiencing
life of the other through contact in public spacesuggests that movements to restrict the
public space and protect residence from disorder is a cultural desire. This desire is
reinforced by actual disorder, irrational fear, and the very policies that seek to quell
disorder and inevitably separate the others from mainstream society.
Johnsen and Fitzpatrick (2010) depart even further from the economic
explanations of revanchism. While conceding that the state was revanchist in tone and
action, and that such language was typically connected closely to economic arguments,
the public discourse was more complicated. There were those who spoke with fear and
revenge, but there were also those who supported exclusionary policies in Great Britain
from a discourse of compassion. In other words, supporters of the regulation of space
truthfully believed that the laws were in the best interest of the homeless or the deviant.
This finding is echoed by the discourse present during the aforementioned Denver
City Council debates over anti-homeless laws in 2011 and 2012. The discourse of fear
threat and compassion were utilized in promoting the anti-homeless camping ban, at that
time, one of the harshest such laws in the United States (Whelley, 2013). Johnsen and
Fitzpatrick (2010) similarly find that this kind of cultural political discourse was a vital
factor in passing legislation in England.
42


Johnsen and Fitzpatrick5 s (2010) political discourse is understood as an extremely
hard to quantify variable, subject to the shifting moods and conceptions of the public.
Johnsen and titzpatrick (2010) used interviews to qualitatively measure public discourse.
For this study, interviews were not an option, but only the aggregation of demographic
variables that may or may not show change in the makeup of a city.
As stated above, the cultural shift hypothesis proposes that affluent professionals
and whites are culturally afraid of or culturally adverse to homeless people. According to
this theory, if a city5s population comprises a higher proportion of white citizens, has
higher incomes, higher education attainment, and is more racially segregated, that city
should have a higher number of anti-homeless lawsas a reflection of the cultural
conflicts often associated with those variables.
Within this theoretical discussion, there exists an unexamined implication that
largely white affluent professional communities may be culturally afraid of the homeless.
In the threat hypothesis I discussed earlier, I proposed that an actual growth in homeless
rates and crime rates might be responsible for the growth in anti-homeless laws. What the
threat theory doesnt account forhoweveris the possibility that rising cultural threats
emerge from new city residents moving into a city (affluent, white professionals), and a
growing cultural panic, and not growing homeless populations or crime rates, is what
really drives anti-homeless laws.
Similar to the unexamined cultural variables and adding to Johnsen and
Fitzpatrick5 s (2010) cultural political discourse perspective, political variables are largely
unexamined in anti-homeless literature. Mitchell (2011) and Smiths (1996) assessment
concerning the relationship between political parties and revanchist policies over history
43


indicates that anti-homeless laws reemerged out of a conservative-led neoliberal
movement during Reagans Presidency and wasat that timedriven by a specific
political party. However Mitchell (1997) indicates the while anti-homeless laws originate
from conservative revanchist conceptionshe writesPerhaps the most stringent of the
newest anti-homeless laws are in stereotypically liberal5 cities of the West Coast
(p.307).
The different arguments of Mitchell (1997) and Smith (1996) regarding the
relationship of anti-homeless laws to political party control deserve more study. To the
extent that changing political party fortunes are connected to anti-homeless laws, I can
conclude that the cultural shift hypothesis might have significance because a shift in
political party support is an additional way to gage changing culture by measuring
changing political ideology.
The cultural shift theory hypothesizes that a demographic of non-minority affluent
professionals are culturally afraid of or culturally adverse to homeless people. The shift
toward a new wealthy demographic in cities is hypothesized to be partly responsible for
the rise in homeless laws, and more so than a rise in homelessness, per se (Florida, 2012).
A new demographic arriving in the city is hypothesized to be fearful or adverse to
homelessness, as its existence is counter to the professional goals and tastes of the urban
yuppie class. The cultural shift hypothesis is that if a city5s population has a higher
proportion of white citizens, has higher incomes, higher education attainment and is more
racially segregated, it should have a greater number of anti-homeless laws.
44


CHAPTER III
METHODOLOGY
All prior studies regarding reasons for the growth in anti-homeless laws, with the
exception of Wyly and Hammel (2003), have utilized qualitative methods. Wyly and
Hammel (2003) quantitatively studied a small sample of cities with regard to one test,
gentrification. Verifying qualitative research through quantitative research is important
for the progression of knowledge concerning the criminalization of homelessness.
Quantitative methods are chosen for this study because there is a void of quantitative
assessment in anti-homeless literature and little is known about the other issues that may
or may not correlate with the prevalence of anti-homeless laws.
This study utilizes a multiple regression analysis tool to complete its quantitative
analysis. StatPlus software, an extension of Microsoft Excel, was used for all multi-
regression tests. An Excel spreadsheet was created to house all independent variable and
dependent variable data. When an imperfection in the database was found, it was fixed
and prompted a random verification of other independent variable data.
The study sample was confined to those cities in the United States (n=102) where
data are available for all dependent and independent variables. All cities in the sample
have implemented anti-homeless laws; however, the specific laws vary across cities in
severity and purpose. The severity and intricacies of an ordinance, and the specific
process a municipality went through in order to implement such laws, waere not the focus
of this study.
The National Law Center on Homelessness and Poverty (NLCHP) produces a
study roughly every two to three years, surveying cities that have implemented anti-
45


homeless laws. TheNLCHP (2011) categorizes anti-homeless laws across 8 categories
and 14 sub-categories, which provides a rough measure of the illegality of the act of
being homeless in any given city. The categories, with their corresponding sub categories,
appear in Table 2. The Other category includes a large array of ordinances that
criminalize anything from sleeping in cars, to washing windshields, to banning the use of
ones own property in public (NLCHP2011).
Table 2: Anti-homeless Law Categorization
Categories Subcategories
Sanitation Bathing in particular places
Urination/ Defecation in public
Begging Begging in public city-wide
Begging in particular places
Aggressivepanhandling
Sleeping Sleeping in public city-wide
Sleeping in particular public places
Camping Camping in public city-wide
Camping in particular public places
Sitting/ Lying Sitting or lying in particular public places
Loitering: Loitering/loafing in particular public places
Loitering/loafing/vagrancy city-wide
Vagrancy Obstruction of sidewalks /public places
Closure of particular public places
Other: Other
In order to quantitatively study the reasons for the prevalence of anti-homeless
laws, each city in the sample needed to be assigned a single number representing the
severity of anti-homeless laws. The total number of anti-homeless laws recorded by
NLCHP (2011) for any given city was used as a measure of the prevalence of anti-
homeless laws in that city. The total number of anti- homeless laws for each city was
calculated by adding all laws recorded in the NLCHP (2011) report on the criminalization
46


of homelessness. The total number of laws per city served as the dependent variable (DV)
for the multi-regression analysis.
The sample size of this study is 102 cities in the United States. The NLCHP
(2011) report on anti-homeless laws had a sample size of 248 cities from the United
States and Puerto Rico. Acknowledging that the political, economic and cultural
environments may be different in Puerto Rico than in the rest of the United States, the
Puerto Rican cities were omitted.
In addition, data was gathered for homelessness and crime rates, and various city
level census data points for 102 cities across 39 states and the District of Columbia. In
short, the sample is confined to those cities where adequate data were available, making
the sample non-random. However, because the sample size is large and the geographic
representation of different states and regions are accounted for, this sample is generally
representative of cities in the United States. Further discussion of representativeness of
the sample will take place in the results section.
The independent variables in this study were chosen based on the three
hypotheses being tested. Data were collected for 144 independent variables in a database
accounting for the different measures of the three chosen hypotheses. In total, fifteen
independent variables were selected and used in the regression model to test the three
hypotheses. See Appendix C for a list of 144 independent variables separated by
hypothesis category. The next three sections contain reasoning concerning the choice of
independent variables to test each hypothesis.
47


Threat of Homelessness and Crime Test
HI: Rising crime and homeless rates will share positive relationships with high
levels of anti-homeless laws across U.S. cities. Testing this hypothesis required that
homeless and crime statistics be tabulated. Following the method of Boushey & Luedtke
(2011), the total rate of crimes and homeless people in an area, and the rate of increase or
decrease in those rates, were plotted against the implementation of anti-homeless laws
during the time period where data is available. Crime statistics originate from the U.S.
Department of Justice Uniform Crime Reporting Statistics Database, last updated in
2011, with yearly reports from municipalities starting in 1985 (DOJ, 2013). Some large
cities, such as Chicago, Illinois, did not report crime statistics to the FBI for the year
2000 and thus were omitted from the sample.
Homeless statistics originate from the U.S. Department of Housing and Urban
Developments (HUD) yearly point-in-time homeless estimates. These estimates are
based on counts reported by Continuum of Care organizations within states. Homeless
data are made available in two forms, state level and Continuum of Care (CoC) level.
CoC organizations are made up of one or more cities and counties that have joined
together to address homelessness. This study uses CoC level data because they are
superior to state data as a measure or homelessness in U.S. cities. Some states, such as
Rhode Island, Nebraska, and Montana, and cities such as Washington (GA), and Tempi
(AZ), do not have CoC data or were not a primary city within the CoC program, and thus
were omitted from the sample.
48


CoC areas vary in size depending on the city and state which they exist. For
example, Los Angeles City and County has its own Continuum of Care program, while
Fort Worth, Texas includes Arlington and Tarrant counties in their CoC.
Some CoC organizations geographic areas changed drastically during the past 10
years, affecting the validity of the homeless counts, and these were omitted from the
study. For exampleSan Diegos CoC grew from encompassing only the city in 2005 to
include the entire county in 2012. Because the size of San Diego County is many times
larger than the city of San Diego, the San Diego CoC count of homeless persons is
inconsistent across the relevant time frame, and therefore San Diego was omitted from
the study.
Rates of homelessness and crime were calculated using the data that were
provided by two government databases. Both government databases indicate limitations
on their ability to verify data because the city, or CoC, submitted data to the national
database without independent verification. However, there are no other data sets as
comprehensive as these two databases (HUD and Department of Justice) and both are
used frequently in academic research. See Appendix C for complete list of threat test
variables.
Neoliberal Growth Test
H2: Economic factors associated with neoliberal growth patterns will share a
strong positive relationship with the growth of anti-homeless laws across U.S. cities.
Contemporary urban theory argues that economic and political neoliberalism is deeply
connected with anti-homeless politics. Wright (1997), Mitchell (2003) and Harvey (2012)
argue that anti-homeless laws are a result of neoliberal policy. Harvey (2012) and
49


Mitchell (2003) insist neoliberal policy resulted in the expansion of private ownership
and explosion of re-developed public spaces, which led to efforts to remove the homeless
from those spaces. Peck (2002) argues that cities are all more or less caught up in this
neoliberal era, and that the politics and economics of neoliberalism are fused, so that anti-
homeless laws are spreading across the country. The independent variables that will be
focused on to test this theory will measure growth of a city in terms of population,
wealthjobsnumbers of people in the creative class2 the high tech sectorproperty
values, rental vacancy rates, income to rent ratios, and median cost of rent, among others.
See Appendix C for complete list of neoliberal test variables.
Cultural Shift Test
H3: Demographic factors indicating political and cultural shifts will hold positive
associations with anti-homeless laws. The goal of this test is to quantify changes in
culture. Changes in income, changes of education attainment, changes in wealth and
racial segregation patterns, population growth, minority presence, political party strength,
and measures of creative class will be used to interpret possible changes in culture and
identify if particular cultural demographic variables associate with anti-homeless laws.
Many of the same measures used to measure gentrification will be used to measure
cultural changes. Meaning, many of the same measures used to investigate neoliberal
growth may also indicate cultural shifts within the city. Measures of racial and political
shifts and racial concentrations will be used to give context to those measures that may
support more than one theory. While each measure is only tested once in the multiple
2 The Creative Class refers highly educated, highly mobile, typically young
professionals working in creative or knowledge based industries (Florida, 2012).
50


regression model, some measures will be relevant in the discussion of more than one
theory.
Racial demographics statistics, dissimilarity statistics, population growth, gini
inequality indexes, measures of racial diversity, median income, and poverty data were
used in this study. Segregation Dissimilarity indexes are a city block by city block
measure of diversity that is aggregated to measure the total integration of two races in
any given city. See Appendix C tor a complete list of independent variables used to test
the cultural shift hypothesis.
The political question for this study is whether the majority political party within
a city has any association with the prevalence of anti-homeless lawspresumably the
conservative Republican Party would be more likely than Democrats to support anti-
homeless laws. Presidential voting statistics from each county were used in the model to
test this possibility. Voting data from 2000 and 2012 were used to calculate the average
voting gaps (in terms of the percent of the local population voting for one party versus the
other) and percent change in voting gap was relied on as proxy for this test. Voting gaps
are calculated by subtracting one partys proportion of the vote from the other. In this
case, I chose to subtract percent Republican vote totals from percent Democratic vote
totals for presidential elections. The difference equals the percent voting gap or the
proportional voting difference between Democrats and Republican. See Appendix C for a
complete listing of independent variables to test the influence of local political party
power. Unfortunately, only county level data were available for the measure of political
party strength. Again, this study used the best available data. In the next section I will
describe the results of the multi-regression analysis.
51


CHAPTER V
RESULTS
The results section is separated into five subsections. The first examines the
descriptive statistics of the study and the results from a simple Pearsons Correlation test.
The Pearsons Correlation tests for existing linear relationships between two sets of data.
The second examines the overall strength and statistical significance of the multiple
regression model. The remaining three sections are devoted to presenting the regression
results of the three theoretical tests: threat, neoliberal growth, and culture shift.
Descriptive Statistics
The sample mean of anti-homeless laws across 102 cities is 10.8 anti-homeless
laws and the standard deviation is 2.68 laws. In the larger sample of 189 US cities the
mean number of laws per city was slightly lower (10.1 laws per city) with a higher
standard deviation (2.99). Similar mean and standard deviations between samples
strengthens generalizability. The highest number of anti-homeless laws in a city was
sixteen laws (Colorado Springs, CO; Los Angeles, CA; Ashville, NC) and a low of two
laws (Fall River, MA). The median of the sample is 11 while the mode is also 11.
The geographic dispersion of cities in the sample has cities of 38 states and
Washington DC. Seventeen of the cities in the sample are in the Northeast, (ME, NH,
VT, NY, NJ, PA, DE, RI) seventeen cities are in the Midwest (OH, IN, IL, MO, KA, NE,
IA, SD, ND, MI, WI, MN), twenty-three cities are in the West (CO, WY, MT, UT, NV,
AZ, NM, CA, OR, WA, ID), and 45 cities are in the South (TX, OK, LO, AL, AK, MS,
FL, TN, KT, WV, MD, VA, NC, SC). Twenty-seven of the southern cities are in the mid-
and south Atlantic regions, while eighteen cities are from the gulf region as well as inland
52


southern states. The sample mean for Northeast cites was the lowest (8.47 laws), while
the Midwest region had the highest (12.06 laws); closely followed by the West (11.78
laws), and the South (10.71). The minimal variation in the dependent variable
geographically3 and statistically may explain why detecting meaningful patterns is
difficult.
Cities in the sample varied greatly in terms of population. Naples, FL (20,091),
was the smallest citywhile New York was the largest (8,128,980) (ACS 2011).Only 18
cities in the sample had populations under one hundred thousand, and the mean
population was 477,555. Twenty-six cities had populations over five hundred thousand,
and seven cities had populations over one million. The majority of the sample (fifty-eight
cities or 56.86%) had populations above one hundred thousand and smaller than five
hundred thousand people. No relationship was found between population sizes or the
changes in population and number of anti-homeless laws.
3 In an attempt to identify geographic trend in anti-homeless laws I included a
measure of temperature. Intuitively, if locations are cold and homeless people are
outside during the harsh weather, a theory of agency, a theory that supports anti-
homeless laws, would be undermined by the fact that surviving in the cold is not
something one would likely choose to do. The average temperature for all cities in the
sample was 58.36 degrees Fahrenheit with an average January temperature of 38.93
degrees Fahrenheit. Temperature was an important variable to consider in order to rule it
out as a possible confounding variable for the rest of the study. In the end, colder cities
varied in the same way that warmer cities varied in terms of the number of anti-homeless
laws. For example, in 2011 Corpus Christi had 8 anti-homeless laws (below average)
with an average yearly temperature of 72 degrees Fahrenheit and an average January
temperature of 5b degrees Fahrenheit. Jacksonville FL, with comparable temperatures
had 15 anti-homeless laws in 2011.With respect to colder cities such as Detroit,
Cleveland, Buffalo, and Portland (OR), all had above average anti-homeless law rates but
lower than average temperatures. Conversely, cities like Fall-River, New York and
Boston had lower than average temperatures and lower than average anti-homeless law
rates. No relationship was found between the number of anti-homeless laws and,
temperature and thus we can dismiss temperature as a possible confounding variable.
53


An examination of descriptive statistics was necessary to understand what types
of cities are in the sample and to assess generalizability. Generalizability is challenged by
the fact that this model is not comprised of a random selection of cities and cities cannot
be placed in controlled settings. However, showing that a large sample or cities (102)
come from different regions of the country and vary in size, location, and diversity,
strengthens the potential generalizability of this model.
The Pearsons Correlation findings appear in Table 3. Note that a P-test less than
.05 indicates statistical significance. The P-Tests measures the probability that the
correlation strength is not the result of a random dispersion of values. A .05 significance
indicates that we are 95% certain that the values are not random, and thus, there a
relationship can be trusted. Three statistically significant correlations exist within the data
that were chosen and used in the multiple regression model.
Table 3: Pearson^ Correlation
Variable Correlation Coefficient P-Test
Average Homeless Population .12727 0.20242
Percent Change in Homelessness -.22976* 0.02018
Percent Change Violent Crime Rate -.23655* 0.01668
Percent Change Property Crime Rare .09108 0.3626
Average Property Crime Rate .11766 0.23889
Average Violent Crime Rate .09108 0.35263
Percent Political Gap Shift .42924*** 0.00001
Average Voting Gap .06919 0.48957
Total Average B+H+AAV Dissimilarity Index -.11537 0.2482
Percent white only -19022 0.05549
Percent Change Median Owner Occupied Value -17543 0.0778
Percent Change in Housing Unit Growth .11703 0.24143
Growth of High Tech Location Quotient .11557 0.24741
Percent Change Median Income -.08242 0.4159
Growth of Aggregate Property Value .02298 0.73457
*p<+05;**p<.01;***p<+001.
54


The percent change in voting gap is by far the most significant (.00001)and most
highly correlated (.42924) of all the independent variables. The percent change in voting
gap is almost twice as correlated than the only other two statistically significant
correlations in the Pearsons Correlation model. The positive nature of this correlation
indicates that the change being measured associated with anti-homeless laws is one
toward Democratic voting patterns. This result indicates that there is a moderate to strong
relationship between cities that are increasingly voting Democratic and a growth in anti-
homeless lawsbut without consideration of any other possible confounding variables.
The percent change in violent crime rates shows the second most significant
relationship (.01b/) and second strongest correlation (-.237) in this model. The negative
nature of the correlation coefficient (-.237) indicates that falling violent crime rates are
correlated with higher homeless law totals. With that said, it is important to remember
that this correlation is half as strong as the Political gap change, and thus the correlation
strength is comparatively small.
The percent change in homelessness populations is the third most significant
(.0202) and third strongest (-.23) correlation in the model. The negative nature of the
correlation coefficient (-.23) indicates that declining homeless rates are related to higher
anti-homeless law rates. However, as in the case for violent crime rate change, the
correlation coefficient for the percent change in homeless populations is comparatively
small. It is also important to note that the percent change in homeless populations and
change in violent crime rates are not significant (P-test = .4249) when correlated with
each other.
55


In the regression model correlations between independent variables are important
for interpretation because it helps give context to covariance and whether the two
independent variables may be effecting each other or affecting the dependent variable in
a similar way. In this case, it is likely that the percent change in violent crime and percent
change in homeless populations are not affected by each other, as they mutually affect the
number of anti-homeless laws in a city, as will be discussed later. (See Appendix E for
complete correlation matrix).
One possible explanation for these results concerning crime and homelessness is
that anti-homeless laws cause a drop in homelessness and at the same time reduce violent
crime rates (as Broken Windows policing theory would suggest). However, correlations
do not prove causality. In addition, asserting that anti-homeless laws cause violent crime
rates and homeless rates to drop is complicated by the fact that homeless rates and violent
crime rates are not significantly correlated with each other. Further study is needed to
uncover whether falling crime rates and/or homeless rates are a result of growing anti-
homeless laws.
Two other results are important to note, as they narrowly missed the .05 P-test
limit. These variables are percent white only and percent change in median
owner/occupied property value. The lack of significance does not necessarily indicate
that a correlation does not exist, but the results simply indicate that there is slightly more
than a 5% chance that the observed correlation is coincidental. In addition, the Pearson
Correlation only measures one relationship at a time, and thus, the results do not take into
consideration that anti-homeless laws may be influenced by a number of factors. The
most important findings of the Pearsons Correlation is that a Democratic shift in politics
56


is associated with higher rates of homeless laws, while two smaller secondary
correlations exist between falling homeless rates and violent crime rates and high levels
of anti-homeless laws.
These results suggest possible relationships between a variety of factors
hypothesized to hold relationships with growing anti-homeless laws. However, the
simple Pearsons Correlation test cannot examine the effect of one variable while holding
other variables constant, nor can it determine the direction of possible causation between
variables (that isa Pearsons Correlation can tell us that anti-homeless laws are
correlated with low rates of homelessness, but it cannot tell us whether the rate of
homelessness drove the passage of the law, or vice versa). To address these limitations,
the multiple regression model used in this study identifies a chosen dependent variable
(number of anti-homeless laws) and allows us to test whether specific independent
variables affect the dependent variable while holding other independent variables in the
model constant. In the next subsection I will discuss the overall regression model results.
Regression Model Results
The regression model results involve an evaluation of the regression model as a
whole. Table 4 contains the results of the overall multi-regression model. The F-Test
(5.12) indicates that the results of this model are not coincidental and thus the results of
this study may be acceptable as long as the model results are statistically significant. The
P-test (4.17E-7) indicates that this model is statistically significant. With a F-Test above
5 and a P-test well below .05, the presence or absence of individual relationships between
the dependent variable and the independent variables are to be trusted according to their
respective P-test values.
57


Table 4: Overall Regression Model Results
Model Results: Regression Statistics _____Value
R Square
Adjusted R Square
F-Test
P-Test
0.47163
0.37947
5.11765
0.0000004167171
The Adjusted R Square Value (.38) indicates the proportion of variance explained
by this model (38%). The .09 difference between the Adjusted R squared and the R
Squared is likely due to the fact that five of the fifteen independent variables do not have
significant relationships.
The Adjusted R Square (.379) expresses what can be learned from this model in
positive terms. Put into negative terms, 62% of the possible variance in all cities within
this model cannot be accounted for. However, the results do explain a significant portion
(37.9%) of total variance, making the results of this model relevant to further scientific
knowledge. Next, I will present the regression findings for all fifteen independent
variables used in the regression model.
Independent Variable Regression Results
Of the fifteen independent variables used (out of 144 tested), ten had significant
relationships with the dependent variable. Five independent variables remain in the model
though they were without statistical significance, because they help to focus the
inferences made from those independent variables that do have significant relationships.
Table 5 shows the multiple regression results for all independent variables, including
regression coefficients and significance values labeled P-test. Note that unlike the
simple Pearsons Correlationthese unstandardized regression coefficients can be greater
or less than 1 or -1 because they exhibit effect rather than correlation. The regression
coefficient can be used to predict the average change in Y (dependent variable), given the
58


change of one unit of X (independent variables), if all other variables are held constant.
With that said, it is inappropriate to make literal inferences concerning the effect of X on
Y because, in reality, independent variables are not held constant and the majority of the
variance is unexplained by this model. In addition, multiple linear regressions are one-
way tests, meaning that the effect on anti-homeless laws is tested while the effect
produced by anti-homeless laws on the independent variables is not. Any effect of anti-
homeless laws on the independent variables is beyond the scope of this study.
Table 5: Multiple Regression Results
Independent Variable Regression Coefficient P-Test
Theory of Threat Test..
Average Homeless Population (2005-2012) 0.00018** 0.00264
Percent Change in Homeless Population (2005-2012) -0.37769* 0.02282
Average Violent Crime Rate (per 100,0; 2000-2010) -0.00019 0.83014
Percent Change in Violent Crime Rate (per 100,000; 2000-2010) -1.70303* 0.03784
Average Property Crime Rate (per 100,000 2000-2010) 0.00026 0.12519
Percent Change Property Crime rate (per 100,000; 2000-2010) 3.50985* 0.01977
Neoliberal Theory Test
Percent Change in Housing Units (2000-2011) 2.75001 0.20062
Percent Change Median Owner/Occupied Value (2000-11) -1.94764* 0.0181
Growth in Aggregate Property Value (2000-2011) -2.7860E-11* .048
Growth of High Tech Location Quotient (2008-2012) 0.46465** 0.0043
Culture Shift Theory Test.
Average Political Voting Gap (+ = Democrat; 2000-2012) -0.35506 0.76451
Percent Change in Political Voting Gap (2000-2012) 7.3437** 0.00438
Average Black/ white + Average Hispanic/ white + Average -0.02843* 0.02023
Asian/ White Dissimilarity Index (1990-2000-2010)
Percent of Population White only (2011) -4.7586* 0.01149
Percent Change in Median Income (2000-2010) 2.75859 0.33269
*p<+05;**p<.01;***p<+001.
Examining trends in independent variable regression data is important in
preparation for analyzing the overall results of the multiple regression model. The
distribution of independent variable values may greatly alter the interpretation of results.
59


The rest of this section is split into 3 subsections to better organize and analyze
the regression results for various independent variables. Each subsection will contain a
presentation of descriptive statistics for each independent variable being tested.
Unfortunately (but as is common in social science regressions) many of the independent
variables are measurements of different categories across varying scales. For example,
the percent change of homeless population rates are a percentage measure (one unit =
1%), while the average homeless population is an aggregate measure of individuals (one
unit =1 homeless individual). Another example, percent change in political gap is
calculated across a twelve-year period while the percent change of homeless rates are
calculated across a eight-year period.
Relative comparisons between individual coefficients of different scales cannot be
made. Those variables that are scaled the same can be compared relatively. For example,
relative effects of measures scaled using percentages, the most common measure in this
model, can be compared if those measures are calculated along the same time frame. The
non-relative comparison of coefficients of different scales is not impossible. Rather than
expressing the differing effects in terms of proportionality, like Pearson Correlation
results (Change in Political Gap is almost twice as correlated to anti-homeless laws as the
change in violent crime rates), I will compare predicted effects by using descriptive
statisticsPearsons Correlation resultsand qualitative discussion of predicted effects.
Theory of Threat Test Results: The first hypothesis that was tested within the multi-
regression analysis was the threat hypothesis. To review, the threat hypothesis states that
rising crime and homeless rates will share positive relationships with high levels of anti-
homeless laws across U.S. cities. As seen in Table 4 (Appendix B), and again in Table 5,
60


the threat hypothesis tested for the presence of relationships between six different
independent variables, while controlling for other variables in the model. Four out of six
threat variables hold significant relationships with the number of anti-homeless laws:
average homeless population, percent change in homeless population, percent change in
violent crime rates, and percent change in property crime rates. The two variables that did
have significant relationships were average property crime rates and average violent
crime rates. According to this model, therefore, the level of crime in a city holds no
relationship with the number of anti-homeless laws in a city, but the change in crime rates
might.
The average homeless population rate, however, shared a positive, statistically
significant relationship with the dependent variable. At first glance, one might conclude
that this result would indicate that the higher a homeless population is, the higher number
of anti-homeless laws that can be expected. This conclusion would be misleading. The
regression coefficient (.00018) and the large variability in homeless rates, as seen by the
range and standard deviation, indicate that this is a very weak association. The regression
model predicts that for a one-unit increase (one unit = one homeless person), there should
on average be a .00018 increase in anti-homeless laws. This result indicates that, on
average, it would take an increase of 5,555.56 homeless persons (1 / .00018) to produce
one new anti-homeless law, keeping all other variables constant. The variability apparent
in the descriptive (summary) statistics supports this finding, but not wholly. On the one
hand, Los Angeles has the highest average homeless rate (62,850) over the last 8 years
and has the highest level of anti-homeless law rates (16), New York has the second
61


0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Ave Homeless Pop (2005-2012)
Figure 1:Residual plot of Average Homeless Population and Anti-Homeless Laws
A citys percent change in homeless population ratemeasuring the growth or
decline of homelessness, had a statistically significant negative relationship with the
highest homeless average (52,413) with half as many (well below average) anti-homeless
laws.
Simply put, homeless population has a very weak effect on anti-homeless laws
according to this model.A 5,555.56 homeless persons increase to one additional law ratio
only holds true if all other factors are held constant. Converselyaverage homeless
populations have the most statistically significant relationship with the number of anti-
homeless laws in the entire model (.00264) warranting further discussion. On the surface
this seem logical, however this is not the case. Figure lisa residual scatter plot with a
logarithmic trend line showing that as the homeless population increases, the effect of
homeless persons on influencing additional laws diminishes. In short, while there is a
statistically significant relationship between anti-homeless laws and average homeless
populations, it is weak, and the effect declines in those cities with more and more
homeless people.
SMS--sol
62


number of anti-homeless laws in that same city. The strength or effect of the relationship
is weak (.38). This model predicts that as homeless rates decrease by 1% over an eight-
year period, there would be an average corresponding .38 increase in anti-homeless laws.
The large range and standard deviations (Table 6) indicate that the results need to be
analyzed further before developing an understanding of what the results mean. A
negative relationship suggests that as homeless rate fall, the number of anti-homeless
laws rise. We cannot conclude, however, that homeless rates fall as a function of anti-
homeless laws because multiple regressions can only predict for the dependent variable
and not independent variables.
It is counter intuitive to think declining homeless rates influence cities toward
passing anti-homeless legislation, but not out of the realm of possibilities. Pmjit (2012)
indicates that anti-squatting legislation in the Netherlands was passed long after the apex
of squatting behavior. This line of reasoning will be discussed further in the discussion
section.
There is a possibility that reverse causation is being observed. Meaning that anti-
homeless laws modestly decrease homelessness. The answer to that question is beyond
the scope of this study, but the possibility of a reverse causation warrants further
discussion. If homeless rates fall in a significant way as a result of anti-homeless laws,
then the advocates of broken windows policing and proponents of anti-homeless laws
may have some justification when arguing that anti-homeless laws force people off the
streets and prevent people from choosing to be homeless in the first place. If this isin
fact, the relationship that exists, then the next question would be the extent to which there
is an increased demand for service after the implementation of anti-homeless laws. Do
63


those cities who employ such laws follow or precede the implementation of anti-
homeless laws with expanded services and housing? Or, are they isolated policies? Do
homeless people move to more remote locations of urban areas after the implementation
of a law making it more difficult to count them? More discussion concerning a possible
reverse causation will appear in the next sections.
For the purposes of this study, there is reason to suspect that anti-homeless laws
are a lagging indicator of homelessness, while keeping in mind the relationship may in
fact be reverse causation or explained by an unknown tertiary factor to which both
homeless rates and anti-homeless law levels are related. In addition, the regression
relationship is very weak when compared to other variables, as we will discuss later in
this section. Overall, this model indicates that homeless rates and changing homeless
populations have little effect on anti-homeless laws when all other variables are held
constant.
A citys percent change in violent crime rate had a statistically significant
negative relationship with the number of anti-homeless laws. The regression coefficient (-
1.70) indicates a modest relationship, while the modest range and standard deviation
support this conclusion. The mean rate of violent crime growth was -12.72 % while the
standard deviation was 33.83%. This indicates that the vast majority of the sample cities
had negative violent crime rate growth. The scale of percent change in violent crime rate
is calculated across an eleven-year span while homeless change rates are calculated
across an eight-year time frame. With this said, the three year difference in calculating
the rate of change for these two variables will most likely not confound a general
comparison between the two as they relate to anti-homeless law. Violent crime rate has a
64


4.5 times greater effect than the percent change in homeless populations on anti-homeless
law levels. Meaning, violent crime rates have more effect than homelessness rates on the
passage of anti-homeless lawsthough the relationship is an unexpectedly negative one.
The model predicts that a 1% decrease in violent crime rates across an eleven-
year span corresponds with a 1.7 unit increase in anti-homeless laws. While this may also
indicate anti-homeless legislation is a lagging indicator of high violent crime rates at
some point in the recent past, we must be wary. Like percent change in homeless
population, this result may be indicating a reverse causal effect (passing more anti-
homeless laws reduces violent crime rates, as advocates of such laws often argue).
However, if this is tme, than most likely the reverse causal effect associated with
homeless rates and violent crime are independent from one another because the percent
change in homelessness and percent change in violent crime have a Pearsons Correlation
strength of -.00168 with a P-test of .9866. This indicates that percent change in
homelessness and percent change in violent crime are not correlated with each other.
Most likely violent crime rates and homeless rates are independent of one another.
Meaning thataccording to the Pearsons Correlation resultsa rise and fall of homeless
rates is not related to a rise or fall in violent crime rates. This small finding simply
provides context for a possible reverse causation. If a reverse causation is present for one
of the two variables, there would not necessarily exist a reverse causation for the other
because the two independent variables observed seem to be highly unrelated. This finding
is also not presented to contest the result of the regression. It is imperative to interpret the
results of this study with the understanding that at times independent variables that effect
the dependent variable may also be or may not be effecting each other.
65


A citys percent change in property crime rate had a positive statistically
significant relationship with the number of anti-homeless laws. The effect (3.51) is twice
as strong as violent crime rate effect and many times stronger than the percent change in
homeless rates. This model predicts that a 1% growth rate in property crime over an
eleven-year period would result in 3.5 new anti-homeless laws. Unlike violent crime rate
change, property crime rate growth seems to share a relationship with high levels of anti-
homeless laws. However, this simplistic view would be misleading. Only eleven of the
one hundred two cities in the sample had positive property crime rate changes over the
past decade. This may indicate that for those cities where property crime fell the slowest
and those where property crime grew are associated with higher anti-homeless law levels
when all other variables are held constant. In addition, the results may indicate simply
that property crime is much more widespread than violent crime and much more likely to
be experienced and witnessed by the majority of city residents as well as commuters and
visitors. Perhaps property crime drives fear of public spaces and homeless persons more
than violent crime rates especially as violent crime rates have continued to fall since the
80s and 90s.
Percent change in property crime rates also has a positive statistically significant
(>.0001)correlation with the percent change in violent crime rate (.5308). The positive
relationship between anti-homeless laws that appears in the regression results stands
contrasted to the negative relationship between the percent change in violent crime rates
and anti-homeless laws. If a reverse causation was present for property crime, then the
results may indicate that anti-homeless legislation would significantly drive up the rate of
66


property crime over an extended period. While not out of the realm of possibilities, that
would stand in opposition to the Broken Windows thesis.
Still, there is evidence that just such a phenomenon might occur. Johnsen and
Fitzpatrick (2010) interviewed at least one homeless individual who indicated anti-
begging laws increased his property crime behavior. The individual stated, uIt pushed me
to do a little bit of shoplifting, petty shoplifting, which I wasn5t happy about (Johnsen and
Fitzpatrick 2010, p.1710). This single interview is not enough to challenge an established
theory like Broken Windows. However, if the act of being homeless is defined as being a
property crime, as anti-homeless laws do, than we would expect property crime rates to
rise after the implementation of a new anti-homeless law. Yet, the Broken Windows
theory predicts that cracking downon minor offences like homelessness is predicted to
have a significant effect on both violent crime and property crime. The data gathered in
this study do not lend support to such a theory.
Putting thoughts of reverse causality aside, this model indicates that homeless
rates share very small, but statistically significant, relationships to anti-homeless laws
while changes in crime rates have more effect when all other variables are held constant.
Falling violent crime rates and homelessness rates and rising property crime rates predict
growth in anti-homeless legislation. The effect of the absolute average homeless
population is extremely small and the logarithmic trend line of the residual plot most
likely indicates that after a certain size is reached in terms of homeless residents, the
effect decreases and becomes even weaker. Issues related to possible reverse causality
will be discussed in part in the discussion section, but will overall require further research
to satisfy questions that have arisen during this section.
67


Table 6: Threat Results Summary Statistics
Independent Variables Mean Median & (Range) Standard Deviation
Average Homeless Population (2005-2012) 3,511.64 1500.75 (194 to 62,850) 8095.81
Percent Change in Homeless Population (2005-2012) +10.54% -5.43% (-85% to+1138%) 137.16%
Average Violent Crime Rate (per 100,000; 2000-2010) 888.88 808.78 (198 to 2231) 423.84
Percent Change in Violent Crime Rate (per 100,000; 2000-2010) -12.72% -18.54% (-69% to 149%) 33.83%
Average Property Crime Rate (per 100,000 2000-2010) 5421.36 5548.28 (1352 to 11763) 1680.62
Percent Change Property Crime rate (per 100,000; 2000-2010) -22.53% -25.79% (-66% to 47%) 18.2%
Neoliberal Growth Test Results: The second hypothesis tested was the neoliberal growth
hypothesis. This hypothesis states that economic factors associated with neoliberal
growth patterns will share a strong positive relationship with high levels of anti-homeless
laws across U.S. cities. Table 4 contains regression results for this section while Table 7
contains descriptive statistics for this subsection. Three of the four independent variables
representing this theory proved to have statistically significant relationships with the
number of anti-homeless laws, however not always in the way that was anticipated:
percent change in median owner/occupied property value (2000-2011), growth of high
tech sector location quotient (HT LQ) (2008-2012), and aggregate growth of property
value (2000-2011).
The independent variable that did not share an association was the percent change
in housing units. Housing unit growth was one measure of population growth and
development. Percent growth in housing units was hypothesized to have a significant
positive effect on the number of anti-homeless laws, but no relationship was found.
68


A citys percent change in the median owner-occupied housing value had a
statistically significant negative relationship with the number of anti-homeless laws. The
regression coefficient (-1.95) suggests that there is a strong relationship and the small
standard deviation and range relative to the mean indicate that outliers do not likely skew
the strength of this relationship. A 1% decline in median owner-occupied housing value
over an eleven year period is expected to have more effect on anti-homeless laws that a
1% decrease in violent crime rates, but less effect than a 1% increase in property crime
rates. Furthermore, the relationship is oddly negative: rising property values result in
fewer anti-homeless laws.
Before we say with confidence that, all things being equal,a 1% decrease in
housing values is expected to produce two anti-homeless laws, we must examine the
descriptive statistics. One hundred percent of the sample had positive growth in terms of
median owner-occupied housing values. The effect that is observed is one that concerns
slower growth, as compared to other cities. As median growth is slower than other cities,
even before reaching the negative, anti-homeless laws are expected to result, if all things
are equal. Therefore, this model may indicate that in those cities where housing prices
rose the slowest over the past decade also tend to have the highest number of anti-
homeless laws.
A citys aggregate property value also shared a negative statistically significant
relationship with the number of anti-homeless laws; however, the strength of the
relationship was much weaker and the significance is much less than percent change in
median owner-occupied housing value. The p-test value for significance was .048, only
two-hundredths below the cutoff of .05. The regression coefficient (-0.00000000002786)
69


at first glance, indicates that the strength of the association is extremely weak. When put
into predicted terms, this model predicts that every $35.9 billion dollars gained in
aggregate value in a city will result in the decrease of one homeless law. With a mean
growth rate of $129.7 billion dollars over eleven-years and a median growth rate of $49.5
billion dollars, the $35 billion dollar effect remains relatively small because the
relationship is negative. All things held constant, one additional anti-homeless homeless
law would result from a -$35 billion dollar loss, a loss that did not occur in the sample.
With that said, if the effect being observed is simply one of slower growth instigating
anti-homeless law implementation, then a $35 billion dollars slower value growth over an
eleven year period is not a outrageous figure.
The predicted decrease in anti-homeless laws associated with rapidly growing
property values may be misleading. Similar to median owner-occupied housing value, the
aggregate value data show that only two cities (Dayton, OH, and Detroit, MI) had
negative growth rates in aggregate terms of the last eleven years. Therefore, like the
owner-occupied results, this model indicates that locations that grew slowest in terms of
aggregate property value, tended to have the highest number of laws, but the effect of
aggregate value has less effect and may be less important than the owner-occupied
results. However, it supports the median owner-occupied hosing value results.
One interesting relationship that was uncovered during the Pearsons Correlation
was the very strong (.879) and statistically significant relationship (.0000) between
aggregate growth and average homeless populations. In addition, median owner-
occupied-housing value held a .255 correlation and a .00983 significance level with
average homeless populations. This suggests that those cities, possibly larger cities,
70


which grew the most (in aggregate value terms), tend to have higher homeless
populations. This conclusion is intuitive, but will require further study to uncover other
variables that predict homelessness.
The growth of a citys high tech sectors location quotient (LQ HT) relative to the
national average had a positive statistically significant relationship with the number of
anti-homeless laws. The location quotient compares concentrations of high-tech sectors
on a national scale. Those locations with a LQ HT equal to one, have a high tech sector
(e.g. microchip company) that is more heavily concentrated that the national average.
While the regression coefficient was modest (0.46465), the P-test indicates that this test
was one of the most statistically significant relationships (0.0043). This model predicts,
holding all other variables constant, that a growth of 1 LQ HT over a 4 year period would
result in one half of one anti-homeless laws.
Table 7: Neoliberal Results Summary Statistics
Independent Variable Mean Median & (Range) Standard Deviation
Percent Change in Housing Units (2000- 2011) + 11.67% + 8.80% (-17% to 61%) + 11.78%
Percent Change in Median Owner/Occupied Housing Value (2000-2011) +79.61% + 77.41% (12% to 188%) + 35.68%
Growth in Aggregate Property Value (2000- 2011) $12968544583 $4955077500 (-921860000 to 312229155000) $34381319 293.95
Growth of Location Quotient of High Tech Sector relative to national mean. (2008-2012) -.147 0 (-4 to + 3) 1.438
The short time period may or may not be minimizing the effect of long-term high
tech growth. More research is needed to answer this question. What is important to note
is that hi tech growth measures do deserve a place in anti-homeless discussion; however,
for this model they have had much less impact than previously expected. Thus, the results
71


indicate that LQ HT shares a moderate relationship with anti-homeless laws. These
results suggest that higher anti-homeless law rates were found in metro areas that also
enjoyed the largest growth of LQ HT relative to all other metro regions in the nation.
Cultural Shift Test Results: The final test of this study was the culture shift hypothesis.
As stated before, this hypothesis predicts that demographic factors indicating political
and cultural shifts will hold positive associations with anti-homeless laws. Table 8
contains descriptive statistics relevant to this subsection, while Table 4 contains a
summary of regression statistics for this section. Three of the five independent variables
studied share statistically significant relationships with the number of anti-homeless laws:
percent change in political voting gap (2000-2012); percent of populations that is White
only (2011); and total combined average of AsianAVhite, Hispanic/ White, and
BlackAVhite dissimilarity indexes (1990-2010).
The two independent variables that did not prove to have significant relationships
were average political voting gap and percent change in median income. Median income
could have been used as a neoliberal test; however, to test the cultural shift hypothesis, I
needed look at what type of demographic change was happening, and this hypothesis
predicted that an influx of wealthy professionals would be a predictor of anti-homeless
laws. Regardless, if this variable is best conceived of as a neoliberal test or cultural test, it
was not significant.
Average political gap, measuring the dominance of a political party in a county,
also had no significant relationship. According to this modela particular political partys
dominance is not important. However, percent change in political party voting strength
presents a very different story.
72


Similar to the Pearsons Correlation resultsthe multiple regression model finds
the percent change in political voting gap to be the most important variable. A citys
percent change in the political voting gap shares a positive statistically significant
relationship to the number of anti-homeless laws, when positive means Democrats enjoy
a majority in the city and negative means Republicans enjoy the majority.
I will reiterate that there was no relationship found between a particular political
party5 s strength and anti-homeless laws. However, the percent change in political gap
result suggests that those cities that became more Democratic since 2000 have the highest
number of anti-homeless laws. The regression coefficient (7.3437) indicates that a 1%
increase in Democratic presidential voting patterns results in 7.34 anti-homeless laws
over a twelve-year period if all other variables are held constant. This finding may
indicate that, in general, Democratic political growth expanded throughout urban areas
while anti-homeless laws mirrored this growth. However, the strength of relationship
indicates that these results are not coincidental.
Descriptive statistics support the existence of a strong relationship because there
is a small degree of variance in the sample data as evident by the small standard deviation
(9.85%) and modest range in values. The P-test value indicates that this relationship is
one of the most statistically significant in the model (0.00438). These results indicate
that anti-homeless laws are highest in those cities that are increasingly becoming
Democratic.
A citys combined average dissimilarity index measurea measure of the degree
of racial segregation in a cityshares a negative, statistically significant relationship with
the number of anti-homeless laws. The regression coefficient is small (-0.02) suggesting
73


that there is a weak relationship with the dependent variable. Dissimilarity statistics are
measured on a vastly different scale. With a maximum score of 300 (if Asian, Black and
Hispanic populations were 100% segregated from Whites) and a minimum score of zero
(if Whites were completely integrated with non-White minorities), the .02 effects
translates into 35.2 dissimilarity units. Putting this into positive terms, every 35.2
dissimilarity index units toward integration is predicted to result in one additional anti-
homeless law.
In reality, cities are not perfectly segregated or integrated. The median and mean
values of dissimilarity index are 134.9 and 133.9 respectively. For the median city, a 35.2
reduction in dissimilarity would mean a 26.9 % change. This result suggests that those
cities where segregation was the lowest had the highest rates of anti-homeless laws,
however, this weak association will need more supporting evidence to make significant
inferences from the result.
A citys proportion of white-only population has a negative statistically
significant relationship with the number of anti-homeless laws in a city. The correlation
coefficient (-4.7586) indicates a strong relationship. Unfortunately there is a large amount
of variance within these data. Fortunately, the P-test (0.011)and the relatively small
standard deviation (17.09%) suggest that it is very unlikely that outliers are responsible
for the strength of the regression coefficient. This model predicts that a city with a 1%
proportional decline in the white population would correspond with a 4.75 additional
anti-homeless laws if all other variables were held constant. Second only to percent
change in voting gap percent White only may be the best predictor in this model. The
74


results indicate that cities where white-only populations were lower tended to have higher
anti-homeless laws.
Table 8: Culture Shift Results Summary Statistics
Independent Variable Mean Median & (Range) Standard Deviation
Average Political Voting Gap (+ = Democrat; 2000-2012) + 11.31% + 9.98% (-43% to 80%) 25.07%
Percent Change in Political Voting Gap (+ = Democrat; 2000-2012) + 8.37% + 9.18% (_14% to 47%) 9.85%
Average Black/ White + Average Hispanic/ White + Average Asian/ White Dissimilarity Index (1990-2000-2010) 133.90 134.93 (75 to 194) 24.63
Percent of Population White only (2011) 61.86% 63.54% (ll%-95%) 17.09%
Percent Change in Median Income (2000-2010) + 23.23% + 22.84% (-6% to + 54%) 9.76%
Overall, this model is statistically significant and accounts for a sizable portion of
variance in anti-homeless law rates within the sample of 102 cities. AboveI presented
results and analysis for each individual independent variable as it relates to the dependent
variable. Of the ten variables that proved to hold significant relationships with anti-
homeless laws, five held weak to moderate relationships (average homeless population,
percent change in homeless population, growth of aggregate housing value, growth of HT
LQ, and total average dissimilarity index), two held moderate effects (percent change in
violent crime, and percent change in median owner/occupied housing value), while three
variables held the strongest association (percent change in property crime, percent white
only and percent change in voting gap). Appendix A contains an additional summary of
results for all independent variable tests.
Developing meaning and theory from these new results requires an examination
of results as they relate to one another and to theory. This conversation will inform
75


conclusions on whether hypotheses are supported or questioned. The next section will
contain this discussion.
Unfortunatelythis regression model only accounts for thirty eight percent of the
total variance in the model. Little inference can be learned from weak to moderate
associations, as they are arbitrary terms of comparison between variables with stronger
effects. The discussion that follows will focus on the five strongest associations, as weak
relationships are less influential as we seek to use this models results to understand the
reality of growth in anti-homeless laws.
76


CHAPTER V
DISCUSSION
The threat hypothesis states that rising crime and homeless rates will share
positive relationships with high levels of anti-homeless laws across U.S. cities. The
theory that homeless rates, crime rates, and anti-homeless law levels are associated is
founded in literature and public discourse. Wilson and Kelling (1982) connect the
homeless person to crime rates in their broken windows theory, stating, .. serious street
crime flourishes in areas in which disorderly behavior goes unchecked. The unchecked
panhandler isin effectthe first broken windowp. 4). Siegel (1997) also makes the
connection between the sight of homeless persons and the perception of crime, stating
that It was the street tax paid to drunk and drug-ridden panhandlers... It was the
threats and hostile gestures of the mentally ill making their homes in the parksSiegel
1997, p.169).
The perception of danger is theorized as fused with the presence of homeless
persons who are suspected in the Broken Windows theory to be provocateurs of crime. If
this theory, which has been employed by mayors such as Mayor Giuliani of NY, is a
valid theory, then there must be a relationship between the perception of crime (homeless
people and other broken windowsand actual crime; both of which are expected to
drive a municipality to pass anti-homeless laws. Results show that there are relationships
between crime, homelessness, and anti-homeless laws however, in strengths and
directions that greatly challenge the validity of the threat hypothesis as stated.
77


First, I will discuss the relationship between homelessness and anti-homeless laws
as seen through the results of this study before moving on to examine the relationship
between crime and anti-homeless laws.
Measures of homelessness in this study were not found to relate to anti-homeless
laws as expected. While average homeless populations did share some positive
relationship with the level of anti-homeless laws, an extremely weak effect was found.
The only inference that can be made by this result is that there needs to exist some base
homeless population in a city for there to be anti-homeless laws. In other wordswithout
a base level of homelessness existing in a city, a city would have no use for and feel no
need to implement anti-homeless laws. In addition, there would be little possibility of
anxiety or fear created from homeless populations that exist in very low numbers.
The weakness of the relationship between size of the homeless population and
number of anti-homeless laws indicates that beyond the possibility that there needs to be
at least a threshold number of homeless persons to inspire the passage of anti-homeless
laws. The average numbers of homeless in a city have little effect on the number of laws
prohibiting homeless behavior.
First, the variance in homeless populations across cities in the sample varies
drastically (194 to 62,850) with a large standard deviationwhile the regression effect
indicates a very weak relationship between these variances and the number of anti-
homeless laws. These two realities indicate, according to this model, that a city would
have to have its homeless populations increase by thousands in order to attribute any
additional law implementation to the homeless population rate if all other variables were
held constant.
78


An increase of 5,000 homeless individuals is not without precedent; however, the
diminishing effect of the homeless levels on laws, the necessity of all other variables
being held constant for an effect to emerge, and the fact that the vast majority of cities in
the sample had less than 5,000 homeless people, supports the interpretation of a weak
result. Average homeless populations ranged from 194-62,850 with a mean of 3,511 and
a median of 1500. In truth this fact means that the relationship is too weak to make
inferences. There may be some relationship that is not being properly examined by this
model, but within this model we must dismiss this variable as having little importance in
explaining the growth of anti-homeless laws.
The second measure used to test an expected relationship between homelessness
and anti-homeless laws was the growth rate of homeless populations from 2005 to 2012.
The threat hypothesis predicted that anti-homeless laws would, in some way, be
connected to cities with increasing homeless populations. In other words, anti-homeless
laws are understood to be a response to an actual increase in homelessness. Evidence
suggests that the reverse its true. A strong negative relationship exists between
homelessness rate changes and anti-homeless laws. Cities Where homeless populations
fell the fastest tended to have a higher number of anti-homeless laws.
To reiterate the conclusions so far: anti-homeless laws are only related to average
homeless populations in so far as there must be a large enough population for homeless
debate to form in the public and private discussion. Also, A citys homeless rate decrease
is related to high levels of anti-homeless laws. Thus, the threat hypothesis is disproven
insofar as it relates to homelessness. Relationships do exist between changes in
79


homelessness numbers and levelshowevernot in the way theory or the threat hypothesis
predicted.
Theoretical questions concerning Broken Windows theory and theories of threat
emerge from these first two results. For example, why do cities5 homeless populations
decrease? One could attribute this to policy. Meaning that effective homelessness
reduction results may accompany the implementation of anti-homeless laws; just as
promised by city leaders prior to the implementation of new anti-homeless laws, as in the
case of Denver, Colorado, and many other cities (Whelley, 2013).
A decline in housing cost, as Quigley, Raphael,& Smolensky (2001) predicts,
could decrease homelessness to some degree. Yet for this to be true housing cost decline
would need to be proven to decrease homeless rates faster than unemployment statistics,
as unemployment grew rapidly as a result of the recession. Regardless, I must seriously
question the validity of the threat hypothesis. More discussion pertaining to the decline of
homelessness will appear later in this section.
Crime rate results offer some defense for the threat hypothesis, however, average
crime rates for both violent and property crimes share no association with anti-homeless
laws. This fact indicates that the level of crime as well as the level of homelessness have
little to do with a political effort to criminalize homeless survival behaviors. This result
challenges established theory of threat, which holds that anti-homeless laws are inspired
by fear of crime.
Furthermore, changes in violent crime rates are negatively associated with high
number of laws. Cities where violent crime fell fastest have the highest number of anti-
homeless laws. This contradicts the theory of threat because it predicts that anti-homeless
80


laws are associated with falling crime as well as falling homeless rates, unless there is a
reverse causation present.
An argument supporting the presence of a reverse causal effect is not without
merit, in so far as change in violent crime rates and homeless rate change. The Broken
Windows theory would definitely support this assertion. However, the results of this
study will not be able to prove or disprove the presence of a reverse causal effect.
With that said, looking at the Pearson5 s Correlation results tells us that homeless
rates and crime rates are uncorrelated with one exception; average property crime rates
have a negative correlation with average homeless populations. These simple correlation
results suggest that in many ways homelessness and crime are not as linked as some have
predicted, and discourse, assumes even if they co-vary within the regression and the
simple correlation. Therefore, it would be hasty to draw support for a reverse causal
effect through a logical and theoretical connection between homeless rates and crime
rates that need to be reexamined.
In addition, because percent change in property crime has a positive regression
affect on anti-homeless laws, as well as a strong positive simple correlation with violent
crimes rates, the presence of a reverse causal effect may also imply that anti-homeless
laws drive up property crime. Either way, Broken Windows theory needs to be re-
examined as it applies to anti-homeless laws. Do anti-homeless laws decrease violent
crime rates and increase property crime rates across an eleven-year span, and if so, how
does theory justify the contradictory nature of these results? Again, the discussion of the
Pearsons Correlation is not meant to question the regression resultsbut to interpret the
81


relationships between the independent variables in order to create theoretical
understanding
Concerning falling homeless rates, some have argued that anti-homeless laws are
not effective in reducing homelessness (NHCLP, 2011; USICH, 2012). Evidence
supporting the effectiveness of Broken Windows policing is also mixed (Hinkle, 2014).
In addition, little evidence suggests that broken windows policing has an effect on violent
crime, but there is evidence of its effect in reducing property crime (Corman and Mocan,
2005). With that said, there is evidence to suggest that the effect on property crime is also
mixed. Arguing these two points, Corman and Mocan (2005) write,
It is important to emphasize that the arrests for felonies have the largest effect on
felony crimes and that the effect of broken-windows policing, although significant
for some crimes, are not universally significant, nor are they of great magnitude.
To put the broken-windows hypothesis in perspective, note that other cities also
experienced significant decreases in crime during the 1990swithout the
dramatic increase in misdemeanor arrests (pp.262-263).
The minimal effect on violent crimes of employing broken windows policing suggests
that anti-homeless laws are not causing a decrease in violent crime. However, more
research is needed.
Percent change in political gap also holds a significant (.034), but moderate (-
.20956), correlation with changes in violent crime according to the simple correlation
results. This means that there is a slight connection between cities with falling violent
crime rates and cities that are becoming Democratic. This demonstrates co-variance and
challenges arguments toward reverse causation because anti-homeless laws may be
affected by the same shifting demographics that affect anti-homeless laws, rather than
anti-homeless laws driving the changes in crime rates.
82


There is also evidence in revanchist literature that anti-homeless laws are lagging
indicators of previous violence, homelessness, and disorder. Pmijt (2012) cites Smith
(1996) and Peck and Tickell (2002), arguing that revanchist policies, such as anti-
homeless laws, are preceded by a retreat of the state welfare services and are a reaction to
Liberal policies of the 1960s. In this vein, it could be that that current political wave
of anti-homeless laws is a function of past policy and urban environments (the liberal
policy of the 19605s and corresponding urban decay), which has provoked a recent wave
of revanchist anti-homeless laws.
Pmijt (2012) indicates that the anti-squatting legislation and the conception that
squatters are dangerous in the Netherlands took place thirty years after violence occurred
between police and squatters and fifteen years after the peak of squatting occurred. Pmijt
(2012) writesthe time lag between the collective memory and 2010 anti-squatting
legislation supports the interpretation that an element of revenge is present^ (p.l117).
Similarly, currently falling homeless and crime rates amid a growth in anti-homeless
laws, may also indicate an element of revenge for past urban wrongs (which is the heart
of much of Smiths argument on current revanchist politics). All this being saida reverse
causal effect remains a real possibility and warrants further study. There exists enough
supporting evidence in the literature and in the significance of this model to continue with
this discussion without further investigation into a reverse causation.
The rate of property crime growth is positively correlated with the number of anti-
homeless laws. As discussed in the results section, the positive relationship between
property crime growth and high levels of anti-homeless law is more appropriately
described as follows: those cities where property crime rates dropped the slowest tend to
83


have the highest number of anti-homeless laws. This result generates two additional
questions to consider. First, how might property crime rates be a better predictor of anti-
homeless laws than violent crime and homelessness? Second, what circumstance would
create an environment where homelessness and violent crime rapidly fell while property
crime fell slowly or slightly increased?
Nathan Glazer (1979) created a hypothesis concerning the perception of crime as
related to the sight of disorder. His 1979 piece titled On Subway Graffiti in New York,
blamed the sight of graffiti not for causing crime per se, but for causing the perception of
crime to increase. This perception of crime is derived from ones insecurity and inability
to control his or her surroundings. Glazer (1979) writes;
He (the subway rider) is assaulted continuously, not only by evidence that every
subway car has been vandalized, but by the inescapable knowledge that the
environment he must endure for an hour or more a day is uncontrolled and
uncontrollable, and that anyone can invade it to do whatever damage and mischief
the mind suggests (p4).
Glazier (1979) continues, explaining that once an association with crime and the
sight of disorder is made, the next logical question is how can it be controlled? He states,
If the linkage is a common one, then the issue of controlling graffiti is not only
one of protecting private property, reducing the damage of defacement, and
maintaining the maps and signs in the subway... but it is also one of reducing the
ever-present sense of fear, of making the subway appear a less dangerous and
unpleasant place to the possible user. And so one asks: why cant graffiti be
controlled? (Glazer1979, p.5).
Glazer (1979) did not devote much of this thesis to the sight of homeless people,
however Wilson and Kelling (1982) made the connection three years later. In order to
better maintain public order, Glazer (1979) recommends an expansion of juvenile
detention, apprehension, and reform, to manage public space. In much the same way,
Wilson and Kelling (1982) recommend managing public property through imposing
84


restrictions on public behavior of the homeless, as they are equated with the disorderly
graffiti artists of Glazers (1979) hypothesis.
If Glazers (1979) hypothesis is trueproperty crime might be a more important
predictor of anti-homeless law rates than violent crime. More importantly, it may explain
why cities that saw the smallest improvement in property crime rates might be seeking
ways to deal with the perception of disorder.
Conversely, Hipp (2013) found that violent crime rates were the strongest
predictor of crime perception from 1970 to where his study ends in 1999. And during that
same time period, he found that property crime had little association with the perception
of crime. Hipp (2013) also indicates that prior to 1970, burglary (property crime) was
most responsible for shaping the public perception of crime.
Perhaps, because violent crime has fallen drastically around the country since the
19805 s (Levitt and Dubner2006)property crime rate change has once again become a
more reliable predictor of the public perception of crime. Property crime rates are higher
than violent crime rates in every city in the sample. Property crime is more visible and is
witnessed by more people on a day-to-day basis. It is logical to think that property crime
drives public conception of safety, however, that question was not addressed by this
study. The results of this model indicate that property crime is a twice more effective
predictor of anti-homeless laws than violent crime, and perhaps, that is due to dynamics
like these.
There is the real possibility that additional anti-homeless laws increase property
crime by virtue of anti-homeless laws being property offences for which people can be
arrested. However, Broken Windows theory suggests that all crime will fall after Broken
85


Windows policing is adopted. The long eleven-year property crime rate calculations and
anti-homeless laws are not necessarily a comprehensive test for broken windows
policing, but as Corman and Mocan (2005) indicate, may be part of a broader trend that is
influenced not only by Broken Windows policing strategies, but by other factors as well.
Therefore, these results most likely represent a political reaction to property crime more
so than other threat measures.
The threat hypothesis, for the most part, can be abandoned. While there are no
relationships between the absolute level of crime and the number of laws, a strong
relationship exists between changes in property crime rates and anti-homeless laws. At
the same time, these results contradict theory in many ways. Rather than a positive
relationship between homeless growth rates and violent crime growth rates, there is a
moderate to weak negative relationship. Homelessness, for the most part, can be
dismissed as an unimportant variable, especially considering the amount of variance
remaining unexplained. Thus a new understanding of threat as it relates to anti-homeless
laws reads as follows: anti-homeless laws are high in those locations where property
crime statistics are unchanging, slowly decreasing, or slightly increasing. It will be
important to discuss the other two hypotheses to see if the clarified threat hypothesis is
supported by other findings, as well as by the literature.
The neoliberal growth results, such as the threat results, challenge the stated
hypothesis. The neoliberal growth hypothesis predicts that economic factors associated
with neoliberal growth patterns will share a strong positive relationship with high levels
of anti-homeless laws across U.S. cities. A city5s number of anti-homeless laws was
expected to hold a positive relationship with all growth measures. The neoliberal
86


hypothesis suggests that because cities are global economic centers, each city must
compete on a global scale making policy choices and instituting priorities designed to
increase property values, attract global business, tourism, investment, etc. Quigley and
Taylor (2004) and others (Mitchell 2003; Wright 1997) argue that this shift in priorities
translates to negative consequences on those living in public places. Cities seeking to
grow at all costs inevitably rnn out of space and are forced to push the poor out to make
way for new development and investment. I hypothesized that such neoliberal growth
dynamics would have the most significant relationship with anti-homeless laws. The
results of the regression model, however, suggest that the hypothesis of neoliberal growth
as a predictor of anti-homeless laws needs significant refinement.
High tech sector growth was a particular variable that proved to have a
relationship meeting some level of expectations. Florida (2012) argues that technology
sectors are one of three main growth sectors in a city that indicate a creative and healthy
growing city. The Milliken Institute (2013) ranks cities by a variety of indicators, many
of which are tech growth indicators. One of these indicators, location quotient of high
tech sector growth (HT LQ), holds a moderate positive relationship with anti-homeless
laws. This result indicates that over the last four years, metropolitan areas that have seen
the most growth of high tech sector concentrations, relative to all other metro areas tend
to have the higher numbers of anti-homeless laws. This result supports the neoliberal
hypothesis because technology is one of the most important growth sectors for which
cities around the world are competing. It is also the only growth measure that correlated
as expected.
87


Unfortunately, the regression effect for HT LQ is moderate. Minimal support for
the neoliberal hypothesis can be given, in that this result does not indicate that the
potential neoliberal causes of anti-homeless laws are dominant. On the other hand, this
result does indicate that growth measures are an important part of explaining anti-
homeless law growth. HT LQ is only one measure of many variables that are related to
growth. According to this models predicted effectsHT LQ most likely has less effect on
anti-homeless laws than changing property crime rates and is similar to the effect of
changing violent crime rates.
Housing values did not associate as expected. Anti-homeless law levels had a
negative relationship with two separate measures of property values. Immediately this
pattern calls into question the validity of the neoliberal hypothesis as stated and demands
theoretical explanations for how anti-homeless laws can have a positive correlation with
one growth measure and negative correlation with another. First, I need to make a point
of clarification. Even though there is a negative relationship, these results do not
necessarily indicate that a decline in housing prices would correspond with an increase in
anti-homeless laws. The negative relationship, in this case, is the result of the fact that in
locations where housing prices grew more slowly over the past decade (housing prices
didn5t actually decline in the cities measured), anti-homeless laws were more prevalent.
Housing value changes negative association with anti-homeless laws challenge
the neoliberal hypothesis because, in reality, the result was almost the opposite of what
was hypothesized. This result indicates that anti-homeless laws are not positively related
to all growth measures, but only some. Perhaps high tech and other globally competitive
industries, such as tourism, or investment measures, such as foreign direct investment,
88


may have a positive association with the level of anti-homeless lawsbut general
property values have the opposite effect. This point borders on conjecture because no
data on tourism or foreign direct investment were tabulated for cities. Furthermore, the
negative effect of the percent change in median owner-occupied housing values on anti-
homeless laws was one of the five strongest relationships, and we must use this result in
the explanation of how growth may or may not affect anti-homeless laws. An additional
explanation would be that the housing crisis and recession could be influencing housing
value data. While this may be tme, there are no results of this study that directly links the
recession and/or recessions to the production of anti-homeless laws. The results indicate
that anti-homeless laws, in part, are a reaction to slowing housing growth.
Looking deeper into the connection between slow growth rates and anti-homeless
laws will require an understanding of how cities grew over the past decade. Frey (2012)
analyzed population growth data from 1980 to 2010 and found while 61 of the nations
100 largest metro areas grew faster in the 1990s than during the 1980s69 grew slower
in the 2000s than in the 1990s. Southern and Western metro areas still grew fastest in the
2000sbut exhibited the greatest growth slow-downs from the prior decadep.1).Frey
(2012) continuesstating Growth slowed considerably during the latter part of the 2000s
especially in bubble economy metropolitan areas (p.l). Frey (2012) indicates that the
fastest growing cities saw the sharpest drop in growth rates.
What is it about slowing growth that could be promoting the implementation of
anti-homeless laws? Frey5s (2012) arguments that growth slowed in the later part of the
decade most significantly in those places most affected by the housing crisis and that
those locations that grew the fastest in the 1990s slowed in the 2000s, provide two lines
89


of reasoning to explore. First, cultural perceptions of homeless people may be changed by
economic insecurity created by the housing market that had the slowest growth rates over
the past decade, and the largest property value losses during the housing crisis may have
driven some cities to implement anti-homeless laws. Second, those cities where housing
markets grew the slowest, and possibly, were hit the hardest by the recession, may have
pursued revanchist policies to address growing economic insecurity.
The first explanation originates from the work of Mitchell (2011), Ryan (1976),
Gans (1995), Barak (1991), and Kawash (1998), who argue that the very presence of
homelessness instills economic anxiety of fear and loss in those of the middle class. This
explanation has many problems in making sense of the data in this studys model
however. First, this explanation does not explain why the perceived threat of homeless
people increases as homelessness decreases. This reality only makes sense if fear is
generated from the experience of property crime and not homelessness, and homelessness
takes the role of a political scapegoat to attack property crime.
Second, it does not explain why tech growth would be associated with anti-
homeless laws. When high-tech industry grows in a city, would that not decrease
economic insecurity? In addition, previous research has shown that economic growth
rates (or declining economic growth) actually have little to no relationship on public
attitudes towards the poor (Kam and Nam, 2008; Hogan, Chiricos and Getz, 2005;
Mughan2007; Kluegel1987).
I reject the first explanation of perceptions of economic insecurity driving anti-
homeless laws through changing perceptions of the homelessness, because it does not
explain the correlation of high tech growth rates with more anti-homeless laws, nor is
90


there sufficient evidence that attitudes toward the poor are affected by housing statistics
or recessions. Economic insecurity created by slowing growth rates could be contributing
to a broader culture of fear, but more information is needed to solidify this conclusion.
The second explanation theorizes that those cities where economic insecurity has
increased the most are the same areas where revanchist policies, like anti-homeless
legislation, have been pushed most successfully. Nichols et al.(2011) used a unique
measure to quantify insecurity. Nichols and other researchers from the Urban Institute
and Rockefeller Foundation created a measure based on actual job variability, wage
variability and, other economic data. Nichols et al.(2011) write:
ESI (Economic Security Index) is a measure of the share of Americans who
experience large income losses. More specifically, it tracks the proportion of
Americans who see their available household incometheir household income
after paying for medical care and servicing their financial debtsdecline by 25
percent or more from one year to the next and who lack an adequate financial
safety net to replace this lost income (p.1).
Nichols et al.(2011) indicate that the recent recession created a sharp increase in
the economic insecurity index, but that it was already rising prior to the economic crisis.
Combining Frey (2012) and Nichols et al.s (2011)findings suggests that the largest
increase in economic insecurity happened in many of the same locations that Frey (2012)
argues saw slowing population growth in the 20005s.
The literature connecting 'punitiveness5 and economic insecurity is important to
note. While Johnson (2001) indicates that individual economic insecurity, as measured by
past events, is only weakly relevant to punitive public attitudes. Johnson (2001) and
Hogan, Chiricos, and Gertz (2007) find that negative measures of insecurity of future
economic possibility are connected to punitive attitudes and resentment toward welfare
recipients. Hogan, Chiricos, and Gertz (2007) write,
91


Full Text

PAGE 1

ANTI HOMELESS LAWS IN U.S. CITIES: A QUANTITATIVE STUDY OF THREE THEORIES by COLLIN JAQUET WHELLEY B.A., University of Dayton, 2006 A thesis s ubmitted to the Faculty of the Graduate School of the University of Colorado Denver in partial fulfil lment of the requirements for the degree of Master of Arts Political Science 2014

PAGE 2

! ii 2014 COLLIN J WHELLEY ALL RIGHTS RESERVED

PAGE 3

! iii This t hesis for the for the Master of Arts degree by Collin Jaquet Whelley h as b een a pproved for the Political Science Program By Tony R. Robinson, Chair Kathryn A. L. Cheever Thorsten H. Sphen April 29 2014

PAGE 4

! iv Whelley, Collin Jaquet (MA, Political Science) Anti Homeless Laws in U.S. Cities: A Quantitative Study of Thr ee T heories Thesis directed by Associate Professor Tony R. Robinson ABSTRACT The purpose of this study is to better understand why cities in the United States are increasingly employing anti homeless laws by testing three established qualitative theories This study uses quantitative methods to investigate the variance in the total number of anti homeless laws in a city with a sample size of 102 cities This study examines 15 independent variables to test the three established theories in question The f irst is a theory of homeless and criminal threat, the second is a theory of neoliberal growth, and the third theory considers the impact of cultural shifts and gentrification within a city. Results indicate that all theories ne ed to be revised, that cultur al political indicators have the most effect on anti homeless law growth and stagnating or rising property crime rates and stagnant or falling housing values also have a significant effect. However, the majority of variance in the number of anti homeless l aws in a given city remains unknow n. The form and content of this abstract has been approved. I recommend its publication. Approved: Tony R. Robinson

PAGE 5

! v DEDICATION I dedicate this thesis to the men, women and children who will be sleeping on the stree t tonight; t o those fleeing violence, pe rsecution, and famine and struggling with illness; t o those whom I have w orked with and those I have not; to those we have lost. Y our life and memory lives in those who choose to bare witness to your suffering, your struggle, your resiliency, and your love. Thank You.

PAGE 6

! vi ACKNOWLEDGEMENT It has been a privilege to be apart of the New Directions in Politics and Public Policy program at the University of Colorado Denver. I have never been pushed to work so hard. Thank you to the professors who have challenged, engaged, and taught me. I would like to offer a special thanks to Kathryn A. L. Cheever, Ph.D and Thorsten H. Sphen, Ph.D for all your work as part of my Thesis Committee and for the hours you spent as my profess ors I want to thank Tony Robinson PhD for inspiring me to continue working on issues related to homelessness and poverty, for your help in framing my questions and help in designing my experiment. One of the best decisions of my academic experience was t o take your class in Seoul, South Korea. My time at the University of Colorado Denver is an experience that has greatly expanded my desire for life long scholarship. Thank you to my friends and family. Thank you Scott for your mentorship, wisdom and friend ship. Thank you to my mother Sarah and father Peter, you have instilled me with roots of support and wings of wonder that I will cherish as long as I live. Thank you t o my brother Patrick and sister Katy, your genuine interest in my life and your far reach ing accomplishments have only ever pushed me to further actualize my own potential. The most important thank s go to my wife Merida Your constant support and affection has allowed me to accomplish more than I ever thought possible. Thank you

PAGE 7

! vii TABLE OF CON TENTS CHAPTER I. INTRODUCTION 1 II ANTI HOMELESS LAW AND THEIR PROLIFERATION 14 Anti Homeless Legislation 14 Theory of Threat 22 Neoliberal Economic Dynamics and Anti Homeless Laws 29 Cultural Shift Theory 3 5 III. METHODOLOGY 4 5 Th reat of Homelessness and Crime Test 48 Neoliberal Growth Test 49 Cultural Shift Test 5 0 IV. RESULTS 5 2 Descriptive Statistics 5 2 R egression Model Results Independent Variable Regression Results 5 7 58 Theory of Threat Test Results 6 0 Neoli beral Growth Test Results 68 Cultu ral Shift Test Results 72 V. DISCUSSION 77 VI. CONCLUSION 104 REFERENCES 110 APPENDIX 117 A. Results Summary 117 B. Regression Summary 118 C. List of Independent Variables 119 D. Standardized Results 12 3 E. Pearson's Correlation Matrix 12 4

PAGE 8

! viii LIST OF TABLES TABLE 1 Categories of Anti homeless laws 1 4 2 Anti homeless Law Categorization 4 6 3. Pearson's Correlation Results 5 4 4. Overall Regression Model Results 5 8 5. Mul tiple Regression Results 59 6. Threat Summary Statistics 6 8 7. Neoliberal Summary Statistics 7 1 8. Culture Shift Summary Statistics 75 9. Summaries of Revised Hypotheses 100

PAGE 9

! ix FIGURE 1. LIST OF FIGURES Residual plot of Ave Homeless Population and Anti Homeless Laws 6 4

PAGE 10

! 1 CHAPTER I INTRODUCTION On May 14 th 2012, Denver Colorado passed a city wide "Unauthorized Camping" ordinance This ordinance prohibits the use of anything but clothes on ones body to protect against the elements for the purpose of shelter and violations are punishable by up to a $99 9 fine. The city of Denver as well as the surrounding metro area, continues to have more homeless people than shelter beds and the availability of shelter beds are not always adequate for those with physical disabilities, chronic mental illnesses, substan ce abuse disorders, for transgendered persons or those without transportation (NAEH 2012 ; Denver City Council, 2012a ; Robinson, 2012 ) Criminalizing the act of sheltering with a blanket or any other protection against the elements in a city without unive rsal access to shelter is in fact criminalizing a group of people who cannot or will not access shelter. This ordinance was passed twenty one days after the United Nations commended the United States Department of Justice and United States Interagency Cou ncil on Homelessness (USICH) for withdrawing support for laws criminalizing behaviors vital to survival for those living on the street, while encouraging more effective and humane policy recommendation s (UN OHCHR, 2012). Despite the UN recommendation to re sist such laws, and despite the federal government withdrawing support from such laws, m any cities like Charlotte NC Tampa FL, and Denver CO continue to pass camping bans and other laws that criminalize survival behaviors of homeless peo ple

PAGE 11

! 2 Anti homele ss laws are laws and ordinances that criminalize the behavior of people living in public places ( such as sleeping, sitting, covering oneself, or even urinating ); which would not otherwise be illegal if a person had a private place to conduct those activiti es (NLCHP, 2011). In other words, these laws criminalize life sustaining activities of those living on the street. The National Law Center on Homelessness and P overty (NLCHP) (2011) surveyed 234 cities in the United States and found that forty percent of c ities ban "camping" in some areas of the city sixteen percent ban camping city wide, thirty three percent ban sitting and lying in particular locations, fifty six percent ban loitering in particular locations, twenty two percent of cities have a city wide ban on loitering, fifty three percent of cities ban begging in particular locations, and twenty four percent of cities ban panhandling city wide. In addition, NLCHP found from a study of 188 cities between 2009 to 2011 that there was a seven percent inc rease in laws prohibiting panhandling, a seven percent increase in laws prohibiting camping in particular places, and a ten percent increase in loitering bans in particular places (NLCHP 2011) A question arises as to the reason why laws that criminalize h omeless behaviors are increasing across cities in the United States even while the U nited N ations the U.S. F ederal government, and countless homeless advocate and aid organizations have spoken out i n opposition specifically to anti homeless laws W hy do cities pass these laws? The purpose of this study is to better understand the political, cultural, and economic, factors that push contemporary cities to adopt anti homeless legislation. Extensive qualitative research has been conducted highlighting the gr owth of anti homeless laws and conjecturing as to the reasons for that growth H owever much of the qualitative work has

PAGE 12

! 3 not been supported by quantitative methods. This study will use quantitative methods to build on and test the theories that already exi st in the literature regarding the promotion of anti homeless laws Denver City Councilwoman Susan Shepard offered testimony an d an explanation as to why she believed the most rec ent urban camping b an was passed in her city The Councilwoman disagreed th at the camping ban was a result of compassion for the homeless and joined the minority in opposition. She stated, Despite all the talk about compassion this is not really about compassion. How many times have we heard the words "we have to keep our Denver core strong" (or) "we can't weaken our downtown business sector or ever ything will fall apart? So is this about compassion or is it an economic strategy that we are talking about here? Personally...I think...this is code. Homeless are bad for business and they are ugly and dirty and stinky and they make our streets look bad and we are sick of dealing with them. So let's just sweep them under somebody else's rug...(Denver City Council, 2012b ). Councilwoman Shepard's assessment, while extremely critical of business interests, hypothesizes that Denver's camping ban is t he resul t of an economic agenda This hypothesis is not without merit. Over the past two decades the United States has experienced a "back to the city" movement where wealthy professionals leav e suburban areas for downtown urban centers. During this same period, as wealth and new urban professionals increasingly mo v e t o cities, laws criminalizing behaviors of homeless populations, referred to as anti homeless laws have also spread (Florida, 201 2; Mitchell 1997) Anti homeless laws are an attempt by city officials to control the development and shape of the city as it grow s Development and re development projects are taking place at great speed across the globe. Many cities in the United States are following the same trajectory of population and development growth. The new spaces that are created

PAGE 13

! 4 as part of the new urban growth dynamics however, are not always produced in a way to meet the needs of entire populations. More often the needs of t h e multinational corporations and power brokers who control investment dominate the creation of new urban space s and as a result, their needs are met before the needs of broader society Turner (2002) writes that public property ownership is shifting to p rivate investors who tend to constrict the democratic use of public space, as the new private spaces are not created for everyone. Turner (2002) focuses on public private partnerships, such as tax increment financing (TIF) and business improvement district s (BID) where public finances subsidize private development in an effort to speed up development. These partnerships result in the private control over large swaths of cities that are at times controlled by a single development firm. One example of explic it transfer of assets from public control to private control is the sale of public housing to private property owners and development corporations S imilar example s exist in New York City's Zuccotti Park and Denver Colorado's Eddie Maestas Park Zuccotti P ark, the creation of a public private partnership, is controlled by a private entity that has some control over the "public" use of that space to the extent it can define what "public" means Denver Colorado's Eddie Maes tas Park, locally referred to as "Tr iangle Park," is now in the control of Denver Urban Gardens. Denver Urban Gardens has funding and city support to completely replace the park that for years gave rest to homeless individuals waiting for food and shelter with an urban g arden designed to b eautify Denver's "skid row" (Meyer, 2013). Harvey (2012) argues that, "The actually existing right to the city, as it is no w constituted, is far too narrowly confined, in most cases in the hands of a small political

PAGE 14

! 5 and economic elite who are in a positio n to shape the city more and more after their own particular needs and hearts desire" (p. 24). Meaning, the development firms and private entities that have gained control of city blocks and parks are now able to regulate th e use of the space that they co ntrol. They regulate the use of space to coincide with their private interests (such as interests in carefully managed spaces attractive to shoppers and tourists) and are legally able to ignore broader social interests (such as the interest in providing pu blic spaces where even very low income people can gather) The increasing privatizatio n of urban space is part of a growing global and U.S. trend, which can be called "urban neoliberalism As part of a pro market turn in governing philosophy felt acros s the globe, urban neoliberalism involves the opening up of new market forces in urban spaces the devolvement and dismantling of the welfare and re distributive systems, such as public housing, and a political and economic commitment to attracting private capital investment above all else (Mitchell, 2003). The neoliberal cities of the United States are seeking to attract wealth through public subsidies of corporate projects, as well as through marketing strategies to brand cities as clean, safe and excitin g places to play (Wyly & Ham mel, 2003). The neoliberal goal is to build a clean, pure, safe environment for tourists shoppers and conventioneers, as well as spaces for those with expendable income to spend in new retail, sports, entertainment and luxury housing venues (Mitchell, 2003; Amster, 2003, Sennatt, 1970; Ferrell, 2001). As part of the neoliberal turn, U.S. cities are also trying to attract direct foreign investment in developments and post industrial industries such as high tech, biotech and fina ncial services (Florida, 2012; Kriznik, 2011).

PAGE 15

! 6 Some scholars argue that t o accomplish these goals more and more cities are actively participating in excluding portions of their own populations from access to urban spaces and use the growth of anti homele ss laws as evidence Those promoting this perspective argue that c ities are choosing to ban behaviors of the homeless and seeking to rid themselves of homeless people s presence in order to create the space s attractive to capital investment Mitchell (19 97) writes, In city after city concerned with livability,' in other words, making urban centers attractive to both footloose capital and to the footloose middle classes, politicians and managers have turned to...a legal remedy that seeks to cleanse the s treets of those left behind by globalization and other secular changes in the economy by simply erasing the spaces in which they must live...(p. 305). The "legal remedies" Mitchell (1997) refers to are anti homeless laws. Today the trend continues; the prevalence of anti homeless laws con tinues to increase across U.S. c ities (NLCHP, 2011; USICH, 2012). Mitchell (1997) Smith (1996) and others argue that a s new city development is centered on building world class consumption spaces and purified playgroun ds for adults with expendable incomes, power brokers in U.S. cities are pushing to clear the homeless off the streets. However, Mitchell's (1997) and Council Woman Shepard's (2012) assessments of the economic causal factors pushing cities to pass anti home le ss laws are not the only theories existing in literature and have, for the most part, not been supported with quantitative evidence. The e conomic focus of M itchell (1997) and Shepard (2012 ) may not properly consider other possible contributing facto rs wi thin cities concerning the promotion of anti homeless laws Other factors, which might drive the increase in anti homeless laws, include crime rates, political party strength, and cultural changes. Cultural changes refer to shifting populations of a city t hat may be driving support and advocacy for a particular

PAGE 16

! 7 policy. The shifting makeup is tracked by observing changing racial groups, education levels, migratory patterns, wealth, and how groups are integrating or segregating across the urban landscape. The growth, decline or dispersion patterns of a particular cultural group may have an effect on the level of support for or acceptance of particular political policies. Alternative explanations for the contemporary rise of anti homeless laws are important to examine because anti homeless laws are not a new creation in the same way that neoliberal politics have dominated contemporary urban history. Anti homeless law s like homeless ness itself, have a long history in this country. While Brenner and Th eodore (20 02) indicate that neo liberal policies and reforms began to blossom across US Cities in the 1970's and 1980's with the elections of Prime Minister Margaret Thatcher and President Ronald Ragan, Wasserman and Clair (2011) indicate that Anti homeless laws or poor laws" can be traced back to the Middle A ges and were present in the United States shortly after the founding of the country Anti homeless laws are not a new political phenomenon, but a re emerging one. A historical perspective suggest s that there may be causal factors besides economic policy that drive the growth of anti homeless laws. Mitchell (2011) argues that historically, vagrancy laws were used to control a large population of poor, escaped and freed slaves, after emancipation, when reserve la bor accumulated as people moved to the cities. Similarly, d uring industrialization, especially during times or recession and depression, large numbers of single men as well as families constructed shantytowns and encampments along rivers and in parks and they were often the subject of past anti homeless laws (Mitchell, 2011).

PAGE 17

! 8 Wasserman and Clair (2011) argue that prior to the recession in 1890, the "homeless" or "hobos" were understood as a migratory workforce with a lifestyle that was somewhat glorified. As migratory employment vanished and the recession of 1890 devastat ed the demand for migratory work, those traveling workers became stuck in one place or another. Wasserman and Clair (2011) interpret this reality of a growing, semi permanent reserve army of labor in many cities as related to the deterioration in the public's conception of homelessness as well as to a re emergence of "poor laws." While large shantytown encampments are much less common in America today anti homeless laws have again re emerg ed. Mitchell (2011) writes that the re emergence of anti homeless legislation in the 1980's and 1990's, a re emergence that continues today, coincided with neoliberal changes in the way government welfare was administered as hard financial times created a cultural atmosphere of "compassion fatigue" for the homeless (p. 943). Conversely, Wasserman and Clair (2011), borrowing from Axelson and Dail (1988) argue that the conceptions of homelessness as deviant, dangerous and needing behavioral control began well before the political and economic developments of the 1980's and 1990's. Wasserman and Clair (2011) write, Current conceptions of homelessness are most directly rooted in the negative attitudes developed in the period, when homelessness transformed from a semi legitimate nomadic lifestyle to a public nuisance that offended the sensibilities of wealthier citizens (p.9 ). Wasserman and Clair (2011) do not explicitly argue that poor laws are a result of cultural changes brought on by increased homeless n ess rates and economic stagnation, but the foundation is laid for a deeper examination of various cultural views of homelessne ss and how they might drive

PAGE 18

! 9 anti homeless legislation, in addition to considering the eco nomic motivators of such laws. Wasserman and Clair (2011) argue that cultural conception s of homeless people, connected to a growing lack of mobility among growing homeless populations at the turn of the century, may share a relationship with the explosion of "poor laws" during that time. Advanc ing a similar "cultural" theory for anti homeless sentiment, Murphy and Tobin (2011) argues that the public conception of homelessness concerning causes and cultural assumptions about who is homeless have swung from structural causes and supportive polic ies to personal responsibility and punitive measures. Although there has been no quantitative study regarding the connection between the public conception of homelessness and anti homeless legislation some qualitative assessments argue that a view of home lessness as largely voluntary promotes municipal efforts to criminalize homelessness (Whelley, 2013). While economic factors may play a large role in shaping anti homeless legislation, further study is needed to examine cultural political and other factor s that might also influence the growth of anti homeless laws. Mitchell (2011) offers an effective qualitative explanation of how challenging urban political and economic environments fostered the growth of anti homeless laws in the 1980's and 1990's. Ho wever, even as the economy recovered in the 1990's and through the boom and bust cycle of the 2000's, the prevalence of anti homeless laws continue d to grow. The specific reason(s) why anti homeless laws are growing in U.S. cities still remains somewhat of a mystery since those laws have tended to expand whether economic times are good or bad, in both flourishing and declining U.S. cities. In

PAGE 19

! 10 addition, Mitchell's (2011) understanding of the neoliberal determinants of the recent burst in anti homeless laws does not necessarily contradict Wasserman and Clair's (2011) cultural perspective, but cannot easily explain why poor laws similarly expanded in a very different time the turn of the century. The following literature review will seek to expand upon several unproven theories of why more and more cities are seeking to criminalize the survival behaviors of those living on the street. The literature review will be separated into two sections. First, the nature of anti homeless laws will be explained in greater detail, and the thesis will offer a brief introduction of two differing perspectives concerning support for and opposition to these laws. The next section will focus on three different theoretical explanations for the recent increase in anti homeless laws The first explanation which I call the "threat hypothesis ," will focus on the possibility that growing homeless populations and crime rates instigate a perception of threat in the minds of some urban residents concerning the homeless T his perception of threat leads localities to pass more laws restricting the behaviors and presence of homeless people in public locations The second theoretical explanation will discuss neoliberal economic growth as a driving force behind the growth of anti homeless laws in that increased global competition for cap ital investment is alleged by some to lead cities to be more restrictive of homeless behaviors in an effort to create downtown spaces more attractive to international investors The last explanation will focus on growing racial and political groups namely w hite and affluent professionals, which may be shifting the city culture towards fear of urban disorder This shift in culture instigates a culture clash between new white and affluent residents and

PAGE 20

! 11 urban stre et culture that existed prior. The clash creates cultural panic and pushes the cities establishment toward the regulation of public spaces (i.e. anti homeless laws). As urban areas attract new residence and racial diversity expands across those same urban landscapes, the cultural makeups of cities are changing. Political tools to control the public landscape may be supported out of a reaction to cultural disorder. Cultural fear may be connected with the threat and neoliberal theories; however little is kno wn about possible effects of shifting culture on anti homeless laws. Most of the existing literature is founded on qualitative analysis in that scholars have presented logically compelling theories explaining the growth in anti homeless laws supported b y analysis of the discourse of urban officials, etc. But there has been very little quantitative testing of these theories a gap that this thesis will work to address. The three established theories investigated in this study are tested using a quantitati ve method through multiple regression analysis. The number of anti homeless laws in one hundred and two American cities will be compared with possible correlating data associated with the three theories presented above The conclusion of the quantitative ana lysis offers some surprises. In the qualitative scholarly literature the neoliberal growth theory is one of the most common explanations accounting for the rise in anti homeless laws. For this study, t he neoliberal theory was hypothesized to hold the strongest relationship to anti homeless laws. However, contrary to prediction, the evidence shows that e conomic growth measures actually have the least powerful relationship with anti homeless laws. F urthermore, slower growth rates seem to have more effec t on anti homeless laws than more robust

PAGE 21

! 12 growth rates. It is hard to attribute the rise in anti homeless legislation to a surge in neoliberal economic growth patterns, therefore. Another common theory is that rising rates of homelessness and rising crime rates create a "threat" in people's minds, helping to explain a surge in anti homeless laws. Again, the hypothesis of this thesis, that the threat theory was expected to hold strong relationsh ips with anti homeless law, is largely challenged T he data do n ot support this hypothesis in its entirety Evidence suggests that the number of homeless persons in a given city and average crime rates have little relevance in a discussion related to the growth in homeless laws. At the same time anti homeless laws m ay be positively correlated with falling homeless rates and falling violent crime rates (however, they are also correlated with rising property crime rates). In the end, this thesis will argue that the quantitative data show that the cultural shift theory proves to have the most statistically significant relationship with anti homeless laws and not in the way first predicted. Rather than complimenting the neoliberal theory with corresponding demographic trends indicating gentrification and segregation of we alth and race, the results indicate a cul ture shift of a different kind. According to the results of this study the number of anti homeless laws in a city holds the strongest relationship with a positive shift in Democratic political voting growth (i.e., the stronger the rate of growth in the Democratic vote, the more anti homeless laws result). This finding is supported by smaller, but significant, cultural associations between a decrease in urban segregation and high percentages of non white residents, with higher numbers of anti homeless laws.

PAGE 22

! 13 Changing cultural landscapes, particularly in the ways listed in the preceding paragraph, may be instigating fear and insecurity that pushes cities to enact policies to control the public space. Pub l ic discourse may remain centered on the homeless and some may insist that this is a homeless issue, however the trends that associate with anti homeless laws are not a reaction to absolute or rising homeless populations. The reasons why cities are implementing anti ho meless laws are disconnected from homelessness because a nti home less laws are reactionary policies to cultural shifts slow economic growth and stagnant or rising property crime rates This conclusion contradicts much of the literature and complicates the interpretation of discourse centered on the economy and homelessness.

PAGE 23

! 14 CHAPTER II ANTI HOMELESS LAW AND THEIR PROLIFERATION Anti Homeless Legislation Anti homeless laws are city ordinances that criminalize the behavior of individuals living in public spa ces (NLCHP, 2011). The National Law Center on Homelessness and Poverty (NLCHP) and the United States Interagency Council on Homelessness (USICH) group these laws into six main categories, as can be viewed in Table 1. Table 1 : Categories of Anti Homeless L a ws Categories of Anti homeless laws as defined by NLCHP (2011) and USICH (2012) (1) Enactment and enforcement of laws that make it illegal to sleep, sit, or store personal belongings in the public spaces of cities without sufficient shelter or affordable housing; (2) Selective enforcement against homeless persons of seemingly neutral laws, such as loitering, jaywalking, or open container ordinances; (3) Sweeps of city areas in which homeless persons live in order to drive them out of those areas, frequ ently resulting in the destruction of individuals' personal property, including important personal documents and medication; (4) Enactment and enforcement of laws that punish people for begging or panhandling in order to move poor or homeless persons out of a city or downtown area; (5) Enactment and enforcement of laws that restrict groups sharing food with homeless persons in public spaces; (6) Enforcement of "quality of life" ordinances related to public activities and hygiene (e.g. public urination) when no public facilities are available to people without housing. (NLCHP, 2011; USICH, 2012) The most prevalent anti homeless laws are panhandling, loitering, and camping ordinances banning these kinds of behavior in particular places (NLCHP, 2011). Ci tywide ordinances that ban the same behaviors across the entire city are less common, but much more restrictive. The NLCHP (2011) found that out of 234 cities, sixteen percent banned camping citywide while forty percent of the cities banned camping

PAGE 24

! 15 in particular places. 1 NLCHP (2011) also found that the prevalence of anti homeless laws continues to increase. As mentioned before, f rom 2009 to 2011 there was a seven percent increase in anti panhandling laws nationwide a seven percent increase in camp ing bans for particular places and a ten percent increase of loitering bans in parts of a city. The NLCHP (2011) also surveyed people experiencing homelessness as well as homeless service providers across 26 states in an attempt to better understand if the se laws were being enforced. Fifty five percent of homeless respondents reported being cited or arrested for camping or sleeping in a public place. As anti homeless laws grow in U.S. cities the theoretical understandings of these laws, as well as the und erstanding of the causes of homelessness, remain divided. Discourse surrounding the support of and opposition to anti homeless laws is generally divided by opposing theoretical frameworks of structure and agency ," as they apply to the causes and solu tions of homelessness. Neale (1997) writes, "A structural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he average temperature for all cities in the sample was 58.36 degrees Fahrenheit with an average January temperature of 38.93 degrees Fahrenheit. Temperature was an important variable to consider in order to rule it out as a possible confoun ding variable for the rest of the study. In the end, colder cities varied in the same way that warmer cit i es varied in terms of the number of anti homeless laws For example, in 2011 Corpus Christ i had 8 anti homeless laws (below average) with a n average yearly temperature of 72 degrees Fahrenheit and an average January temperature of 56 degrees Fahrenheit. Jacksonville FL, with comparable temperatures had 15 anti homeless laws in 2011. With respect to colder cities such as Detroit,

PAGE 25

! 16 explanation of homelessness locates the reasons for homelessness beyond the individual, in wider social and economic factors" (p.49). A structural approach holds that the structure of society is re sponsible for making people homeless through such mechanisms as a poorly functioning economy, inadequate affordable housing stock, and soaring medical expenses t hat drive people into bankruptcy Under this structural perspective, because society is respons ible for creating the environment causing people to be homeless it is therefore society's obligation to change the environment in such a way as to allow homeless people to improve their living conditions and become self sustaining. And, because the homel ess are largely not to blame for their situation, the behaviors that they must partake in on a day to day basis, are unavoidable behaviors of survival. Any attempt to criminalize their behaviors of survival is effectively criminalizing the person who has n o other choice but to be homeless. Mitchell (1998) writes in support of the structural viewpoint : "if homeless people can only live in public, and if the things one must do to live are not allowed in public space, then homelessness is not just criminalized ; life for the homeless is made impossible" (p.10). Conversely, an "agency" understanding of homelessness allows one to argue that anti homeless laws do not criminalize people but rather the behavior in which these peopl e choose to partake Wright (199 7) argues that authoritarian responses to homelessness, like anti homeless laws, emerge from an understanding that homelessness is a result of poor and illegitimate choices. Therefore, society has less of an obligation to ensure access to housing or shelte r. Most importantly is the fact that agency arguments support legislative tools that attempt to push people toward "culturally appropriate"

PAGE 26

! 17 public behavior because the "inappropriate behavior" of living on the streets is viewed as voluntary Wasserman an d Clair (2011) expand on this understanding of agency by including arguments drawn from human service medical and substance treatment perspectives These perspectives may initially contradict discourse s of agency, but inevitably provide cover and ammuniti on for agency and anti homeless law supporters alike For example, individual pathological explanations of homelessness, such as substance abuse or mental illness, place the cause of homelessness in the deficiency of the individual. This deficiency is ofte n understood as a wrong cho ice (substance abuse) or due to a deficient mental health or capacity (for which individuals are also often blamed for, at least to the extent they "choose" to live on the streets and not get adequate personal help) Those who "c hoose" not to seek treatment allegedly choose their substances or their illness over shelter and help In other words everyone, even the sick, is responsible for his or her own situation from an "agency" perspective The view of human service workers and mental health professional s may often be structural in that many of these people do not seek to "blame" homeless people for their situation, but the argument used to support a medicalized viewpoint is often adopted and used by governing officials in supp ort an argument of agency that allows for new laws restricting homeless behavior in public places (Wasserman and Clair, 2011, Mitchell 2011; 1997) Even the common substance treatment slogan "take ownership promotes this agency understanding (Wasserman a nd Clair, 2011). While taking ownership of one s health an wellbeing is vital to ending homelessness o n an individual basis, the barriers to becoming self sufficient and

PAGE 27

! 18 empowered to make healthy choices are complexities that are overlooked by those who m ostly understand homelessness as a choice. Lyon Callo (2000) explains the medicalization conundrum stating, On the one hand, recent efforts may have facilitated increased services to reform, treat, and retrain individualized homeless people, and such eff orts to improve the lives of some individuals who are homeless. On the other hand, however, the "continuum of care" (medicalized and individualized) approach also does not fundamentally address questions of access to and distribution of resources in the c ommunity..."disease" within the discourses of "helping" actually obliterates discussion of alternative explanations and thus hinders developments aimed at resolving homelessness through altering class, race, or gender dynamics. When homelessness is individ ualized and medicalized, those concerns remain peripheral to the central work of normalizing perceived shortcomings or deviancy within homeless people (p. 330). Agency perspectives stem from an understanding that homeless people are homeless because they are deficient and make bad choices. In this way the agency perspective logically support s the implementation of anti homeless laws t o help reform homeless behavior. The normal perception of shortcomings and deviancy also supports a view of homelessness as criminal. If homeless are considered deviant and criminal, as Lyon Callo, ( 2000), Wright (1997), Amste r (2003) chronicle than there emerges an argument concerning the degree of responsibility for deviant behavior and the right of the public to restrict it Mitchell (2003; 2011) names this dichotomous understanding of homelessness a discourse of "deserving" and undeserving poor." Mitchell (2011) connects the "undeserving" perspective with anti homeless laws by stating, The response was a criminalization of homeless people in many cities. Laws w ere passed that outlawed everything from sleeping outdoors, to sitting in sidewalks, to free food giveaways. A traditional divi sion between "deserving poor" ( women, children, and those who behaved themselves in ways d ominant society deemed sufficiently grateful to charity) and "undeserving poor" (men of working age, those who lived on the streets or in encampments and refused to enter shelters or rehabilitation p rograms) was reasserted with a vengeance (934)

PAGE 28

! 19 Once the behavior of the individual homeless person is deemed deviant than the discourse of personal responsibility and "un de servedness" forms and from that understanding a logical progression toward anti homeless legislation may develop The m ost important concl usion to take from these arguments is that, according to th e agency perspective, homeless people are predominantly the cause of their own situation Homelessness is the result of poor and illegitimate choices. From this perspective, policies regulating the "deviant" behavior of the homeless people are justified. Believing that homeless people are broadly responsible and are principle agents of their own homelessness supports the viewpoint that people are voluntarily homeless and their behaviors may be viewe d as criminal because those behaviors (such as sleeping or sheltering in public places) are inconsistent with the goal of social order or the goal of city development. By prohibiting "bad" behavior, the city forces people to make better choices. Sawhill (2 003) writes "unless the poor adopt more mainstream behaviors, and public policies are designed to move them in this direction, economic divisions are likely to grow" (p.1). Similarly, Matthew Arnold states, "without order there can be no society; and with out society there can be no human perfection" (as quoted in Mitchell, 2003, p. 14). Many homeless advocates, mental health and service providers, the Obama administration and the United Nations, who all offer a structural perspective on homelessness, crit icize this "agency" approach, and call for the end to anti homeless laws (NLCHP, 2011; USICH, 2012; UN 2012). In addition, recent scientific evidence suggests that authoritarian responses to homelessness are ineffective and counter productive to the goal o f decreasing the presence of homelessness in a city (USICH 2012, Cooter, Meanor

PAGE 29

! 20 & Soli, 2012;). The USICH (2012) cites NLCHP (2009) and Roman and Travis (2004) indicating that, Rather than helping people to regain housing, obtain employment, or access need ed treatment and services, criminalization creates a costly revolving door that circulates individuals experiencing homelessness from the street to the criminal justice system and back. Sweeps can also result in the destruction of the personal property of people experiencing homelessness, including identification documents and medication. It can be much more difficult to secure employment, benefits and housing with a criminal record. Many of these measures include criminal penalties for their violation; the refore, they actually exacerbate the problem by adding additional obstacles to overcoming homelessness (p 8). In this same tradition, a Denver Colorado community group, Denver Homeless Out Loud (DHOL) recently produced an evaluative report on the effect s of the urban camping ban on the homeless population. DHOL surveyed over 500 homeless individuals across the city The conclusions of the report indicated that since the city passed a restrictive ban on public "camping," "most respondents have not been a ble to access dependable shelter...(and) the majority of respondents say their life has become more challenging, more stressful, and less safe since the ban was enacted" (Robinson, 2012, pp.8 9). The report also indicates that 73% of respondents are turned away from shelter "with some frequency ... 66% of respondents who used to live downtown say they now usually sleep in more hidden and unsafe location (s) ...(and) 37% say they have sometimes chosen not to cover themselves from the elements (such as using a b lanket) due to the camping ban" (Robinson, 2012, p.8). Regardless of concerns of counter productivity cities across the country continue to implement anti homeless laws The literature indicates that the discourse of agency and medicalization conti nues t o dominate public debate. Whether or not the discourse of

PAGE 30

! 21 agency has grown is a question that has not been quantitatively examined, but is assumed and supported by qualitative assessments of the rise of anti homeless legislation Trying to understand why the discourse of agency has dominated public debate is similar to an attempt to understand why cities are increasingly using ant i homeless laws because assumptions of agency are fundamental in the promotion of anti homeless laws. However, trying to chart t he ascension of the agency discourse is a different question that will not be examined in the present study. In addition, there is a possibility that some come to the conclusion that anti homeless laws are appropriate through an individualized and medical iz ed understanding of homelessness This understanding is not necessarily formed from an "agency" understanding of homeless people's behavior. Rather, it may be argued that causes beyond a homeless person's control have led him or her into homelessness (su ch as mental illness or alcoholism), and yet it may still be wise to have a lattice of local laws that "force" that homeless person to seek the kind of indoors help (social workers, drug and detox counseling ) that might improve their lives, rather than mak ing it easy for troubled people to live on the streets For example, behavior modification, a strategy used to set boundaries between care givers and patients, uses the construction and destruction of boundaries to improve a patien t's focus on the most im portant task of care and wellbeing. Some argue that anti homeless laws are these kinds of artificial boundaries that push people toward help. Anti homeless laws may be considered boundaries' in the sense that they protect homeless people from the devia nt behavior that created their homelessness or keep them homeless. Even in this approach, however, the underlying assumption s are that

PAGE 31

! 22 people have the agency to access care or need to be forced in to accepting care as well as the availability of adequate care and services. The structural viewpoint opposes anti homeless laws because they c riminalize a class of people by criminalizing survival behaviors because supporters of this viewpoint argue homeless individuals are not wholly responsible for the situat ion in which they find themselves The agency perspective supports anti homeless laws because people either need to be pushed into making better decisions or their choice of homelessness is illegitimate and should be criminalized. Beyond the different unde rstandings of homelessness that support or object to criminalizing homeless behavior, there are economic, political, and cultural explanations of why the prevalence of anti homeless laws has grown. This next section will focus on three largely qualitative theories that were developed to explain the reasons why cities implement anti homeless laws. The three theories explored here are first, the homeless threat which suggests that the growing presence of homeless people and crime push cities to implement a nti homeless laws. Second I explore the possibility that the political and economic effects of the neo liberalization of city space might explain anti homeless laws, as concerns for economic competitiveness may drive political decisions to criminalize hom elessness. Lastly I will examine whether the growth of affluent professionals and a growing proportion and segregation of white residents within a city may have any effect on a city's level of support for anti homeless laws Theory of Threat

PAGE 32

! 23 The theor y of threat argues that anti homeless laws, tools used to establish and maintain order in a city, are a function of disorder. Historically, vagrancy laws were used to control poor populations. Mitchell (2011) argues, speaking about the creation of anti ho meless laws prior to the 1960's, that rising homeless populatio ns at the turn of the century were met with cultural panic in the emerging middle class. Mitchell (2011) writes, Three crucial results of the rise of this massive "reserve army" of migratory a nd often homeless men (and some women) are particularly important for understanding contemporary homeless ness in America. First, it generated a serious moral panic among the emerging bourgeoisie, especially in times of economic crisis when the bourgeoisie' s own precarious social stability was at risk. A result, vagrancy laws were reinforced and new policing methods sometimes quite violent were introduced to contain and corral the wandering poor while separatin g the worthy from the unworthy (p 936 ) These An ti homeless laws remained in place in many cities until the 1960's, when a series of federal, state and U.S. Supreme court cases struck down vagrancy, loitering and other anti homeless laws (Mitchell, 2011). Between the 1960s and 1980s, most cities saw a r elaxation of "public space" anti homeless laws, including groundbreaking cases allowing even mentally ill homeless people to refuse to go to shelters, as long as they were not an immediate threat to themselves or others. Siegel (1997) argues, "sometime in the 1980's we came to the end of our national romance with cities and the public spaces that define them" ( p. 169). Siegel (1997) also argues that those romantic and nostalgic notions concerning free and open and creative public spaces w ere replaced with f ear and avoidance as his conception of culture and safety within cities deteriorated (Siegel, 1997). Mitchell (2011) touches on two key variables that are possibly pushing contemporary cities to implement new anti homeless laws. The quote above suggest s

PAGE 33

! 24 that prior to the 1980's, there was a relationship between a rise in homelessness, fear being instilled in the middle class, and a rise in anti homeless laws. This theory argues echoed by Siegal (1997), that a rise in the homeless population creates a cul tural reaction across the middle class, which is self conscious of losing new social status, and is seeking to implement political policies that will solidify their way of life (Ryan (1976); Gans 1995; Barak 1991; Kawash 1998) Anti homeless laws are then a reaction to a real or perceived threat in that the presence of homeless people threatens the middle and upper class way of life. Mitchell (2011) also indicates that the threat instigated by the presence of homeless men and women is amplified during times of national economic hardship. Although Mitchell (2011) moves on to hypothesize about other factors that have pushed cities to implement anti homeless laws after the 1980's the theory of threat warrants further discussion. The perception of homelessness has been and continues to be associated in the minds of many with criminality. It is this association that is largely responsible for fostering perceptions of threat. Amster (2003) writes that for the past six centuries, homelessness has been associated with criminality. Homelessness may not b e seen as a crime itself, but is considered to be related to crime (Amster, 2003). This understanding of homelessness remains present in contemporary discourse advocating for laws regulating public space, like anti h omeless laws. Underneath this discourse, there exists a language of fear of losing ones city to chaos and disorder. In this vein, Wilson and Kelling's (1982) famous Broken Windows theory has significantly influenced public discourse and policy concerning homelessness and anti

PAGE 34

! 25 homeless law over the past three decades. Wilson and Kelling (1982) developed the Broken Windows theory during a period of growing crime rates in urban areas in New York City in the 70's and 80's. As an increase in crime occurred wit hin the same period as a n increase of homelessness in urban areas, Wilson and Kelling (1982) hypothesized that if a city wanted to significantly decrease major crime, prevent the further deterioration of the city environment and stop "white flight," cities needed to focus on policing minor offences such as the presence of the homeless in public areas It is minor offences such as panhandling and public sleeping that Wilson and Kelling (1982) believe wou ld escalate into larger crimes. Wilson and Kelling (19 82) argued that there is a causal relationship between the visibility of disorder, a lack of attention to small scale disorderly behaviors, the deterioration of community intervention and outlook, and rising crime. Broken W indows theory hypothesizes that a llowing disorder to exist on a small scale creates an environment whereby disorder become s more widespread, possibly violent, and destructive to the community that deteriorates in the midst of disorder. According to this theory, the presence of disorder do es not force people to commit crimes, but allows crime to flourish because disorder has undermined the population's ability to care for the environment within which they live. People who pass by care less and less about the environment and therefore the destruction of that environment becomes less offensive. In addition those who do care about the environment become discouraged by the presence of disorder, presume a lack of governability, and lose faith that positive civic action will create positive ch ange.

PAGE 35

! 26 Hinkle (2014) creates a four staged flow chart ex plaining the escalation of the Broken W indows hypothesis. Stage one is "disorder goes untreated and is noticed by residents," stage two is c itizens become fearful and withdraw from t he community," s tage three is "i nformal social control decreases and/or is perceived to be low by criminals," and stage four is "d isorder and crime increase as criminals increase their activity in the area" (p.214). Those who are destructive flourish while those who seek to build and create must find other locations and markets hospitable to their desires. When those productive people leave, the city will fail. In short, this theory suggests that minor crime (or even non criminal "disorder") can destroy cities by escalatin g into major crime and by dissuading productive populations from staying in the community Included in minor offences are the behaviors that homeless people participate in (such as gathering on street corners, sleeping in public, being in parks after curfe ws, etc.) as they live in public spaces. Wilson & Kelling (1982) support regulating homeless behavior because the homeless person is an example of disorder. By regulating the homeless person from being seen the city is saving itself from destruction. Wil son and Kelling (1982) write: The citizen who fears the ill smelling drunk, the rowdy teenager, or the importuning beggar is not merely expressing his distaste for unseemly behavior; he is also giving voice to a bit of folk wisdom that happens to be a cor rect generalization namely, that serious street crime flourishes in areas in which disorderly behavior goes unchecked. The unchecked panhandler is, in effect, the first broken window (p. 4) Wilson & Kelling's (1982) hypothesis uses a language based on f ear and threat. This same fear and threat of a city descending in to chaos manifests in discourse that support s anti homeless law In the last 30 years many elected officials have shared the same perception as Wilson and Kelling regarding a homeless threa t. City officials, similar

PAGE 36

! 27 to Wilson & Kelling fear that t he homeless presence threatens the future economic and cultural possibilities of cities (Smith, 1996). Wright (1997) for example, documents how the officials of Chicago and San Jose mobilized agai nst the homeless around the discourse of fear. As an example of city officials utilizing the discourse of homeless threat in their support of anti homeless laws, Denver Colorado's newly passed camping ban is a case in point. Denver adopted a citywide ca mping ban ordinance on May 14 2012 an ordinance that restricts homeless people anywhere in the city from sheltering themselves in any way from the weather with anything other than their basic clothing. Denver's Mayor, Michae l Hancock, and a newly elected Councilman Albus Brooks, started working on some form of a camping ban ordinance soon af ter taking office (Meyer, 2011) based on their perception of a rising homeless threat. Councilman Albus Brooks stated, "We believe that numbers (the rise of homeless po pulations in Denver) have been increasingly alarming so this is about public safety, sanitation, and the protection of our right of w ay" (Denver City Council April 24, 2011 ). The Mayor voicing a similar perspective stated, "We only have one Downtown. We can't afford to lose our city core,"(Meyer, 2011 ). In support of the camping ban, Denver City Councilman Charlie Brown also argued that the camping ban " is a fight for sanity, it is time we fight to change this culture of chaos in our city" ( Denver C ity Council, May 14, 2012). These three elected officials in Denver Colorado spoke about the fear of homeless people eroding the very fabric of the city environment by their presence. Similarly, former New York City Mayor Rudolf Giuliani not only explicitl y used Broken Windows theory in his argument in support of the regulation of space and

PAGE 37

! 28 cracking down on minor crimes but also utilized a language of fear and threat. Giuliani stated: We have made t he "Broken Windows" theory an integral part of our law enf orcement strategy. This theory says that the little things matter. As James Q. Wilson describes it, "If a factory or office window is broken, passersby observing it will conclude that no one cares or no one is in charge. In time, a few will begin throwing rocks to break more windows. Soon all the windows will be broken, and now passersby will think that, not only no one is in charge of the building, no one is in charge of the street on which it facesso more and more citizens will abandon the street to thos e they assume prowl it. Small disorders lead to larger ones, and perhaps even to crime." There's a continuum of disorder. Obviously, murder and graffiti are two vastly different crimes. But they are part of the same continuum, and a climate that tolerates one is more likely to tolerate the other. ( in Coman & Mocan, 2005, p. 237). In addition, scholar Fred Siegel (1997) uses this theory to justify criminalizing the behaviors of the homeless by equating homelessness with criminality in contemporary politica l theory. Siegel (1997) writing about the threat felt by citizens of New York City, states: What unnerved most city dwellers, however, was not crime per se but, rather, the sense of menace and disorder that pervaded day to day life. It was the gang of tou gh exacting their daily tribute in the coin of humiliation. It was the street tax' paid to drunk and drug ridden panhandlers It was the threats and hostile gestures of the mentally ill making their homes in the parks (Siegel, 1997, p.169). Siegel (1997 ), and others argue that the sight of disorder destroys the moral fabric and the core of cities. Disorder is defined in such a way that equates the sight of poverty and homeless people with the sight of disorder. Siegel attempts to separate the behavior it self from the nature of the person, in order to criminalize the behavior without admitting he is criminalizing the existence of a homeless person ; however, it is a half hearted attempt. As Wright (1997) suggests, this kind of discourse inevitably proposes that the homeless people are "dangerous" to society and must be pushed out of sight (p. 16).

PAGE 38

! 29 The language of f ear and threat is connected to the presence or the perception of public disorder. The language of fear and threat concerning homeless persons pe rsists throughout public discourse and political argument To explore to what extent such fears an d such discourse actually drive the expansion of anti homeless laws, this thesis will quantitatively test whether a growth in local homelessness and crime rat es have a relationship with the implemen tation of anti homeless laws. If the "threat" theory offers a good explanation for the rise in anti homeless laws, we should see that as rates of homelessness and crime actually increase in a city (thus increasing p erceptions of threats), a growth in anti homeless laws would follow. For this reason, t heory suggests that anti homeless laws will show a positive correlation with homelessness and crime rates in cities. Similar theories have been tested in other politic al science subject areas. Boushey and Luedtke (2011) tested a threat hypothesis to explain anti immigrant laws across state legislatures. They found a significant relationship between the percent rise in immigrant populations with a rise in the number of state immigration bills that were directed at controlling the influx of immigrants in that state (Boushey & Luedtke, 2011). The threat hypothesis entails an assumption that t he threat of rising disorder in the form of increasing homelessness (or a growth in actual crime rates) will inspire a municipality's government to enact anti homeless laws As opposed to this approach, I will explore in the next section how anti homeless laws perhaps are less a function of growing homelessness and crime, but rather are a function of the political and economic effects of the neo liberalization of cities. If no significant relationship is found between homelessness, crime rates, and anti homeless laws, then perhaps the language of threat is

PAGE 39

! 30 merely a tool used by politi cal actors and power brokers to drum up support among the populace for anti homeless laws that offer perceived financial returns to powerful city players. However, due to the richness of the literature and records of public discourse that supports the the ory of threat, I hypothesized that the measures of threat used would hold strong positive relationships with anti homeless laws. Bel ow I will discuss such economic factor that may be correlated with a growth in anti homeless laws Neoliberal Economic D yna mics and Anti Homeless L aws According to a n economic appro ach, c ultural variables such as the psychological reaction to the rise in homelessness and poverty are not sufficient to explain the re emerging growth of anti homeless laws. Exploring underlyin g structural conditions in the current urban political economy might supplement a focus on public disorder, in an of itself Cities tod ay are very different places tha n i n the 1960s. Globalization, neo liberal economic policies, and revanchist regimes have significantly shifted political, economic, cultural, and institutional relationships of the contemporary U.S. c ity, and this changing economic order is argued by some to be the true catalyst of rising anti homeless legislation. Revanchism refers to a re actionary and cruel policy of dispossession and control, enforced by the wealthy and well placed against the social rabbl e of their day ( Smith, 1996 ). Reva n chist policies are political reactions to perceived economic threats, more than a reaction to a fear of disorder and crime, per se. Smith (1996) understands revanchist policies to be less about fear and more about economic control. Smith (1996) argues that revanchism is a vengeful policy implemented and supported by the wealthy in order to regain econo mic control of the city

PAGE 40

! 31 that was "taken" from them in the 1960's and 1970's and in an effort to make large profits from controlling the reshaping of urban space "Taking back" inner city spaces from the unrest of the 1960s is not only a cultural strategy therefore, but can be a very profitable strategy as rent gaps are found and exploited through the upward development of once degraded spaces, previously inhabited by low income and homeless people. Peck and Tickell (2002) write, that in the 1980s, the nation state became the principal anchoring point for institutions of (gendered and racialized) social integration and (limited) macro economic management, neo liberalization was inducing localities to compete by cutting social and environmental regulatory standards and eroding the political and instit utional collectivities upon which more progressive settlements have been constructed in the past (p. 385). This shift dispersed power from the federal government to the state and then the state continued to devolve regulatory power to the municipalities. U.S. cities therefore, are now the key locales of competition that compete for capital whereas the broader nation stat e used to be in previous decades This period of political and economic change coincides with Mitchell's (2011) observation of the re emergence of anti homeless laws coinciding with the deconstruction of progressive social welfare programs at the municipal level amid a poorly functioning economy. Before I continue to unpack these political and economic arguments, it is necessary to further define three pertinent terms: globalization, neo liberalism, and revanchism. Globalization refers to the growing global interconnectedness of local, national and international economies (Kahler, 2004). Neo liberalism refers to political policies of free market capitalism that allow globalization to enter every part of the local economy. The trend of neo liberalization in U.S. cities includes the globalization of local

PAGE 41

! 32 markets including housing and developmen t market s a trend that leads city planning to become obsessed with global marketability (Harvey, 2012). Harvey (2012) exp lains that the shift toward neo liberalism has turned municipalit ies' focus away from social welfare policies of the New Deal/Great S ociety era and toward attracting wealth and producing spaces desirable to global investors. Harvey (2012) writes, "the politics of capitalism are affected by the perpetual need to find profitable terrains for capital surplus production and absorption" (p. 5). Revanchism connects globalization, neo liberalization and anti homeless law expansion together by explaining how the arguably cruel neoliberal policies of economic elites (such as laws banning public sleeping) are adopted in an effort to compete global ly, and as a specific reaction to a previously more progressive era Revanchists are supporters of the global neo liberalization of cities and control the power of city governments turning it toward authoritarian responses to homelessness City Officials adopt anti homeless laws because those laws allow powerful actors in a city to more effectively to pursue global investment. Revanchists seek a clean and safe environment to attract global capital under the neoliberal global economy in order to maintain th eir quality of life and grow local profit opportunities According to this "revanchism" framework, municipalities controlled by either party today have joined the neoliberal pursuit of creating clean and purified space which has necessarily involved the m in a harsh attack on the lives of the very poor There has been little study of political parties and whether different parties have differing levels of support for anti homeless laws; however revanchist theory suggests that all major political parties will support the implementation of anti homeless laws, since such laws

PAGE 42

! 33 are a reflection of the global era we live i n and are driven by structural factors of capital expansion that affect all political parties. Evidence of econo mic neoliberalism (including it s harsh revanchis t aspects) around the globe dominates contemporary theory regarding the explanation for the growing prevalence of anti homeless laws. Beckett & Herbert (2008) write, "cities that depend upon capital investment, tourism, retail, and subu rban shoppers for their economic well being, the environment on commercial streets has become the subject of much official attention" (p.17). Furthermore as cities focus on economic growth strategies that attract capital and depart from Keynesian policies of the past they must develop marketing strategies that support this goal (Beckett & Herbert, 2008, Mitchell 1997; Mitchell 2011). Similarly, Kriznik (2011) observes that contemporary cities are primarily focused on a global competition for capital and, as a result, must resort to marketing strategies that are primarily focused on attracting wealth. The re creation of spaces used as marketing tools, involves restructuring the meaning and purpose of space. Kriznik (2011) writes, Yet by reconstructing th e meaning of a place, city marketing not only promotes its qualities but also legitimates the interests of dominant economic or political groups" (p. 295). The se marketing strategies have been expanded to the legal regulation of space including a growth i n anti homeless law s Beckett & Herbert (2008) write "from a political economic perspective, the intensifications of urban social control measures stems from the ascendance of neoliberal global capitalism and the related transformation of urban economies" ( p. 16). Again, social control measures include anti homeless laws. The economic perspective

PAGE 43

! 34 understands the rise of anti homeless laws as a function of the degree to w hich a city has neo liberalized, meaning the degree to which the city is engaged in the pursuit of global capital. In this vein, Qu igley, Raphael and Smolensky (2001 ) and Raphael ( 2010) found a positive relationship between homeless populations and the median cost of rent an d a median rent to income ratio (meaning that as cities become econ omically more successful and experience rising income and rent levels, anti homeless laws expand) Qu igley, Raphael and Smolensky (2001 ) found that rental vacancy had a negative correlation with homeless populations and a positive relationship with a compe titive housing market. Qu igley, Raphael and Smolensky (2001 ) make a strong argument for economic contributing factors and solutions to homelessness. Quig ley, Raphael and Smolensky (2001 ) write that their study indicates, ... r elatively small changes in ho using market conditions can have substantial effects upon rates of homelessness. Consider, for example, a reduction in the rate of homelessness by one fourth. The quantitative results suggest that this could be achieved in the national sample of housing ma rkets by a one percentage point increase in the vacancy rate (from an average of 8.4) combined with a decrease in average mon thly rent to income ratios from 17.5 to 16.8. (p. 51) Quigley, Raphael and Smolensky's (2001 ) findings suggest there may be a conn ection between the revanchist policies and metropolitan growth that unintentionally promote homelessness b ecause there is a connection between the housing market and homelessness. Testing this theory c an be achieved through a national sampling of measures associated with neoliberal growth and revanchist policies present in Qu igley, Raphael and Smolensky 's (2001 ) article as well as by considering a range of other growth measures associated with neoliberal economic growth that may correlate with the growth in anti homeless laws If the level of homelessness is associated with neoliberal

PAGE 44

! 35 growth dynamics as predicted, there should be a covariance with these kind of economic indicators Therefore there should be a relationship between levels of development, th e rise of median rent, other economic factors, and the expansion of anti homeless laws. Economic growth indicators were expected to hold the strongest relationships with anti homeless laws An argument that economic and political factors influenc e the adop tion of anti homeless ordinances does not mean that there is no role for cultural explanations. Smith (1996) writes in this regard that the neo liberal discourse in urban areas often becomes focused on the ugliness of society, the disorder, the filth, and the violence that takes place on the street. Therefore, r evanchism includes culture in its definition while holding that the economic factors ultimately drive social control policy such as anti homeless law. To better separate the neoliberal hypothesis fro m the cultural shift hypothesis, however, further discussion of the neoliberal growth hypothesis will focus purely on economic indicators One significant problem for testing neoliberalism is that it is an elusive variable in that there is no real test fo r it. Neoliberalism is a broad term, without a clear set of operational variables that define it Although later in this thesis I do attempt to isolate economic variables that are associated with neoliberal priorities the limitations of quantitatively mea suring the broad concept of economic "neo liberalism" cannot be wholly overcome Before discussing the design of the experiment developed to empirically test these theories, I will turn to the final theory explaining the growth of anti homeless law s to be e xplored in this thesis Cultural Shift Theory

PAGE 45

! 36 The Cultural S hift theory is unique because it does not usually appear explicitly in literature discussing anti homeless laws Theories of threat and neoliberalism emerge as the most prominent defenses of ant i homeless laws However, cultural changes may occur as demographics shift as cities grow and prosper or decline These changes m ay influence the growth of anti homeless laws. The separation of the cultural s hift theory from theories focusing on crime rate s and economic imperatives is important for this study because culture changes resulting from neoliberal growth or as the result of factors not discussed, may be more responsible for the enacting of anti homeless laws than variables discussed in the threat and neoliberal sections Essentially, the cultural s hift hypothesis proposes that affluent professionals and new white residents moving into cities are culturally afraid of or culturally adverse to homeless people. This section will discuss three unexam ined cultural and political factors that may be associated with the rise in anti homeless laws The first is an examination of cultural changes associated with the growth in cities of a middle class and a young professional and upwardly mobile demographi c that may be grasping for power and supporting policies that support their private interests and cultural viewpoints Similarly, cultural fear of homelessness may grow as a function of new city residen ts, as o pposed to growth in homeless populations. Fer rell (2001) chronicles a cla sh of culture between residents that move to redeveloping areas and the resid ents of abandoned buildings, riverbanks, and train yards. The second possibility is that new urban professionals may experience fear when being faced with poverty and diversity on a larger scale then they are used too and are grasping to take control of their new environment This may simply be a fear of urban life

PAGE 46

! 37 and associated disorder separate from the f ear generated by the actual presence of homel ess persons. Sennett (1970) argued that urban residents were retreating from a respect and admiration for the creative potential of public spaces to a place of fear and avoidance of difference and change. The third angle of the "cultural s hift" theory is the possibility that party politics has a role to play in anti homeless legislation. Even though political dynamics are usually kept separate from cultural theories, changing political compositions are a method by which one can examine shifts in culture. Mitchell (1997) notes that many of the most restrictive cities in terms of homeless behavior are thought of as "Liberal" cities. Smith (1996) charts a much more conservative le d movement to criminalize homeless behavior and restrict public space. Neither a uthor presents quantitative evidence, however both present effective qualitative assessments. There is reason to believe that anti homeless law has a political component because their re emergence was formed from a neoconservative movement charted by both Smith and Mitchell. Yet Mitchell 's (1997) work is in some ways contradictory to Smith (1996) indicating that party politics might have less to do with the passing of such o rdinances as anti homeless laws. However both agree that Broken Windows policing originated from a conservative Republican political viewpoint and therefore some influence of party is still expected to remain. Symptoms of neoliberal economic growth cannot be sufficiently examined without discussing the cultural shifts that may occur as a city grows. C hange that is associated with growth namely gentrification, which entails demographic and cultural change may help explain the growth of anti homeless law.

PAGE 47

! 38 Wyly and Hammel (2005 ) hypothesize that a change in demography that reshapes cult ural and political dynamics within cities may be respo nsible for the increase of anti homeless laws. L ocal authorities in any city usually move quickly against street people doing any of these things; but our reasoning is that the policies are formalized only under certain circumstances, and that gentrification is one of the processes that helps to broaden support for explicit, city wide quality of life' ordinances" (pp. 9 10). If gentrification is one of the processes that broaden support for anti homel ess laws then there is an implicatio n of cultural change occurring as urban upper classes grow in a city As a city gentrifies the new influx of residents and reorganization of where they live, may create a culture supportive of anti homeless laws. Wyly and Hammel (2005 ) define gentrification by stating that, Gentrification is fundamentally about the reconstruction of the inner city to serve middle and upper class interests. When it avoids direct displacement, the process usually involves middle class or developer subsidies that cannot be seen in isolation from cutbacks in housing assistance to the poor and other attacks on the remnants of the we lfare state ( p. 5). Gentrification involves economic change and demographic change. Some of these changes may include changes in income levels racial makeup and segregation patterns in a city But as these factors change, there are likely to be powerful cultural changes in a city as well. Ferrell (2001) chronicles the shift and clash of culture between residents of abandoned public spaces homeless, graffiti artists, and other stree t dwellers and the new residents of redeveloped lofts and visitors of sports complexes that replaced them. Ferrell (2001) also indicates how the new residents of renovated spaces who moved to the new location to find an "authentic" and "cultural" urban experience immediately seek to cleanse the building and surrounding area of the culture that existed before their arrival, namely the street culture. Ne w residents inevitably support mo re

PAGE 48

! 39 development projects to fill the blighted areas surrounding them and fill the abandoned public spaces with new private "cultural" spaces. The new spaces, once home to the homeless and runaway youth, became intolerant of "street" culture. Ferrell (2001) also suggests that t he clash was not simply one sided. Graffiti artists and persistent squatters fought to maintain some hold on their old stomping grounds by moving to unseen parts and painting graffiti on buildings and passing trains to remind new and o ld residents of the culture that was supplanted I f anti homeless laws are a function of a city's changi ng culture (associated with neo liberal growth and revanchist policies ) they would also share a relationship with gentrification. If gentrification is associated with anti homeless legislation as i s the case with Wyly and Hammel 's ( 2005 ) study then the results would promote further investigation in order to uncover whether the shift in cultur al values, or the economic pursuit of neoliberal growth oppo rtunities was really the most responsible for the surge in anti homeless laws Though I recognize how intertwined these variables are in reality, t his study seeks to investigate the factors of neoliberal growth separate from the cultural and demographic c hanges that may be associated with such growth, in order to form a better understanding of the factors that may be most responsible for growing anti homeless laws. Wyly and Hammel (2005 ) agree with Mitchell's (1997) assessment of gentrification as a key aspect in explaining anti home less laws. Wyly and Hammel (2005 ) found that "gentrification is generally correlated with one strand, explicit anti homeless laws, but most of the variation among cities comes from the broader urban context in with reinvestmen t and revanchism have emerged" (p.11 12). Some of the

PAGE 49

! 40 variables playing out in a broader urban context" are racial segregation, wealth segregation and i nequality (Wyly and Hammel, 2005 ). Unfortunately Wyly and Hammel's (2005 ) sample size of cities stu died was to o small to make larger generalizations concerning culture, economics, and gentrification. In addition they measured segregation by observing racial segregation of one race at a time. There was no overall measure of segregation that might give a general understanding of how white and non white s are segregated within a city In short, no firm generalizations about segregation could be made from their work because different races segregate differently and make up different percentages of the popula tions for different cities. For example, the Asian population of San Francisco is larger and integrated differently than the Asian population of Boston or Denver and thus an overall segregation of minority group s proved hard to quantify in a standardized way from city to city Revanchism is an important term to discuss in greater detail because it is both connected to the culture of the city as well as to a broader discourse on anti homeless laws. Smith (1996) and Mitchell (1997, 2011) understand economic factors as primary drivers of revanchist policies. Revanchism is a politic of revenge on the lower classes by the upper classes. The upper class is seeking to "take back" a city stolen fr om them and seize economic potential. The lower classes include the homeless, but also include other populations that don't fit into a "traditional" picture of a revanc h ist s vision of a healthy city. Smith (1996) writes, More than anything the revanchist city expresses a race/class/gender terror felt by middle and ruling class whites who are suddenly stuck in place by a ravaged property market, the threat and reality of unemployment, the decimation of social services, and the emergence of minority and immigrant groups, as well as women, as powerful urban actors. It porten ds a vicious reaction against minorities, the

PAGE 50

! 41 working class, homeless people, the unemployed, women, gays and lesbians, immigrants ( p. 207 ) The cultural undertones and non economic factors that Smith (1996) touches upon are commonly considered as "a euph emism for class motivated warfare' on poverty deviance" (Papayanis 2000: as cited in Van Eijk, 2010). In other words, non economic theories explaining the em ergence of revanchist policies, such as anti homeless laws, are sometimes dismissed as secondary a rguments to the real motivating factor of economic class warfare However, Pruijt (2012), Johnsen and Fitzpatrick (2010), and Van Eijk (2010) refuse to dismiss cultural factors so quickly. Pruijt (2012) evaluated r evanchist theory by following the Dutch anti squatting movement. While Pruijt (2012) ultimately affirms the thesis of revanchism, Pruijt lend s a voice to the cultural dynamics of revanchism Pruijt (2012) indicates that policy makers use a discourse of exaggerated "moral panic" that draws out a discussion of "culture war". The "culture war" dynamic highlights the fact that more variables th a n economic variables may be in play for any given city. Pruijt (2012 ) writes, "The case of anti squatting legislation, as in Van Eijk's (2010) analysis of ur ban policy in Rotterdam, shows that rev anchism can have a strong cultural component and that it is not exclusively bound to the economic logic of gentrification and competition between cities as suggested by Smith (1996, 1998)" (p.1118). The culture war c omplements Ferrell's (2001) observations and is in turn supported by Richard Sennett's (1970) seminal work, The Uses of Disorder Sennett (1970) hypothesized that urban culture is turning away from a respect for the cultural creativity born in urban public spaces of disorder for a pursuit of unattainable purity or

PAGE 51

! 42 structured space fueled by a juvenile fear of disorder, change, and otherness; the city itself. Smith (1970) writes, By defining the "outsider" and the "otherness the "we" become solidified wit h other "we's" to separate them selves from the reality of pain and the toughness of life... it is bred out of the way that human beings learn at a certain point in their growth how to lie to themselves, in order to avoid the pain of perceiving the unexpecte d, the new, the "otherness" around them ( p. 39). Sennett's assessment of the origin of avoidance, a cultural fear of experiencing life of the "other" through contact in public space, suggests that movements to restrict the public spac e and protect residen ce from disorder is a cultural desire. This desire is reinforced by actual disorder irrational fear, and the very policies that seek to quell disorder and inevitably separate the "others" from mainstream society. Johnsen and Fitzpatrick (2010) depart ev en further from the economic explanations of revanchism. While conceding that the state was revanchist in tone and actio n and that such language was typically connected closely to economic arguments, the public discourse was more complicated. There were t hose who spoke with fear and revenge, but there were also those who supported exclusionary policies in Great Britain from a discourse of compassion. In other words, support er s of the regulation of space truthfully believed that the laws were in the best in terest of the homeless or the devi a nt This finding is echoed by the discourse present during the aforementioned Denver City Council debates over anti homeless laws in 2011 and 2012 The discourse of fear, threat and compassion were utilized in promoting the anti homeless camping b an at that time, on e of the harshest such laws in the United States (Whelley, 2013). Johnsen and Fitzpatrick (2010) similarly find that this kind of cultural political discourse was a vital factor in passing legislation in Engla nd.

PAGE 52

! 43 Johnsen and Fitzpatrick's (2010 ) political discourse is understood as an extremely hard to quantify variable subject to the shifting moods and conceptions of the public. Johnsen and Fitzpatrick (2010 ) used interviews to qualitatively measure public discourse. For this study, interviews were not an option, but o nly the aggregation of demographic variables that may or may not sh o w change in the makeup of a city. As stated above the cultural shift hypothesis proposes that affluent professionals and whi tes are culturally afraid of or culturally adverse to homeless people. According to this theory, if a city's population co mprises a higher proportion of w hite citizens, has higher incomes, higher education attainment and is more racially segregated, that city should have a higher number of anti homeless laws as a reflection of the cultural conflicts often associated with those variables Within this theoretical discussion there exists an unexamined implication that largely w hite affluent professional co mmunities m ay be culturally afraid of the homeless. In the threat hypothesis I discussed earlier, I proposed that an actual growth in homeless rates and crime rates might be responsible for the growth in anti homeless laws. What the threat theory doesn't a ccount for however, is the possibility that rising cultural thre at s emerge from new city residents moving into a city (affluent, white professionals), and a growing cultural panic, and not growing homeless populations or crime rates, is what really drives anti homeless laws. Similar to the unexamined cultural variables and adding to Johnsen and Fitzpatrick 's (2010) cultural political discourse perspective, political variables are largely unexamined in anti homeless literature. Mitchell (2011) and Smith' s (1 996 ) assessment concerning the relationship between political parties and revanchist policies over history

PAGE 53

! 44 indicates that anti homeless laws reemerged out of a conservative led neoliberal movement during Reagan's Presidency and wa s at that time drive n by a specific political party However Mitchell (1997) indicates the while anti homeless laws originate from conservative revanchist conceptions, he writes, "Perhaps the most stringent of the newest anti homeless laws are in stereotypically liberal cit ies of the West Coast (p.307). The different arguments of Mitchell (1997) and Smith (1996) regarding the relationship of anti homeless laws to political party control deserve more study. To the extent that changing political party fortunes are connected to anti homeless laws, I can conclude that the cultural s hift hypothesis might have significance because a shift in political party support is an additional way to gage changing culture by measuring changing political ideology. The cultural shift theory h ypothesizes that a demographic of non minority affluent professionals are culturally afraid of or culturally adverse to homeless people. T he shift toward a new wealthy demographic in cities is hypothesized to be partly responsible for the rise in homeless laws and more so than a rise in homeless ness, per se ( Florida, 2012) A new demographic arriving in the city is hypothesized to be fearful or adverse to homelessness as its existence is counter to the professional goals and tastes o f the urban yuppie cl ass. The cultural shift hypothesis is that if a city' s population has a higher proportion of white citizens, has higher incomes, higher education attainment and is more racially segregated, it should have a greater number of anti homeless laws.

PAGE 54

! 45 CHAPTER III M ETHODOLOGY All prior studies regarding reasons for the growth in anti homeless laws with the exception of Wyly and Hammel (2003 ), have utilized qualitativ e methods. Wyly and Hammel (2003 ) quantitatively studied a small sample of cities with regard t o one test, gentrification. Verifying qualitative research through quantitative research is important for the progression of knowledge concerning the criminalization of homelessness. Quantitative methods are chosen for this study because there is a void of quantitative assessment in anti homeless literature and little is known about the other issues that may or may not correlate with the prevalence of a nti homeless laws. This study utilizes a multiple regression analysis tool to complete its quantitative a nalysis. StatPlus software, an extension of Microsoft Excel, was used for all multi regression tests. An Excel spreadsheet was created to house all independent variable and dependent variable data. When an imperfection in the database was found, it was fi xed and prompted a random verification of other independent variable data. The study sample was confined to those cities in the United States (n=102) where data are available for all dependent and independent variables A ll cities in the sample have imple mented anti homeless laws; however the specific laws vary across cities in severity and purpose. The severity and intricacies of an ordinance and the specific process a municipality went through in order to implement such laws waere not the focus of thi s study. The National Law Center on Homelessness and Poverty (NLCHP) produces a study roughly every two to three years surveying cities that have implemented anti

PAGE 55

! 46 homeless laws. The NLCHP (2011) categorizes anti homeless laws across 8 categories and 14 s ub categories which provides a rough measure of the illegality of the act of being homeless in any given city. The categories with their corresponding sub categories appear in Table 2 The "Other" category includes a large array of ordinances that crimi nalize anything from sleeping in cars, to washing windshields, to banning the use of one s own property in public (NLCHP, 2011). Table 2 : Anti homeless Law Categorization Categories Subcategories Sanitation Bathing in particular places Urination/ Defe cation in public Begging Begging in public city wide Begging in particular places "Aggressive" panhandling Sleeping Sleeping in public city wide Sleeping in particular public places Camping Camping in public city wide Camping in particular publi c places Sitting/ Lying Sitting or lying in particular public places Loitering: Loitering/loafing in particular public places Loitering/loafing/vagrancy city wide Vagrancy Obstruction of sidewalks /public places Closure of particular public places Other: Other In order to quantitatively study the reasons for the prevalence of anti homeless laws each city in the sample needed to be assigned a single number representing the severity of anti homeless laws. The total number of anti homeless laws rec orded by NLCHP (2011) for any given city was used as a measure of the prevalence of anti homeless laws in that city The total number of anti homeless laws for each city was calculated by adding all laws recorded in the NLCHP (2011) report on the c riminal ization

PAGE 56

! 47 of homelessness. The total number of laws per city served as the dependent variable (DV) for the multi regression analysis. The sample size of this study is 102 cities in the United States. The NLCHP (2011) report on a nti homeless laws had a sampl e size of 248 cities from the United States and Puerto Rico. Acknowledging that the political, economic and cultural environments may be different in Puerto Rico than in the rest of the United States the Puerto Rican cities were omitted. In addition dat a was gathered for homelessness and crime rates and various city lev el census data points for 102 citie s across 39 states and the District of Columbia In short, the sample is confined to those cit i es where adequate data were available making the sample non random Howe ver, because the sample size is large and the geographic representation of different states and regions are accounted for this sample is generally representative of cities in the United States. Further discussion of representativeness of t he sample will take place in the results section. The independent var iables in this study were chosen based on the three hypotheses being tested. Data were collected for 144 indep endent variables in a database accounting for the different measures of th e three chosen hypotheses. In total, fifteen independent variables were selected and used in the regression model to test the three hypotheses. See Appendix C for a list of 144 independent variables separated by hypothesis category. The next three sections contain reasoning concerning the choice of independent variables to test each hypothesis.

PAGE 57

! 48 Threat of Homelessness and Crime Test H1: Rising crim e and homeless rates will share positive relationship s with high levels of anti homeless laws across U.S. ci ties. Testing this hypothesis required that homeless and crime statistics be tabulated. Following the method of Boushey & Luedtke (2011), the total rate of crimes and homeless people in an area and the rate of increase or decrease in those rates, were plo tted against the implementation of anti homeless laws during the time period where data is available. Crime statistics originate from the U.S. Department of Justice Uniform Crime Reporting Statistics Database last updated in 2011 with yearly reports from municipalities starting in 1985 (DOJ, 2013). Some large c ities such as Chicago Illinois did not report crime statistics to the FBI for the year 2000 and thus were omitted from the sample. Homeless statistics originate from the U.S. Department of Housi ng and Urban Development 's (HUD) yearly point in time homeless estimates. These estimates are based on counts reported by Continuum of Care organizations within states. Homeless data are made available in two forms, state level and Continuum of Care (CoC) level. CoC organizations are made up of one or more cit i es and counties that have joined together to address homelessness. This study uses CoC level data because they are superior to state data as a measure of homelessness in U.S. cities. Some states such as Rhode Island, Nebraska, and Montana, and cit i es such as Washington (GA) and Tempi (AZ) do not have CoC data or were not a primary city within the CoC program, and thus were omitted from the sample

PAGE 58

! 49 CoC areas vary in size depending on the city and st ate which they exist. For example Los Angeles City and C ounty has its own C ontinuum of C are program, while Fort Worth Texas includes Arlington and Tarrant counties in their CoC. Some CoC organizations' geographic areas changed drastically during the pas t 10 years affecting the validity of the homeless counts and these were omitted from the st udy. For example, San Diego's Co C grew from encompassing only the city in 2005 to include the entire county in 2012. Because the size of San Diego County is many t imes larger than the city of San Diego the San Diego CoC count of homeless persons is inconsistent across the relevant time frame, and therefore San Die go was omitted from the study. Rates of homelessness and crime were calculated using the data that w e re provided by two government databases. Both government databases indicate limitations on their ability to verify data because the city or CoC submitted data to the national database without independent verification However, there are no other data set s as comprehensive as these two databases (HUD and Department of Justice) and both are used frequently in academic research. See Appendix C for complete list of threat test variables. Neoliberal Growth Test H2: E conomic factors associated with neoliberal growth patterns will share a strong positive relationship with the growth of anti homeless laws across U.S. cities Contemporary urban theory argues that economic and political neoliberalism is deeply connected with anti homeless politics Wright (1997), Mitchell (2003) and Harvey (2012) argue that anti homeless laws are a result of neoliberal policy. Harvey (2012) and

PAGE 59

! 50 Mitchell (2003) insist neoliberal policy resulted in the expansion of private ownership and explosion of re developed public spaces which led to efforts to remove the homeless from those spaces Peck (2002) argue s that cities are all more or less caught up in this neoliberal era, and that th e politics and economics of neo liberalism are fused so that anti homeless laws are spreading across t he country The independent variables that will be focused on to test this theory will measure growth o f a city in terms of population, wealth, job s numbers of people in the creative class 2 ," the h igh tech sector, property values, rental vacancy rates, i ncome to rent ratios, and median cost of rent among others See Appendix C for complete list of n eoliberal test variables Cultural Shift Test H3: D emographic factors indicating political and cultural shift s will hold positive associations with anti hom eless laws The goal of this test is to quantify changes in culture. Changes in income, changes of education attainment, changes in wealth and racial segregation patterns, populat ion growth minority presence, political party strength and measures of crea tive class will be used to interp ret possible changes in culture and identify if particular cultural demographic variables associate with anti homeless laws. Many of the same measures used to measure gentrification will be used to measure cultural changes. Meaning, many of the same measures used to investigate neoliberal growth may also indicate cultural shifts within the city. Measures of racial and political shifts and racial concentrations will be used to give context to those measures that may support m ore than one theory. While each measure is only tested once in the multiple !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!! @ ;1,!<#+,$/'A,!#3$..*!+,-,+.!1')13?!,952$/,98!1')13?!%06'3,8!/?&'2$33?!?05()! &+0-,..'0($3.!40+7'()!'(!2+,$/'A,!0+!7(043,9),!6$.,9!'(95./+',.!BC30+'9$8!@D"@E:

PAGE 60

! 51 regression model, some measures will be relevant in t he discussion of more than one theory. R acial demographics statistics dissimilarity statistics, population growth gini inequ ality indexes measures of racial diversity, median income, and poverty data were used in this study. Segregation Dissimilarity indexes are a city block by city block measure of diversity that is aggregated to measure the total integration of two races in any given city. See Appendix C for a complete list of independent variables used to test the cultural shift hypothesis The political question for this study is whether the majority political party within a city has any association with the prevalence of anti homeless laws presumably the conservative Republican Par ty would be more likely than Democrats to support anti homeless laws P residential voting statistics from each county were used in the model to test this possibility Voting data from 2000 and 20 12 were used to calculate the average voting g ap s (in terms of the percent of the local population voting for one party versus the other) and percent change in voting gap was relied on as proxy for this test. Voting gaps are calculated by subtracting one p arty's proportion of the vote from the other. In this case, I chose to subtract percent Republican vote totals from percent Democratic vote totals for p residential elections The difference equals the percent voting gap or the proportional voting differenc e between Democrats and Republican. See Appendix C for a complete listing of independent variables to test the influence of local political party power Unfortunately, only county lev el data were available for the measure of political party strength Again this study u sed the best available data. In the next section I will describe the results of the multi regression analysis.

PAGE 61

! 52 CHAPTER V R ESULTS The results section is separated into five subsections. The first examines the descriptive statist ic s of the st udy and the results from a simple Pearson's Correlation test The Pearson's Correlation tests for existing linear relationships between two sets of data. The second examines the overall strength and statistical significance of the multiple regression model The remaining three sections are devoted to presenting the regression results of the three theoretical tests : t hreat, neoliberal growth and culture shift. Descriptiv e S tatistics The sample mean of anti homeless laws across 102 cities is 10.8 anti homele ss laws and the standard deviation is 2.68 laws. In the larger sample of 189 US cities the mean number of laws per city was slightly lower (10.1 laws per city) with a higher standard deviation (2.99). Similar mean and standard deviations between samples st rengthens generalizability. The highest number of anti homeless laws in a city was sixteen laws (Colorado Springs CO; Los Angeles, CA ; Ashville, NC) and a low of two laws ( Fall River MA). The median of the sample is 11 while the mode is also 11. The geo graphic dispersion of cities in the sample has cities of 38 states and Washington DC. Seventeen of the cit i es in the sample are in the Northe ast (ME, NH, VT, NY NJ, PA, DE, RI) seventeen cities are in the Midwest (OH, IN, IL, MO, KA, NE, IA, SD, ND, MI, WI, MN), twenty three cities are in the W est (CO, WY, MT, UT, NV, AZ, NM, CA, OR, WA, ID ), and 45 cities are in the South (TX, OK, LO, AL, AK, MS, FL, TN, KT, WV, MD, VA, NC, SC). Twenty seven of the southern cit i es are in the mid and south Atlantic region s while eighteen cities are from the gulf region as well as inland

PAGE 62

! 53 southern states. The sam ple mean for Northeast cites was the lowest ( 8.47 laws) while the Midwest region had the highest ( 12 .06 laws); closely followed by the West (11.78 laws) and the S outh (10.71). The minimal variation in the dependent variable geographicall y 3 and statistically may explain why detecting meaningful patterns is difficult. Cities in the sample varied greatly in terms of population. Naples FL (20,091) was the smallest ci ty while New York was the largest (8,128,980) (ACS 2011). Only 18 cities in the sample had populations under one hundred t housand and the mean population was 477,555. Twenty six cities had populations over five hundred thousand and seven cities had popu lations over one million. The majority of the sam ple (fifty eight cities or 56.86%) had populations above one hundred thousand and smaller than five hundred thousand people. No relationship was found between population sizes or the changes in population an d number of anti homeless laws. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!! F G(!$(!$//,%&/!/0!'9,(/'-?!),0)+$&1'2!/+,(9!'(!$(/' > 10%,3,..!3$4.!G!'(2359,9!$! %,$.5+,!0-!/,%&,+ $/5+,:!G(/5'/'A,3?8!'-!302$/'0(.!$+,!2039!$(9!10%,3,..!&,0&3,!$+,! 05/.'9,!95+'()!/1,!1$+.1!4,$/1,+8!$!/1,0+?!0-!$),(2?8!$!/1,0+?!/1$/!.5&&0+/.!$(/' > 10%,3,..!3$4.8!40539!6,!5(9,+%'(,9!6?!/1,!-$2/!/1$/!.5+A'A'()!'(!/1,!2039!'.!(0/! .0%,/1'()!0(,!40539!3'7,3?! 2100.,!/0!90:! The average temperature for all cities in the sample was 58.36 degrees Fahrenheit with an average January temperature of 38.93 degrees Fahrenheit. Temperature was an important variable to consider in order to rule it out as a possible confoun ding variable for the rest of the study. In the end, colder cities varied in the same way that warmer cit i es varied in terms of the number of anti homeless laws For example, in 2011 Corpus Christ i had 8 anti homeless laws (below average) with a n average yearly temperature of 72 degrees Fahrenheit and an average January temperature of 56 degrees Fahrenheit. Jacksonville FL, with comparable temperatures had 15 anti homeless laws in 2011. With respect to colder cities such as Detroit, Cleveland, Buffalo, and Portland ( OR ), all had above average anti homeless law rates but lower than average temperature s Conversely, cities like Fall R iver, New York and Boston had lower than average temperatures and lower than average anti homeless law rates. No relati onship was found between the number of anti homeless laws and temperature and thus we can dismiss temperature as a possible confounding variable.

PAGE 63

! 54 A n examination of descriptive statistics was necessary to understand what typ es of cities are in the sample and to assess generalizability. Generalizability is challenged by the fact that this model is not comprised of a random selection of cities and cities cannot be placed in controlled settings. However, showing that a large sample of cities (102) come from diff erent regions of the country and vary in size, location and diversity strengthens the potential generalizabi lity of this model The Pearson's Correlation findings appear in Table 3. Note that a P test less than .05 indicates statistical significance. The P Tests measures the probability that the correlation strength is not the result of a random dispersion of va lues. A .05 significance indicates that we are 95% certain that the values are not random and thus there a relationship can be trusted Three statistically significant correlations exist within the data that were chosen and used in the multiple regressio n model Table 3: Pearson's Correlation Variable Correlation Coefficient P Test Average Homeless Population .12727 0.20242 Percent Change in Homelessness .22976* 0.02018 Percent Change Violent Crime Rate .23655* 0.01668 Percent Change Property C rime Rare .09108 0.3626 Average Property Crime Rate .11766 0.23889 Average Violent Crime Rate .09108 0.35263 Percent Political Gap Shift .42924*** 0.00001 Average Voting Gap .06919 0.48957 Total Average B+H+A/W Dissimilarity Index .11537 0.2482 Per cent white only 19022 0.05549 Percent Change Median Owner Occupied Value 17543 0.0778 Percent Change in Housing Unit Growth .11703 0.24143 Growth of High Tech Location Quotient .11557 0.24741 Percent Change Median Income .08242 0.4159 Growth of A ggregate Property Value .02298 0.73457 *p < .05; **p < .01; ***p < .001.

PAGE 64

! 55 The percent change in voting gap is by far the most significant (.00001) and most highly correlated (.42924) of all the independent variables. The percent change in voting gap i s almost twice as correlated tha n the only other two statistically significant correlations in the Pearson's Correlation model. The positive nature of this correlation indicates that the change being measured associated with anti homeless laws is o ne towar d Democratic voting patterns This result indicates that there is a moderate to strong relationship between cities that are increasingly voting Democratic and a growth in anti homeless laws but without consideration of any other possible confounding varia bles The percent change in violent crime rates shows the second most significant relationship (.0167) and second strongest correlation ( .237) in this model. The negative nature of the correlation coefficient ( .237) indicates that falling violent crime rates are correlated with higher homeless law totals. With that said, it is important to remember that this correlation is half as strong as the Political gap change and thus the correlation strength is comparatively small. The percent change in homelessn ess populations is the third most significant (.0202) and third strongest ( .23) correlation in the model. The negative nature of the correlation coefficient ( .23) indicates that declining homeless rates are related to higher anti homeless law rates. Howe ver, as in the case for violent crime rate change, the correlation coefficient for the percent change in homeless populations is comparatively small. It is also important to note that the percent change in homeless populations and change in violent crime r ates are not significant ( P test = .4249) when correlated with each other

PAGE 65

! 56 In the regression model correlations between independent variables are important for interpretation because it helps give c ontext to covariance and whether the two independent vari ables may be effecting each other or affecting the dependent variable in a similar way. In this case it is likely that the percent change in violent crime and percent change in homeless populations are not affected by each other as they mutually affect t he number of anti homeless laws in a city as will be discussed later. (See Appendix E for complete correlation matrix). One possible explanation for these results concerning crime and homelessness is that anti homeless laws cause a drop in homelessness a nd at th e same time reduce violent crime rates (as Broken Windows policing theory would suggest) H owever correlations do not prove causality. In addition, asserting that anti homeless laws cause violent crime rates and homeless rates to drop is complicat ed by the fact that homeless rates and violent crime rates are not significantly correlated with each other F urther study is needed to uncover whether falling crime rates and/ or homeless rates are a result of growing anti homeless laws. Two other results are important to note as they narrowly missed the .05 P test limit These variables are p ercent white only and percent change in median owner/occupied property value. The lack of significance does not necessarily indicate that a correlation does not exis t, but the results simply indicate that there is slightly more than a 5% chance that the observed correlation is coincidental In ad dition, the Pearson Correlation only measures one relationship at a time and thus the results do not take into considera tion that anti homeless laws may be influenced by a number of factors. The most important findings of the Pearson's Correlation is that a Democratic shift in politics

PAGE 66

! 57 is associated with higher rates of homeless laws while two smaller secondary correlation s exist between falling homeless rates and violent crime rates and high levels of ant i homeless laws. These results suggest possible relationships between a variety of factors hypothesized to hold relationships with growing anti homeless laws. However, t he simple Pearson's Correlation test cannot examine the effect of one variable while hol ding other variables constant, nor can it determine the direction of possible causation between variables (that is, a Pearson's Correlation can tell us that anti homele ss laws are correlated with low rates of homelessness, but it cannot tell us whether the rate of homelessness drove the passage of the law, or vice versa). To address these limitations the multiple regression model use d in this study identifies a chosen d ependent variable (number of anti homeless laws) and allows us to test whether specific independent variables affect the dependent variable while holding other independent variables in the model constant. In the next subsection I will discuss the overall r egression model results. Regression Model R esults The regression model results involve an evaluation of the regression model as a whole. Table 4 contains the results of the overall multi regression model The F Test (5.12) indicates that the results of th is model are not coincidental and thus the results of this study may be acceptable as long as the model results are st atistically significant. The P t est (4.17E 7) indicates that this model is statistically significant. With a F Test above 5 and a P test w ell be low .05 the presence or absence of individ ual relationships between the dependent variable and the independent variables are to be trusted according to their respective P test values.

PAGE 67

! 58 Table 4 : Overall Regression Model Results Model Results: Regres sion Statistics Value R Square 0.47163 Adjusted R Square 0.37947 F Test 5.11765 P Test 0.0000004167171 The Adjusted R Square Value (.38) indicates the proportion of variance explained by this model (38%) The .09 difference between the Adjusted R sq uared and the R Squared is likely due to the fact that five of the fifteen independent variables do not have significant relationships. The Adjusted R Square (.379 ) express es what can be learned from this model in positive terms. Put into negative terms, 62% of the possible variance in all cities within this model cannot be accounted for However, the results do explain a significant portion (37.9%) of total variance making the results of this model relevant to further scientific knowledge. Next, I will present the regression findings for all fifteen independent variables used in the regression model. Independent Variable Regression Results Of the fifteen independent variables used (out of 144 tested) ten had significant relationships with the dependent variable. Five independent variables remain in the model though they were without statistical significance because they help to focus the inferences made from those independent variables that do have significant relationships. Table 5 shows the multiple regression results for all independent variables including regression coefficients and significance values label ed P test Note that unlike the simple P e a rson 's C orrelation, these unstandardized regression coefficients can be greater or less than 1 or 1 because they exhibit effec t rather than correlation. The r egression coefficient can be used to predict the average change in Y (dependent variable) given the

PAGE 68

! 59 change of one unit of X (independent variables) if all other variables are held constant. With that said, it is inappropriate to make literal inferences concerning the effect of X on Y because in reality independent variables are not held constant and the majority of the variance is unexplained by this model. In addition, multiple linear regressi ons are one way tests, meaning that the effect on anti homeless laws is tested while the effect produced by anti homeless laws on the independent variables is not A ny effect of anti homeless laws on the independent variables is beyond the scope of this st udy. Table 5: Multiple Regression Results Independent Variable Regression Coefficient P Test Theory of Threat Test : Average Homeless Population (2005 2012) 0.00018** 0.00264 Percent Change in Homeless Population (2005 2012) 0.37769* 0.02282 Average Violent Crime Rate (per 100,000; 2000 2010) 0.00019 0.83014 Percent Change in Violent Crime Rate (per 100,000; 2000 2010) 1.70303* 0.03784 Average Property Crime Rate (per 100,000 2000 2010) 0.00026 0.12519 Percent Change Prope rty Crime rate (per 100,000; 2000 2010) Neoliberal Theory Test : 3.50985* 0.01977 Percent Change in Housing Units (2000 2011) 2.75001 0.20062 Percent Change Median Owner/Occupied Value (2000 11) 1.94764* 0.0181 Growth in Aggregate Property V alue (2000 2011) 2.7860E 11* .048 Growth of High Tech Location Quotient (2008 2012) 0.46465** 0.0043 Culture Shift Theory Test : Average Political Voting Gap (+ = Democrat; 2000 2012) 0.35506 0.76451 Percent Change in Political Voting Gap (2000 2012) 7.3437** 0.00438 Average Black/ white + Average Hispanic/ white + Average Asian/ White Dissimilarity Index (1990 2000 2010) 0.02843* 0.02023 Percent of Population White only (2011) 4.7586* 0.01149 Percent Change in Median Income (2000 2010) 2.75859 0.33269 *p < .05; **p < .01; ***p < .001. Examining trends in independent variable regression data is import ant in preparation for analyzing the overall results of the multiple regression model The distribution o f independent variable values may greatly alter the interpretation of results.

PAGE 69

! 60 The rest of this section is split into 3 subsections to better organize and analyze the regression results for various independent variables Each subsection will contain a pre sentation of descriptive statistics for each independent variable being tested. Unfortunately (but as is common in social science regressions) m any of the independent variables are measurements of different categories across varying scales. For example t he percent change of homeless population rates are a percentage measure (one unit = 1%) while the average homeless population is an aggregate measure of individuals (one unit = 1 homeless individual) Another example, percent change in political gap is ca lculated across a twelve year period while the percent change of homeless rate s are calculated across a eight year period. Relative comparisons between individual coefficients of di fferent scales cannot be made Those variables that are scaled the same ca n be compared relatively For example, relative effects of measures scaled using percentages, the most common measure in this model, can be compared if those measures are calculated along the same time frame. The non relative comparison of coefficients of different scales is not imp ossible. Rather than expressing the differing effects in terms of proportionality, like Pearson Correlation results (Change in Political Gap is almost twice as correlated to anti homeless laws as the change in violent crime rate s) I will compare predicted effects by using descriptive statistics, Pea rson's Correlation results, and qualitative discussion of predicted effects Theory of Threat Test Results : The first hypothe sis that was tested within the m ulti regression analysis w as the threat hypothesis To review, the threat hypothesis states that risin g crime and homeless rates will share positive relationships with high levels of anti homeless laws across U.S. cities As seen in Tab le 4 (Appendix B) and again in Table 5,

PAGE 70

! 61 the t hreat hypothesis tested for the presence of relationships between six different independent variables while controlling for other variables in the model. Four out of six threat variables hold significant relationships with the number of anti homeless laws : average homeless population, percent change in homeless population, percen t change in violent crime rates, and percent change in property crime rates. The two variables that did have significant relationships were average property crime rates and average violent crime r ates. According to this model therefore, the level of crime in a city holds no relationship with the number of anti homeless laws in a city but the change in crime rates might The average homeless population rate, however, shared a posi tive statistically significant relationship with the dependent variable. At first glance, one might conclude that this result would indicate that the higher a homeless population is the higher number of anti homeless laws that can be expected. This concl usion would be misleading. The regression coefficient (.00018) and the large variability in homeless rates as seen by t he range and standard deviation indicate that this is a very weak association. The regression model predicts that for a one unit increa se (one unit = one homeless person) there should on average be a .00018 increase in anti homeless laws. This result indicates that on average it would take an increase of 5,555.56 homeless persons (1 / .00018) to produce one new anti homeless law keepi ng all other variables constant. The variability apparent in the descriptive (summary) statistics supports this finding but not wholly On the one hand Los Angeles has the highest average homeless rate (62,850) over the last 8 years and has the highest l evel of anti homeless law rates (16), New York has the second

PAGE 71

! 62 highest homeless average (52,413) with half as many (well below average) anti homeless laws. Simply put, ho meless population has a very wea k effect on anti homeless laws according to this model A 5,555.56 h omeless persons increase to one additional law ratio only holds true if all other factors are held constant. Conversely, average homeless populations have the most statistically significant relationship with the number of anti homeless laws i n the entire model (.00264 ) warranting further discussion. On the surface this seem logical however this is not the case. Figure 1 is a residual scatter plot with a logarithmic trend line showing that as the homeless population increases the effect of ho meless persons on influencing additional laws diminishes. In short, while there is a statistically significant relationship between anti homeless laws and average homeless populations, it is weak and the effect declines in those cities with more and more homeless people. Figure 1: Residual plot of Ave rage Homeless Population and Anti Homeless Laws A city's percent change in homeless population rate measuring the growth or decline of homelessness had a statistically significant negative relationship with the 1 3 5 7 9 11 13 15 17 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Total Laws Ave Homeless Pop (2005-2012)

PAGE 72

! 63 number of anti homeless laws in that same city. The strength or effect of the relationship is weak (.38). This model predicts that as homeless rat es decrease by 1% over an eight year period there would be an average corresponding .38 increase in anti homeless laws. T he large range and standard deviations (Table 6 ) indicate that the results need to be analyzed further before developing an understanding of what the results mean A negative relationship suggests that as homeless rate fall, the number of anti homeless laws rise. We cannot conclude, however, that homeless rates fall as a function of anti homeless laws because multiple regressions can only predict for the dependent variable and not independent variables. It is counter intuitive to think de clining homeless rates influence cities toward passing anti homeless legislation, but not out of the realm of possibilities. Prujit (2012) indicates that anti squatting legislation in the Netherlands was passed long after the apex of squatting behavior. Th is line of reasoning will be discussed further in the discussion section. There is a possibility that reverse causation is being observed. Meaning that anti homeless laws modestly decrease homelessness. T he answer to that question is beyond the scope of th is study, but the possibility of a reverse causation warrants further discussion If homeless rates fall in a significant way as a result of anti homeless laws the n the advocates of broken windows policing and proponents of anti homeless laws may have so me justification when arguing that anti homeless laws force people off the streets and prevent people from "choosing" to be homeless in the first place. If this is in fact the relationship that exists then the next question would be the extent to which there is an increased demand for s ervice after the implementation of anti homeless laws Do

PAGE 73

! 64 those cities who employ such laws follow or precede the implementation of anti homeless laws with expanded services and housing ? O r are they isolated policies? Do homeless people move to more remote locations of urban areas after the implementation of a law making it more difficult to count them? More discussion concerning a possible reverse causation will appear in the next section s For the purposes of this study there is reason to suspect that anti homeless laws are a lagging indicator of homelessness, while keeping in mind the relationship may in fact be reverse causation or explained by an unknown tertiary factor to which both homel ess rates and anti homeless law levels are related In addition, the regre ssion relationship is very weak when compared to other variables, as we will discuss later in this section. Overall, t his model indicates that homeless rates and changing homeless populations have little effect on anti homeless laws when all oth er variables are held constant. A city's percent change in violent crime rate had a statistically significant negative relationship with the number of anti homeless laws. The regression coefficient ( 1.70) indicates a mod est relationship while the modest range and standard deviation support this conclusion. The mean rate of violent crime growth was 12.72 % while the standard deviation was 33.83%. This indicates that the vast majority of the sample cities had negative vi olent crime rate growth The scale of percent change in violent crime rate is calculated across an eleven year span while homeless change rates are calculated across an eight year time frame. With this said, the three year difference in calculating the rat e of change for these two variables will most likely not confound a general comparison between the two as they relate to anti homeless law. Violent crime rate has a

PAGE 74

! 65 4.5 times greater effect than the percent change in homeless populations on anti homeless l aw levels Meaning, violent crime rates have more effect than homeless ness rates on the passage of anti homeless laws though the relationship is an unexpectedly negative one The model predicts that a 1% decrease in violent crime rates across an eleven ye ar span correspond s with a 1.7 unit increase in anti homeless laws. While this may also indicate anti homeless legislation is a lagging indicator of high violent crime rates at some point in the recent past, we must be wary. L ike percent change in homeless population this result may be indicating a reverse causal effect (passing more anti homeless laws reduces violent crime rates, as advocates of such laws often argue) However, if this is true than most likely the reverse causal effect associated with ho meless rates and viole nt crime are independent from one another because the p ercent change in homelessness and percent change in violent crime have a Pearson's Correlation strength of .00168 with a P test of .9866. This indicates that percent change in ho melessness and percent change in violent crime are not correlated with each other Most likely violent crime rates and homeless rates are independent of one another. Meaning that, according to the Pearson's Correlation results a rise and fall of homeless rates is not related to a rise or fall in violent crime rates. This small finding simply provide s context for a possible reverse causation. I f a reverse causation is present for one of the two variables there would not necessarily exist a reverse causati on for the other because the two independent variables observed seem to be highly unrelated. This finding is also not presented to contest the result of the regression. It is imperative to interpret the results of this study with the understanding that at ti mes independent variables that e ffect the dependent variable may also be or may not be e ffecting each other.

PAGE 75

! 66 A city's percent change in property crime rate had a positive statistically significant relationship with the number of anti homeless laws. The e ffect (3.51) is twice as strong as violent crime rate effect and many times stronger than the percent change in homeless rates. This model predicts that a 1% growth rate i n property crime over an eleven year period would result in 3.5 new anti homeless law s. Unlike violent crime rate change, property crime rate growth seems to share a relationship with high levels of anti homeless laws. However, this simplistic view would be misleading. Only eleven of the one hundred two cities in the sample had positive pr operty crime rate changes over the past decade. This may indicate that for those cities where property crime fell the slowest and those where property crime grew are associated with higher anti homeless law levels when all other variables are held constant In addition, the res ults may indicate simply that property crime is much more widespread than violent crime and much more likely to be experienced and witnessed by the majority of city residents as well as commuters and visitors. Perhaps property crime d rives fear of public spaces and homeless person s more than violent crime rates especially as violent crime rates have continued to fall since the 80's and 90's. Percent change in property crime rates also has a positive statistically significant ( >. 0001 ) correlation with the percent change in violent crime rat e (.5308). The positive relationship between anti homeless laws that appears in the regression results stands contrasted to the negative relationship between the percent change in violent crime rates and anti homeless laws. If a reverse causation was present for property crime, then the results may indicate that anti homeless legislation would significantly drive up the rate of

PAGE 76

! 67 property crime over an extended period. While not out of the realm of possi bilities, that would stand in opposition to the Broken Windows thesis. Still, there is evidence that just s uch a phenomenon might occur. Johnsen and Fitzpatrick (2010) interviewed at least on e homeless individual who indicated anti begging laws increased his property crime behavior. The individual stated, "It pushed me to do a little bit of shoplifting, petty shoplifting, which I wasn't happy about (Johnsen and Fitzpatrick 2010, p.1710). T his single interview is not enough to challenge an established the ory like Bro ken Windows However if the act of being homeless is defined as being a property crime, as anti homeless laws do, than we would expect property crime rates to rise after the implementation of a new anti homeless law. Yet the Broken Windows th eory predicts that "cracking down" on minor offences like homelessness is predicted to have a significant effect on both violent crime and property crime The d ata gathered in this study do not lend support to such a theory. Putting thoughts of reverse c ausality aside, this model indicates that homeless rates share very small but statistically significant relationships to anti homeless laws while changes in crime rates have more effect when all other variables are held constant. Falling violent crime ra tes and homeless ness rates and rising property crime rates predict growth in anti homeless legislation. The effect of the absolute average homeless population is extremely small and the logarithmic trend line of the residual plot most likely indicates that a fter a certain size is reached in terms of homeless residents, the effect decreases and become s even weaker. Issues relate d to possible reverse causality will be discussed i n part in the discussion section, but will overall require further research to sa tisfy questions that have arisen during this section

PAGE 77

! 68 Table 6 : Threat Results Summary Statistics Independent Variable s Mean Median & (Range) Standard Deviation Average Homeless Population (2005 2012) 3,511.64 1500.75 (194 to 62,850) 8095.81 Perce nt Change in Homeless Population (2005 2012) +10.54% 5.43% ( 85% to +1138 % ) 137.16% Average Violent Crime Rate (per 100,000; 2000 2010) 888.88 808.78 (198 to 2231) 423.84 Percent Change in Violent Crime Rate (per 100,000; 2000 2010) 12.72% 18.54% ( 6 9% to 149%) 33.83% Average Property Crime Rate (per 100,000 2000 2010) 5421.36 5548.28 (1352 to 11763) 1680.62 Percent Change Property Crime rate (per 100,000; 2000 2010) 22.53% 25.79% ( 66% to 47%) 18.2% Neoliberal Growth Test Results : The second h ypothesis tested was t he n eoliberal growth hypothesis. This hypothesis states that e conomic factors associated with neoliberal growth patterns will share a strong positive relationship with high levels of anti homeless laws across U.S. cities Table 4 con tains regression result s for this section while Table 7 contains descriptive statistics for this subsection. Three of the four independent variables representing this theory proved to have statistically significant relationships with the number of anti hom eless laws however not always in the way that was anticipated : percent change in median owner/occupied property value (2000 2011), growth of high tech sector location quotient (HT LQ) (2008 2012), and aggregate growt h of property value (2000 2011) The independent variable that did not share an association was the percent change in housing units. Housing unit growth was one measure of population growth and development. Percent growth in housing units was hypothesized to have a significant positive effect on the number of anti homeless laws, but no relationship was found

PAGE 78

! 69 A city's percent change in the median owner occupied housing value had a statistically significant negative relationship with the number of anti homeless laws. The regression coefficient ( 1.95) suggests that there is a strong relationship and the small standard deviation and rang e relative to the mean indicate that outliers do not likely skew the strength of this relations hip. A 1% decline in median owner occupied housing value over an el even year period is expected to have more effect on anti homeless laws that a 1% decrease in violent crime rates but less effect than a 1% increase in property crime rates. Furthermore, the relationship is oddly negative: rising property values result in fewer anti homeless laws. Before we say with confidence that, all things being equal, a 1% decrease in housing values is expected to produce two anti homeless laws we must examine the descriptive statistics. One hundred percent of the sample had positiv e growth in terms of median owner occupied housing values. The effect that is observed is one that concerns slower growth as compared to other cities As median growth is slower than other cities even before reaching the negative, anti homeless laws are expected to result if all things are equal. Therefore this model may indicate that in those cities where housing prices rose the slowest over the past decade also tend to have the highest number of anti homeless laws. A city's aggregate property value also shared a negative statistically significant relationship with t he number of anti homeless laws; however the strength of the relationship was much weaker and the significance is much less than percent change in median o wner occupied housing value The p test value for significance was .048, only two hundredths below the cutoff of 05. The regression coefficient ( 0.00000000002786)

PAGE 79

! 70 at first gla n ce, indicates that the strength of the association is extremely weak. When put into predicted terms, this mode l predicts that every $ 35.9 billion dollars gained in aggregate value in a city will result in the decrease of one homeless la w. With a mean growth rate of $ 129.7 billion dollars over eleven years and a median growth rate of $ 49.5 billion dollars, the $ 35 billion dollar effect remains relatively small because the relationship is negative All things held constant, one additional anti homeless homeless law would result from a $35 billion dollar loss, a loss that did not occur in the sample. With that said, if the effect being observed is simply one of slower growth instigating anti homeless law implementation then a $35 billion dollars slower value growth over an eleven year period is not a outrageous figure. The predicted decrease in anti homeless laws associated with rapidly growing property values may be misleading. S imilar to median owner occupied housing value, the a ggregate value data show that only two cities (Dayton, OH and Detroit, MI) had negative growth rates in aggregate terms of the last ele ven years Therefore, like the owner occupied results, this model indicates that locations that grew slowest in terms of aggregate property value, tended to have the highest number of laws but the effect of aggregate value has less effect and may be less important than the owner occupied results. However it supports the median owner occupied hosing value results. One interesting relationship that was uncovered during the Pearson's Correlation was the very strong (.879) and statistically significant relati onship (.0000) between aggregate growth and average homeless populations. In addition, median owner occupied housing value held a .255 correlation and a .00983 significance level with average homeless populations. This suggests that those cities, possibly la rger cities,

PAGE 80

! 71 which grew the most (in aggregate value terms), tend to have higher homeless populations. This conclusion is intuitive, but will require further study to uncover other variables that predict homelessness The growth of a cit y' s high tech s ector s location quotient (LQ HT) relative to the national average had a positive statistically significant relationship with the number of anti homeless laws. The location quotient compares concentration s of high tech sectors on a national scale. Those lo cations with a LQ HT equal to one have a high tech sector (e.g. microchip company ) that is more heavily concentrated that the national average. While the regression coefficient was modest (0.46465), the P test indicates that this test was one of the most statistically significant relationships (0.0043). This model predicts holding all other variables constant, that a growth of 1 LQ HT over a 4 year period would result in one half of one ant i homeless laws. Table 7: Neoliberal Results Summary Statistics I ndependent Variable Mean Median & (Range) Standard Deviation Percent Change in Housing Units (2000 2011) + 11.67% + 8.80% ( 17% to 61%) + 11.78% Percent Change in Median Owner/Occupied Housing Value (2000 2011) +79.61% + 77.41% (12% to 188%) + 35.68% Growth in Aggregate Property Value (2000 2011) $12968544583 $4955077500 ( 921860000 to 312229155000) $34381319 293.95 Growth of Location Quotient of High Tech Sector relative to national mean. (2008 2012) .147 0 ( 4 to + 3) 1.438 The short time period may or may not be minimizing the effect of long term high tech growth. More research is needed to answer this question. What is important to note is that hi tech growth measures do deserve a place in anti homeless discussion ; however for this model they h ave had much less impact than previously expected. Thus the results

PAGE 81

! 72 indicate that LQ HT shares a moderate relationship with anti homeless laws. These results suggest that higher anti homeless law rates were found in metro areas that also enjoyed the large st growth of LQ HT relative to all other metro regions in the nation. Cultural Shift Test Results : The final test of this study was the culture shift hypothesis. As stated before this hypothesis predicts that demographic factors indicating political and cultural shifts will hold positive associations with anti homeless laws Table 8 contains descriptive statistics relevant to this subsection while Table 4 contains a summary of regression statistic s for this section. Three of the five independent variabl es studied share statistically significant relationships with the number of anti homeless laws : percent change in p olitical voting gap (2000 2012); percent of populations t hat is White only (2011); and total combined average of Asian/White, Hispanic/ Whit e, and Black/White dissimilarity index es (1990 2010) The two independent variables that did not prove to have significant relationships were average political voting gap and percent change in median income. Median i ncome could have been used as a neolibe ral test; however to test the cultural shift hypothesis I needed look at what type of demographic change was happening and this hypothesis predicted that an influx of wealthy professionals would be a predictor of anti homeless law s Regardless if this variable is best conceived of as a neoliberal test or cultural test it was not significant. Average political gap measuring the dominance of a political party in a county, also had no significant relationship. According to this model a particular politi cal party 's dominance is not important. However, perc ent change in political party voting strength presents a very different story.

PAGE 82

! 73 Similar to the Pearson's Correlation results, the multiple regression model finds the percent change in political voting gap to be the most important variable A city's percent change in the political voting gap shares a positive statistically significant relationship to the number of anti homeless laws when positive means Democrat s enjoy a majority in the city and negative me ans Republican s enjoy the majority I will reiterate that there was no relationship found between a particular political party's strength and anti homeless laws. However, t he percent change in political gap result suggests that those cities that became mo re Democratic since 2000 have the highest numbe r of anti homeless laws. The regression coefficient (7.3437) indicates that a 1% increase in Democratic presidential voting patterns results in 7.34 anti homeless laws over a twelve year period if all other va riables are held constant This finding may indicate that, in general, Democratic political growth expanded throughout urban areas while anti homeless laws mirrored this growth. However, the strength of relationship indicates that these results are not coi ncidental. Descriptive statistics support the existence of a strong relationship because t here is a small degree of variance in the sample data as evident by the small standard deviation (9.85%) and modest range in values. The P test value indicates that this relationship is one of the most statistically significant in the model (0.00438). These results indicate that anti homeless laws are highest in those cities that are increasingly becoming Democratic. A city's combined average dissimilarity index mea sure a measure of the degree of racial segregation in a city shares a negative statistically significant relationship with the number of anti homeless laws. The regression coefficient is small ( 0.02 ) suggesting

PAGE 83

! 74 that there is a weak relationship with the dependent variable. D issimilarity statistics are measured on a vastly different scale. With a maximum score of 300 (if Asian, Black and Hispanic populations were 100% segregated from Whites) and a mini mum score of zero (if Whit es were completely integrated with non White minorities) the .02 effects translates into 35.2 dissimilarity units Putting this in to positive terms, every 35.2 dissimilarity index units toward integration is predicted to result in one additional anti homeless law. In reality citie s are not perfectly segregated or integrated. The median and mean values of dissimilarity index are 134.9 and 133.9 respectively. For the median city, a 35.2 reduction in dissimilarity would mean a 26.9 % change This result suggests that those cit i es wher e segregation was the lowest had the highest rates of anti homeless laws, however this weak association will need more supporting evidence to make significant inferences from the result A city's proportion of white only population has a negative statis tically significant relationship with the number of anti homeless laws in a city. The correlation coefficient ( 4.7586) indicates a strong relationship. Unfortunately there is a large amou nt of variance within these data Fortunately, the P test (0.011) an d the relatively small standard deviation (17.09%) suggest that it is very unlikely that outliers are responsible for the strength of the regression coefficient. This model predicts that a city with a 1% proportional decline in the white population would c orrespond with a 4.75 additional anti homeless laws if all other variables were held constant. Second only to "percent change in voti ng gap, " p ercent White only may be the best predictor in this model. T he

PAGE 84

! 75 results indicate that cit ies where w hite only p opulations were lower tended to have higher anti homeless laws. Table 8 : Culture Shift Results Summary Statistics Independent Variable Mean Median & (Range) Standard Deviation Average Political Voting Gap (+ = Democrat; 2000 2012) + 11.31% + 9.98% ( 43% to 80%) 25.07% Percent Change in Political Voting Gap (+ = Democrat; 2000 2012) + 8.37% + 9.18% ( 14% to 47%) 9.85% Average Black/ White + Average Hispanic/ White + Average Asian/ White Dissimilarity Index (1990 2000 2010) 133.90 134.93 (75 to 194) 24. 63 Percent of Population White only (2011) 61.86% 63.54% (11% 95%) 17.09% Percent Change in Median Income (2000 2010) + 23.23% + 22.84 % ( 6% to + 54%) 9.76% O verall this model is statistically significant and accounts for a sizable portion of varian ce in anti homeless law rates within the sample of 102 cities Above I presented results and analysis for each individual independent variable as it relates to the dependent variable. Of the ten variables that proved to hold significant relationships with anti homeless laws, five held weak to moderate relationships (average homeless population, percent change in homeless population, growth of aggregate housing value, growth of HT LQ, and total average dissimilarity index), two held moderate effects (percen t change in violent crime, and percent change in median owner/occupied housing value), while three variables held the strongest association (percent cha nge in property crime, percent w hite only and percent change in voting gap ) Appendix A contains a n addi tional summary of results for all independent variable tests Developing meaning and theory from these new results requires an examination of results as they relate to one another and to theory. This conversation will inform

PAGE 85

! 76 conclusion s on whether hypothe ses are supported or questioned. The next section will contain this discussion Unfortunately, this regression model only accounts for thirty eight percent of the total variance in the model. Little inference can be learned from weak to moderate associati ons, as they are arbitrary terms of comparison between variables with stronger effect s The discussion that follows will focus on the five strongest associations, as weak relationships are less inf luential as we seek to use this model 's results to understa nd the reality of growth in anti homeless laws

PAGE 86

! 77 CHAPT ER V DISCUSSION The threat h ypothesis states that rising crime and homeless rates will share positive relationships with high levels of anti homeless laws across U.S. cities The theory that homeless rates, cr ime rates and anti homeless law levels are associated is founded in literature and public discourse. Wilson and Kelling (1982) connect the homeless person to crime rates in their broken windows theory, stating, "serious street crime flourishes in areas in which disorderly behavior goes unchecked. The unchecked panhandler is, in effect, the first broken window" (p. 4) Siegel (1997 ) also makes the connection between the sight of homeless persons and the perception of crime stating that "It was t he street tax' paid to drunk and drug ridden panhandlers It was the threats and hostile gestures of the mentally ill making their homes in the parks (Siegel, 1997 p. 169) The perception of danger is theorized as fused with the presence of homeless pe r sons who are suspected in the Broken W indows theory to be provocateurs of crime. If this theory w h ich has been employed by mayors such as Mayor Giuliani of NY, is a valid theory then there must be a relationship between the perception of crime (homeless people and other broken windows ) and actual crime; both of which are expected to drive a municipality to pass anti homeless laws. Results show that there are relationships between crime, homel essness, and anti homeless laws however in strengths and di r ections that greatly challenge the validity of the threat hypothesis as stated

PAGE 87

! 78 First, I will discuss the relationship between homelessness and anti homeless laws as seen thr ough t he results of this study before moving on to examine the relationship betwe en crime and anti homeless laws Measures of homelessness in this study were not found to relate to anti homeless laws as expected. While a verage homeless populations did share some positive relationship with the level of anti homeless laws, a n extremely weak effect was found The only inference that can be made by this result is that there needs to exist some base homeless population in a city for t here to be anti homeless laws. In other words, without a base level of homelessness existing in a city a ci ty would have no use for and feel no need to implement anti homeless laws. In addition, there would be little possibility of anxiety or fear created from homeless populations that exist in very low numbers The weakness of the relationship between size o f the homeless population and number of anti homeless laws indicates that beyond the possibility that there needs to be at least a threshold number of homeless persons to inspire the passage of anti homeless laws T he a verage n umbers of homeless in a city have little effect on the number of laws prohibi ting homeless behavior First t he variance in homeless populations across cities in the sample varies drastically (194 to 62,850) with a large standard deviation, while the regression effect indicates a ve ry weak relationship between these variances and the number of anti homeless laws Th ese two realities indicate, a ccording to this model, that a city would have to have its homeless populations increase by thousands in order to attribute any additional la w implementation to the homeless population rate if all other variables were held constant

PAGE 88

! 79 An increase of 5,000 homeless individuals is not without precedent ; however the diminishing effect of the homeless levels on laws, the necessity of all other vari ables being held constant for an effect to emerge, and the fact that the vast majority of cities in the sample had l ess than 5,000 homeless people supports the interpretation of a weak result. Average homeless populations ranged from 194 62,850 with a mea n of 3,511 and a median of 1,500. In truth this fact means that the relationship is too weak to make inferences There may be some relationship that is not being properly examined by this model but within this model we must dismiss this variable as having little importance in explaining the growth of anti homeless laws. The second measure used to test an expected relationship between homelessness and anti homeless laws was the growth rate of homeless populations from 2005 to 2012. The threat hypothesis pr edicted that anti homeless laws would, in some way, be connected to cities with increasing homeless populations In other words, anti homeless laws are understood to be a response to an actual increase in homeless ness Evidence suggests that the reverse it s true A strong negative relationship exists between homeless ness rate changes and anti homeless laws. Cities W here homeless popul ations fell the fastest tended to have a higher number of anti homeless laws. To reiterate the conclusions so far: anti hom eless laws are only related to average homeless populations in so far as there must be a large enough population for homeless debate to form in the public and p rivate discussion. Also, A city's homeless rate decrease is related to high levels of anti homel ess laws. Thus, the threat hypothesis is disproven insofar as it relates to homelessness. Relationships do exist between changes in

PAGE 89

! 80 homelessness numbers and levels however not in the way theory or the threat hypothesis predicted. Theoretical question s concerning Broken W indows theory and theories of threat emerge from these first two results. For example, why do cities homeless populations decrease ? One could attribute this to policy. Meaning that effective homeless ness reduction results may accompan y the implementation of anti homeless laws; just as promised by cit y leaders prior to the implementation of new anti homeless laws, as in the case of Denver Colorado and many other cities (Whelley, 2013 ). A decline in housing cost, as Quigley, Raphael, & S molensky (2001) predicts could decrease h omelessness to some degree. Yet for this to be true housing co st decline would need to be proven to decrease homeless rates faster than unemployment statistics as unemployment grew rapidly as a result of the reces sion. Regardless, I must seriously question the validity of the threat hypothesis. More discussion pertaining to the decline of homelessness will appear later in this section C rime rate results offer some def ense for the threat hypothesis, however, a vera ge crime rates for both violent and property crime s share no association with anti homeless laws. This fact indicates that the level of crime as well as the level of homelessness have little to do with a political effort to criminalize homeless survival be haviors. This result challenges established theory of threat which holds that anti homeless laws are inspired by fear of crime Furthermore c hanges in violent crime rates are negatively associated with high number of laws. Cities where violent crime fel l fastest have the highest number of anti homeless laws. This contradicts the theory of threat because it predict s that anti homeless

PAGE 90

! 81 laws are associated with falling crime as well as falling homeless rates unless there is a reverse causation present. An argument supporting the presence of a reverse causal effect is not without merit in so far as change in violent crime rates and homeless rate change The Broken Windows theory would definitely support this assertion. However, the results of this study wi ll not be able to prove or disprove the presence of a reverse causal effect. With that said, l ooking at the Pearson's C orrelation results tells us that homeless rates and crime rates are uncorrelated with one exception; average property crime rates have a negative corre lation with average homeless populations These simple correlation results suggest that in many ways homelessness and crime are not as linked as some have predicted and discourse assumes even if they co vary within the regression and the s imple correlation. Therefore, it would be hasty to draw support for a reverse causal effect through a logical and theoretical connection between homeless rates and crime rates that need to be reexamined. In addition, b ecause percent change in property cri me has a positive regression affect on anti ho meless laws as well as a strong positive simple correlation with violent crimes rates the presence of a reverse causal effect may also imply that anti homeless laws drive up property crime. E i ther way Broken Wi n dows theory needs to be re examined as it applies to anti homeless laws. Do anti homeless laws decrease violent crime rates and increase property crime rates across an eleven year span and if so, how does theory justify the contradictory nature of the se results ? Again, the discussion of the Pearson's Correlation is not meant to question the regression results, but to interpret the

PAGE 91

! 82 relationships between the independent variables in order to create theoretical understanding Concerning falling homeless ra tes, some have argued that anti homeless laws are not effe ctive in reducing homelessness (NHCLP, 2011; USICH, 2012) Evidence supporting the effectiveness of Broken W indows policing is also mixed (Hinkle, 2014) In addition little evidence suggest s that b roken windows policing has an effect on v iolent crime, but there is evidence of its effect in reducing property crime (Corman and Mocan, 2005) With that said there is evidence to suggest that the effect on property crime is also mixed. Arguing these two points, Corman and Mocan (2005) write, It is important to emphasize that the arrests for felonies have the largest effect on felony crimes and that the effect of broken windows policing, although significant for some crimes, are not universally significan t, nor are they of grea t magnitude. To put the broken windows hypothesis in perspective, note that other cities also experienced significant decreases in crime during the 1990's, without the dramatic increase in misdemeanor arrests (pp.262 263). The minim al effect on violent crimes of employing broken windows policing suggests that anti homeless laws are not causing a decrease in violent crime. However, more research is needed. Percent change in political gap also holds a significant (.034), but moderate ( .20956) correlation with change s in violent crime according to the simple correlation results This means that there is a slight connection between cities with falling violent crime rates and cities that are becoming Democratic. This demonstrates co vari ance and challenges arguments toward reverse causation because anti homeless laws may be affected by the same shifting demographics that affect anti homeless laws rather than anti homeless laws driving the changes in crime rates

PAGE 92

! 83 There is also evidence in revanchist literature that anti homeless laws are lagging indicators of previous violence, homelessness, and disorder. Pruijt (2012) cites Smith (1996) and Peck and Tickell (2002) arguing that revanchist policies, such as anti homeless laws are prece ded by a retreat of the state welfare services and are a reaction to "Liberal" policies of the 1960's. In this vein, it could be that that current political wave of anti homeless law s is a function of past policy and urban environments ( the liberal policy of the 1960's and corresponding urban decay) which has provoked a recent wave of revanchist anti homeless laws Pruijt (2012) indicates that the anti squatting legislation and the conception that squatters are dangerous in the Netherlands took place thi rty years after violence occurr ed between police and squatters and fifteen years after the peak of squatting occurred. Pruijt (2012) writes, "the time lag between the collective memory and 2010 anti squatting legislation supports the interpretation that a n element of revenge is present" (p.1117) Similarly, currently falling homeless and crime rates amid a growth in anti homeless laws may also indicate an element of revenge for past urban wrongs (which is the heart of much of Smith's argument on current revanchist politics) All this being said, a reverse causal effect remains a real possibility and warrants further study. There exists enough supporting evidence in the l iterature and in the significance of this model to continue wit h this discussion w itho ut further investigation in to a reverse causation The rate of property crime growth is positively correlated with the number of anti homeless laws. A s discussed in the results section, the positive relationship between property crime growth and high leve ls of anti homeless law is more appropriately described as follows: those cities where property crime rates dropped the slowest tend to

PAGE 93

! 84 have the highest number of anti homeless laws. This result generates two additional questions to consider First, how mi ght property crime rat e s be a better predictor of anti homeless laws than violent crime and homelessness? S econd what circumstance would create an environment where homelessness and violent crime rapidly fell while property crime fell slowly or slightly i ncreased ? Nathan Glazer (1979) created a hypothesis concerning the perception of crime as related to the sight of disorder. His 1979 piece titled On Subway Graffiti in New York blamed the sight of graffiti not for causing crime per se, but for causing th e perception of crime to increase This perception of crime is derived from one s insecurity and inability to control his or her surroundings. Glazer (1979) writes; He (the subway rider) is assaulted continuously, not only by evidence that every subway car has been vandalized, but by the inescapable knowledge that the environment he must endure for an hour or more a day is uncontrolled and uncontrollable, and that anyone can invade it to do whatever damage and mischief the mind suggests (p4). Glazier (197 9) continues explaining that once an association with crime and the sight of disorder is made the next logical question is how can it be controlled ? He states, If the linkage is a common one, then the issue of controlling graffiti is not only one of prot ecting private property, reducing the damage of defacement, and maintaining th e maps and signs in the subway but it is also one of reducing the ever present sense of fear, of making the subway appear a less dangerous and unpleasant place to the possible u ser. And so one asks: wh y can't graffiti be controlled? (Glazer 1979, p.5). Glazer (1979) did not devote much of this thesis to the sight of homeless people, however Wilson and Kelling (1982) made the connection three years later. In order to better mai ntain public order, Glazer (1979) recommends an expansion of juvenile detention, apprehension and reform to manage public space. In much the same way, Wilson and Kelling (1982) recommend managing public property through imposing

PAGE 94

! 85 restrictions on public b ehavior of the homeless as they are equated with the disorderly graffiti artists of Glazer s (1979) hypothesis. If Glazer s (1979) hypothesis is true property crime might be a more important predictor of anti homeless law rates than violent crime. More importantly it may explain why cities that saw the smallest improvement in property crime rates might be seeking ways to deal with the perception of disorder. Conversely, Hipp (2013) foun d that violent crime rates were the strongest predictor of crime p erception from 1970 to where his study ends in 1999 A nd during that same time period he found that property crime had little association with the perception of crime. Hipp (2013) also indicates that prior to 1970 b urglary (property crime) w as most respo nsible for shaping the public perception of crime. Perhaps, because violent crime has fallen drastically around the country since the 1980's ( Levitt and Dubner, 2006 ) property crime rate change has once again become a more reliable predictor of the publi c perception of crime. Property crime rates are higher than violent crime rates in every city in the sample. Property crime is more visible and is witnessed by more people on a day to day basis. It is logical to think that property crime drives public conc eption of safety, however that question was not addressed by this study. The results of this model indicate that property crime is a twice more effective predictor of anti hom eless laws tha n violent crime and perhaps that is due to dynamics like these There is the real possibility that additional anti homeless law s increase property crime by virtue of anti homeless laws being property offences for which people can b e arrested However, Broken Windows theory suggests that all crime will fall after Broke n

PAGE 95

! 86 Windows policing is adopted T he long eleven year property crime rate calculations and anti homeless laws are not necessarily a comprehensive test for broken windows policing, but as Corman and Mocan (2005) indicate, may be part of a broader trend that i s influenced n ot only by Broken W indows policing strategies but by other factors as well. Therefore, these results most likely represent a political reaction to property crime more so than other threat measures The threat hypothesis, for the most part, can be abandoned. While there are no relationships between the absolute level of cr ime and the number of laws, a strong relationship exist s between changes in property crime rates and anti homeless laws. At the same time these results contradict theory in many ways. Rather than a positive relationship between homeless growth rates and violent crime growth rates, there is a moderate to weak negative relationship. Homelessness, for the most part, can be dismissed as an unimportant variable especially consid ering the amount of variance remaining unexplained. Thus a new understanding of threat as it relates to anti homeless laws reads as f o llows : anti homeless laws are high in those locations where property crime statistics are unchanging slowly decreasing o r slightly increasing I t will be important to discuss the other two hypotheses to see if the clarified threat hypothesis is supported by other findings as well as by the literature. The neoliberal growth results, such as the threat results, challenge the stated hypothesis. T he n eoliberal growth hypothesis predicts that e conomic factors associated with neoliberal growth patterns will share a strong positive relationship with high levels of anti homeless laws across U.S. cities A city's number of anti home less laws was expe cted to hold a positive relati onship with all growth measures The neoliberal

PAGE 96

! 87 h ypothesis suggests that because cities are global economic centers each city must compete on a global scale making polic y choices and instituting priorities d esigned to increase property values, attract global business, tourism, investment, etc. Quigley and Taylor (2004) and others (Mitchell 2003; Wright 1997) argue that this shift in priorities translates to negative consequences on those living in public pla ces Cities seeking to grow at all costs inevitably run out of space and are forced to push the poor out to make way for new development and investment. I hypothesized that such neoliberal growth dynamics would have the most significant relationship with a nti homeless laws. T he results of the regression model, however, suggest that the hypothesis of neoliberal growth as a predictor of anti homeless laws needs significant refinement. High tech sector growth was a partic ular variable that proved to have a r elationship meeting some level of expectations Florida (2012) argues that technology sectors are one of three main growth sectors in a city that indicate a creative and healthy growing city. The Milliken Institute (2013 ) ranks cities by a variety of indic ators, many of which are tech growth indica tors. One of these indicators, location quotient of high tech sector g rowth (HT LQ) holds a moderate positive relationship with anti homeless laws. This result indicates that over the last four years, m etropolita n areas that have seen the most growth of high tech sector concentrations relative to all other metro areas tend to have the high er number s of an ti homeless laws. This result supports the neoliberal hypothesis because technology is one of the most importa nt growth sectors for which cities around the world are competing It is also the only growth measure that correlated as expected.

PAGE 97

! 88 Unfortunately, the regression effect for HT LQ is moderate. Min imal support for the neoliberal hypothesis can be given in t hat this result does not indicate that the potential neoliberal causes of anti homeless laws are dominant. On the other hand, this result does indicate that growth measures are an important part of explaining anti homeless law growth. HT LQ is only one mea sure of many variables that are related to growth. A ccording to this model s predicted effects, HT LQ most likely has less effect on anti homeless laws than changing property crime rates and is similar to the effect of changing violent crime rates Housin g values did not associate as expected Anti homeless law levels had a negative relationship with two separate measures of property va lues. Immediately this pattern calls in to question the validity of the neoliberal hypothesis as stated and demands theoret ical explanations for how a nti homeless laws can have a positive correlation with one growth measure and negative correlation with another First, I need to make a point of clarification. Even though there is a negative relationship these results do not n ecessarily indicate that a decline in housing prices would correspond with an increase in anti homeless laws. T he negative relationship in this case is the result of the fact that in locations where housing prices gre w more slowly over the past decade (h ousing prices didn't actually decline in the cities measured) anti homeless laws were more prevalent Housing value changes negative association with anti homeless laws challenge the neoliberal hypothesis because in reality the result was almost the opposite of what was hypothesized. This result indicates that anti homeless laws are not positively related to all growth measures but only some Perhaps high tech and other globally competitive industries such as tourism or investment measures, such as foreign direct investment,

PAGE 98

! 89 may have a positive association with the level of anti homeless laws but general property values have the opposite effect This point borders on conjecture because no data on tourism or foreign direct investment were tabulated f or cit i es. Furthermore, t he negative effect of the percent chang e in median owner occupied housing values on anti homeless laws was one of the five strongest relationships and we must use this result in the explanation of how growth may or may not affect anti homeless laws. An additional explanation would be that the housing crisis and recession could be influencing housing value data While this may be true, there are no results of this study that directly links the recession and / or recessions to the pro duction of anti homeless laws. The results indicate that anti homeless laws, in part, are a reaction to slowing housing growth. Looking deeper into the connection between slow growth rates and anti homeless laws will require an understanding of how citie s grew over the past decade Frey (2012) analyzed population growth dat a from 1980 to 2010 and found "w hile 61 of the nation s 100 largest metro areas grew faster in the 1990's than during the 1980s, 69 grew slower in the 2000s than in the 1990s. Southern and Western metro areas still grew fastest in the 2000s, but exhibited the greatest growth slow downs from the prior decade" (p. 1). Frey (2012) continues stating "Growth slowed considerably during the latter part of the 2000s, especially in bub ble econo my metropolitan areas" (p.1). Frey (2012) indicates that the fastest growing cities saw the sharpest drop in growth rates What is it about slowing growth that could be promoting the implementation of anti homeless laws? Frey 's (2012) arguments that grow th slowed in the later part of the decade most significantly in those places most affected by the housing crisis and that those locations that grew the fastest in the 199 0s slowed in the 2000s provide two lines

PAGE 99

! 90 of reasoning to explore First, cultural per ceptions of homeless people may be changed by economic insecurity created by the housing market that had the slowest growth rates over the past decade and the largest property value losses during the housing crisis may have dri ve n some cit ies to implement anti homeless laws. Second, t hose cities where hou sing markets grew the slowest and possibly were hit the hardest by the recession may have pursue d revanchist policies to address growing economic insecurity. The first explanation originates from the w ork of Mitchell (2011), Ryan (1976), Gans (1995), Barak (1991), and Kawash (1998) who argue that the very presence of homelessness instills economic anxiety of fear and loss in those of the middle class. This explanation has many problems in making sense of the data in this study's model however First, this explanation does not explain why the perceived threat of homeless people increases as homelessness decreases. This reality only makes sense if fear is generated from the experience of property crime and not homelessness, and homelessness takes the role of a political scapegoat to attack property crime. Second it does not explain why tech growth would be associated with anti homeless laws When high tech industry grows in a city would that not decr ease economic insecurity? In addition previous research has shown that economic growth rates (or declining economic growth) actually have little to no relationship on public attitudes towards the poor (Kam and Nam, 2008; Hogan, Chiricos and Getz, 2005; Mu ghan, 2007 ; Kluegel, 1987). I reject the first explanation of perceptions of economic insecurity driving anti homeless laws through changing perception s of the homeless ness b ecause it does not explain the correlation of high tech growth rates with more a nti homeless laws, n or is

PAGE 100

! 91 there sufficient evidence that attitudes toward the poor are affected by housing statistics or recessions Economic insecurity created by slowing growth rates could be contributing to a broader culture of fear, but more informatio n is needed to solidify this conclusion. The second explanation theorizes that those cities where economic insecurity has increased the most are the same areas where revanchist policies, like anti homeless legislation have been pushed most successfully. Nichols et al. (2011) used a unique measure to quantify insecurity. Nichols and other researchers from the U rban Institute and Rockefeller Foundation created a measure based on actual job variability, wage variability and, other economic data. Nichols et a l. (2011) write : ESI (Economic Security Index) is a measure of the share of Americans who experience large income losses. More specifically, it tracks the proportion of Americans who see their "available household income" their household income after payi ng for medical care and servicing their financial debts decline by 25 percent or more from one year to the next and who lack an adequate financial safety net to replace this lost income (p. 1). Nichols et al. (2011) indicate that the recent recession crea ted a sharp increase in the economic insecurity index, but that it was already rising prior to the economic crisis. Combining Frey (2012) and Nichols et al.'s (2011) findings suggest s that the largest increase in economic insecurity happened in many of the same locations that Frey (2012) argues saw slow ing population growth in the 2000's. The literature connecting punitiveness and economic insecurity is important to note. While Johnson (2001) indicates that individual economic insecurity as measured by pa s t events is only weakly releva nt to punitive public attitudes. Johnson (2001) and Hogan, Chiricos, and Gertz (2007) find that negative measures of insecurity of future economic possibility are connected to punitive attitudes and resentment toward welf are recipients. Hogan, Chiricos, and Gertz (2007) write,

PAGE 101

! 92 It would appear that punitiveness toward criminals is, in fact part of a general constellation of resentment toward, a scapegoating of, what Gans has termed the "undeserving poor." This particular c onflation is consistent with what Garland (2001) has described as the "politics of reaction." The latter, is argued, has crystallized in support for "more aggressive controls for an underclass' that was perceived to be disorderly, drug prone and dangerou s" and resentment toward "policies that appeared to be nefit the undeserving poor'" (G arland, 2001, p.148). Our measure of punitive attitudes taps int o "more aggressive controls," and our measure of blame gives expression to "resentme nt" of the undeservin g poor" (p. 405) Both explanations are highly conjectural because not enough information is available to connect the results of Nichols et al. 's (2011) measure of economic insecurity to anti homeless laws and there are some generalizability concerns for Hogan, Chiricos, and Gertz 's (2007) and Johnson 's (2001) studies on punitiveness and economic insecurity However, a ggregate property value also shared a negative relationship with anti homeless laws. This association reaffirms that those cities where prop erty values rose the slowest over the past decade tended to have higher numbers of anti homeless laws and thus supports a new hypothesized association between economic insecurity driven by stagnant property markets with the growth of anti homeless laws The negative relationship between housing values challenges the neoliberal hypothesis in some ways. Rather than anti homeless laws being a result of a well thought out economic policy that expands as growth expands there seems to be a contemporary cultura l reaction to rising or stagnant property crime rates and slow housing value growth. On one ha n d the fact that homelessness and violent crime rate declines have been shown to positively effect the number of anti homeless laws may indicate that a Pruijt's (2012) and Smith's (1996) element of "revenge" for historic wrongs is a factor more than current rates of homelessness or c r ime The presence of revenge supports the neoliberal economic strategy argument.

PAGE 102

! 93 On the other hand, Hogan, Chiricos, and Gertz (200 7) indicate that the current economic environment can influence current punitive attitudes toward the poor This means that it is more likely that anti homeless laws are a res ponse to contemporary trends than a revengeful commitment to take back the city f rom the poor. These results call into quest ion Mitchell's (1997; 2011; 2003) assessments of the genocidal politics' of cleansing the street of the homeless. Mitchell s view overlooks a more localized and more complicated interaction between culture, fear, current economic trends, urban environments and tolerance for the other' (Johnsen and Fitzpatrick, 2010). Neoliberal growth, as tested in this model, is not the dominating influential factor contrary to expectations. Theoretical explanations for this r esult conflict and do not present clarity. Only tech sector growth is positively correlated with high anti homeless laws. Yet t he economic growth hypothesis cannot be completely abandoned. There are many key variables that have not been measured such as f oreign direct investment, tourism, and scale of redevelo pment across cities that fit in to the neoliberal growth hypothesis. However, according to this model, neoliberal growth is less explanatory than threat and culture shift theories. In the next subsect ion I will continue an at tempt to fill in theoretical gap s that have been generated through the discussion and revision of the threat and neoliberal hypotheses by focusing on the most important findings within the culture shift test The final hypothesis t ested in this study was the culture shift hypothesis. As stated this hypothesis predicts demographic factors indicating political and cultural shifts will hold positive associations with anti homeless laws The results of this section are the most reveali ng. Measures of cultural shifts and gentrification were expected to

PAGE 103

! 94 share positive relationships with the number of anti homeless laws echoing th e results of Wyly & Hammel (2005 ). I hypothesized that that political party strength (as measured by the local voting patterns of the electorate) would hold some small rel ationship with Republican Party dominance and anti homeless laws As with each test before, the results of this study are no t what were expected. Culture s hift, more than any other hypothesis, se ems to have the greatest relationship with anti homeless laws, but in ways counter to the stated expectations. Wyly & Hammel (2005 ) tested a number of gentrification measures including median income an d dissimilarity statistics for B lack, Asian and Hisp anic separately. For this study I chose to add percentage of the population that was white only as a measure of diversity (rather than breaking out the racial groups separately) and to combine a dissimilarity index for all three decades and for three eth no racial groups Combining dissimilarity index measures, I anticipated, would give a broader measure of the influence of racial diversity and segregated populations than that which was found in Wyly & Hammel 's (2005 ) research. Median income was an import ant measure for this study and for Wy ly and Hammel (2005 ). It was important because median income statistics provide evidence of gentrification beyond race. Measures of housing and property value were expected to co vary with anti homeless law, however th at is not the result that was found. For median income there was no relationship found. Percent change in political voting gap is the most important finding in this study. It is the most correlated measure in the sample according to the Pearson's Correl ation

PAGE 104

! 95 and is the variable with the most effect on anti homeless laws according the regression model results. This resul t does not indicate that cities that are more Democratic pass more laws. No relationship was found between party strength and anti home less laws. The results indicate changing political ideologies within a city (toward the Democrats) lead that city to implement anti homeless laws. The changing political ideologies could be a result of ne w populations moving to cities old residents shifti ng from Republican and Independent to Democratic, or a combination. Regardless, this finding indicates a significant change in political culture is the most important predictor of anti homeless laws A city's w hite only population percentage also shares on e of the strong est effects on anti homeless laws according to the model. Its significant negative relationship with the nu mber of anti homeless laws predicts that cities that are less w hite will tend to have higher laws. What this measure does not indicat e is whether anti homeless laws are related to an i nflux of new non white residents or a result of the faster growth of long standing non white population s Like Political party, this result is very surprising because it is the direct opposite of what was hypothesized Dissimilarity index results share some similarities with percent White only A negative relationship with dissimilarity a segregation measure, indicates that those cities that are less segregated tend to have higher numbers of anti homeless laws. However the effect on anti homeless laws is weak and only limited inferences can be drawn. We must refocus on the two strongest results, percent non white percent change in political voting gap and their relationship with other significant variab les.

PAGE 105

! 96 Political shifts and non white populations have an im portant relationship to discuss. Democrats still maintain a large advanta ge over Republicans with regard to minority voting statistics in presidential politics. For the study sample, average politic al gap held a moderate to strong ( .46293) correlation with percent w hite only ( P test = .0000) The s e data supports the long known pattern that non white voters tend to vote Democrat. Unfortunately the results do not indicate that the increase in anti homeless laws are affected by an increase in non white Democrats moving to the city, but there is evidence to suggest that cities across the country are becoming less white during a decade when anti homeless laws are increasing. On average, the cities in t his sample show a small decline in total average dissimilarity indexes between 1990 and 2010 ( 4.73). Put into context with the rapidly diversifying urban areas around the country, the dissimilarity indexes show that the integration of cities is not taking place as fast as cities are becoming non white. Meaning that in spite of a natural pressure to integrate created by the volume of minorities movi ng to and being born in cit i es, w hite dominated neighborhoods are likely persistent and growing as well. In t his way, gentrification still has a n important part to play in the understanding of anti homeless laws. As hypothesized by Wyley and Hammel (2005), perhaps the way in which the forces of globalization causing gentrification affect cities is different for e ach neighborhood and each ethnic group within that city. Lee, Iceland, Sharp (2012) found in a three decade long study, similar to Frey (2011) that almost all areas in the country diversified more rapidly in the last ten years than ever before. Many citi es across the country are becoming les s and less w hite. While

PAGE 106

! 97 this does not prove causality, it fits nicely with the trends toward Democratic politics and non white population that have a demonstrated effect on anti homeless laws according to this model. H owever this point describes the pattern but doesn't indicate why larger populati ons of non white residents and swiftly growing Democratic voting patters would precede the passing of anti homeless laws. Magnet (2007) hypothesizes that cities where traditio nal civil rights politics and political organizations exist have managed to fend off th e continuous implementation of Broken W indows policing. However, those locations without established civil rights political machines have little defense against such l aws and in fact in those locations minorities become some of the advocates for Broken Windows policing Magnet (2007) writes, How, then, in Denver, did a bunch of liberal Democrats successfully institute New York style policing reform, with overwhelming support not just from homeowners and the business community, but from poor minorities who live in blighted neighborhoods? For one thing, as Nevel observes, "This city doesn't have as sharp elbowed grievance groups" as New York's, which threaten violence (a s city councilman Charles Barron did after the Bell shooting) and throw feces at effigies of the mayor (as the hatchet job film Giuliani Time lovingly documents). With three times as many Hispanics as African Americans in Denver, many of the city's activi st groups that speak or claim to speak for minorities lack the loathing for and suspicion of the police that simmer in many of today's black partisans... Policymakers found that, when vitriol or demagoguery did not drown out the voices of ordinary poor and minority residents, they said that they wanted safe neighborhoods, and for the city to use tactics that experience showed would work best to protect them from the tyranny of lawlessness. In Denver, that desire proved so strong that it erupted in grassroot s advocacy for New York style policing from the very people whom Al Sharpton and his ilk depend on to oppose it minorities and the poor. (p.1) Cities like Memphis Washington DC, New York, and Boston all have lower than averag e anti homeless law levels an d are home to significant number of civil rights and

PAGE 107

! 98 labor organizations. In contrast, cities like Denver, Colorado Springs, Las Angeles, Portland, and Seattle have higher than average anti homeless law levels. While much of Magnet's (2007) argument is co njecture, Bonilla Silva (2004) has a similar explanation of how civil rights politics can be supplanted by policies like Broken W indows policing. Bonilla Silva (2004) predicts that the power of civil rights politics will dwindle within the minority and majority communities as more and more minorities move up the social ladder and as growing ethno racial groups lobby for issues important to their particular communities rather than minorities as a whole. Bonilla Silva (2004) and Magnet (2007) offer some e xplanation of how growing diversity would correspond with growing Broken Windows policing, but most points are highly conjectural. In order to understand how non white and growing Democratic voting patterns better we must con sider broader patterns in the d ata. The relationship found between a culture shift and anti homeless laws follows a trend that is consistent across all measures. Those measures that hold significant relationships with the dependent variable are, in most cases, measures of change. Change in homeless rates, crime rates, home values and growth of the tech sector all follow this trend. Only average homeless population size (weak), average dissimilarity index (weak to moderate), and percent white only (strong) share relationships with the de pendent variable and are not measures of change. The most important variable in this study, percent change in political voting gap, supports the importance of shifting dynamics in cities that pas s anti homeles s laws. These realities support an argument th at the mere presence of change or disorder may provoke insecurity within a city. That insecurity created by a messy mixing of race and politics

PAGE 108

! 99 and associated changing tectonics within the city, may explain why cities are implementing laws designed to co ntrol public space. In other words, the perception of disorder which cities try to counter act with anti homeless laws, may be created by the changing and mixing of cultures within cities. The results of this study have flipped expectations on their head. Neoliberal growth measures were expected to have the highest effect on the number of homeless laws passed in a city cultural shifts associated with neoliberal growth the second highest, and crime and homeless statistics the weakest. What the literature and the results reveal about anti homeless law growth as it relates to a culture shift is that there may be a relationship between rates of urban change along a host a measures and the growth of anti homeless laws. Change provokes fears of instability, and laws to strictly regulate and control public space are the result This culture clash may be the true catalyst behind Prujt's (2012) finding of culture war s and moral panic driving anti homeless politics The culture clash may be reinforced by neolibe ral economics, not necessarily by revengeful policies targeting the poor but by the unintended cultural consequences of constant redevelopment and a changing urban landscape. If disorder pushes cities to criminalize homelessness then the constant physica l change that occurs in a growing urban environment may play its role While revanchist policy may also be employed to secure high tech sectors, more relevant are factors of growth that demonstrate change. The results indicate that changing demographic s r elated to race and political shifts within a constantly reshaping environment effect policy output. Yet this does not hold true for those cities that saw a rise in Republican voting statistics. This may be because

PAGE 109

! 100 only a small minority of cities within th e sample saw Republican gains and those gains were very small compared to the Democratic gains. The sample ranged from a fourteen percent gain in R epublican voting to a forty seven percent gain in Democratic voting patters, a mean of 8.37% growth of D emoc ratic voting patterns and a standard deviation of 9.85% c onfirming that the vast majority of the sample saw growth in Democratic voting patterns. This sample may be overlooking the possible effects of a large Republican gain that did not occur in this s ample and that may or may not have an important effect that is being overlooked. Put into the context of the modern city, changing political and racial demographics has some influence on the culture of fear that exists in the city. But at the same time th ose fears created by racial and political shits are reinforced by other measures that are not quick to change, such as property crime rates and property values. In the same way that culture clash and economic insecurity are reinforce d by each other the re sults also point to a presence of disorder measur ed in the stagnation or rise of property crime reinforcing a cultural feeling of fear and perception of disorder Table 9 contains a summary of conclusions interpreted from the results. Table: 9: Summarie s and Revised Hypotheses Threat 2.0 Property crime growth promotes the passing of anti homeless laws more than any other measure of threat. Neoliberal 2.0 E conomic insecurity resulting from stagnant housing markets contributes to a culture of fear, while some growth measures, like High Tech growth, predict anti homeless law growth. Culture Shift 2.0 Cultural anxiety is provoked by constant ly increasing minority populations, new urban professionals a growing Democratic party, and a rapidly changing urban landscape disrupt ing feelings of security. Combined with fear generated by property crime growth and stagnant housing markets a cultural threat emerges and pushes cities to implement laws to restrict behaviors in public space while grasping to control per ceptions of social order.

PAGE 110

! 101 Together these theories paint a picture of what types of cities implement anti homeless laws. On e implication is that an increase in minority presence may foster perception s of disorder that threatens established residents and new residents alike, as predicted by both the "threat" a nd the "culture shift" hypothese s Property crime and slowing economic growth may influence pe rspectives of security. New and or growing minority groups may be shifting the culture of such a city. Ni chols et al. (201 1) indicate in fact, that economic insecurity grew disproportionately among minority groups in recent years Hogan, Chiricos, and Gertz (2007) argue that Women and non white minorities share the strongest connection between economic insec urity (in future terms), punitive attitudes, and resentment of the underclass. In addition, Wright (1997) argues that the presence of non white populations creates a perception of disorder. His assertio n is supported by the effect non white populations hav e on anti homeless laws Wright (1997) eloquently mak es this connection borrowing fro m Goldberg (1993), when he writes: Degeneracyhas historically been associated with nonwh ites, the very poor, the mark o f the pathological other' (Goldberg 1993a, 55), a sign of disorder that must be contained. In the hypermodern city, stratified by racialized, gendered, and class orientated social physical spaces, within growing sets of polarized landscapes, t he paranoia of losing power assumes the image of becoming the other, to be avoided like the plague' (Goldberg 1996, 55). As Goldberg (1993b) points out, city social space is racialized space, which the racial poor were simultaneously rendered peripheral in terms of urban location and marginalized in terms of power' (Goldberg 1993, 188, quoted in W right 1997, p. 77). Hopkins (2011) tests a similar assertion. Ho pkins (2011 ) sought to test the extent to which the increase in diversity has a negative or positive affect on social welfare spending and criminal justice spending, among other categories. While he found that no association existed between diversity and social welfare spending, substantial evidence

PAGE 111

! 102 exists that diversity positively effects criminal justice spending. In other words, as a city becomes more dive rse that city spends more money on criminal justice This supports my new cultural shift hypothesis whereby diversity a ffects a cit y' s increasing desire to control the public space. Yet the new theory of change causing moral and cultural panic is li mite d in two key ways. First, mo st of the variance is unaccounted for and thus large generalizations cannot be made from the results. Second, some of the measures show that stagnation and slow growth, such as property value and property crime, affect the num ber of anti homeless laws in a city i ndicating that the stagnation or slow growth of some measures reinforce cultural panic or there is something else happening that might be explained if all variance was accounted for. Neoliberal growth dynamics may be associated with the rising sense of fear that accompanies growing diversity, in that modern growth and globalization is associated with rising economic insecurity, which feeds into inse curity over growing diversity. The new understanding of neoliberalism 's effects on anti homeless laws may indicate that cities that have stagnant housing growth rates, after experiencing extremely high growth rates in the previous decade, begin to slow in terms of housing growth and redevelopment. This period of slow growth m ay provoke power brokers and investors to implement revanchist policies to renew local growth patterns by focusing on high t ech i ndustries However, changes in terms of political and racial demographics do point toward a theory of change, bu t rather than saying all change causes moral panic it is cultural changes supported i n part by Ferrell (2001), Sennett (1970), Prujt (2012) Johnsen and

PAGE 112

! 103 Fitzpatrick (2010) and Van Eijk's (2010) arguments, that cause moral panic. The results of this study support Wrig ht (1997) and Goldberg's (1993) assertion that the perception of disorder grows in the presence of diversity and poverty. The data do show that cities with the most integrated communities, and high est non w hite populations (that is, the highest experiences of diversity) tend to have the most anti homeless laws. The findings of this study support the argument that perc eived disorder disconnected from homelessness may be driving political leaders to fear losing the core of their cit i es, as was the case in New York City and Denver Colorado where leaders blamed disorder and a culture war for driving the need for anti homeless laws This is the fear that may be producing anti homeless sentiment. In short, according to this study, the push to pas s anti homeles s legislation is based on perception s of cultural, criminal and economic threat s that reinforce each other.

PAGE 113

! 104 CHAPTER VII CONCLU S ION Dwayne Hudson, a self described homeless person, testified before Denver City council in opposition to the passing of the new camping ban ordinance on April 30 th 2012 Mr. Hudson eloquently stated, As I look at this diverse body...the po sition and power that Mayor Han cock and several members of this council have are a result of protest movements and civil rights movement s...It's ironic today that...the members that h a ve been empowered by those types of protests would try to sweep this horrible picture of people being homeless under the rug, (and) create a program by jailing that population. Let's Jail these people for the sake of some business community who is afraid in the same way that the white class was afr aid that if blacks moved here or if blacks moved there it would hurt their business. There is a problem...here and it's not go ing to be salved by trying to s weep thi s under the rug and criminalize something...before that park was there, urban camping was an American Heritage. And the Indians didn't say. Could you move your covered wagon becau se you are killing my landscape? They didn't say you can camp here but unc over your wagon.' And here we are, as soon as we get in the position of power...(we are) doing things that stop other people from doing what we did to get t here. (Denver City Council, April 30, 2012 ) Mr. Hudson touches on the similarities between racial struggles and homeless struggles in terms of perception and fear. He admits to believing that the purpose of this law is to address economic and cultural fear of the "other," rather than the expressed wish to help the homeless. Mr. Hudson's quote fits nic ely with the findings that anti homeless laws are more about other factors contributing to fear than homelessness itself If no other conclusion is acknowledged from this study, the fact that homeles sness has little to no effect on the level of laws that a re designed to control them is revealing. Regardless if one supports anti homeless laws for revanchist interests or from some idea that homeless people need to be corralled into service by the police, the results of this study point toward

PAGE 114

! 105 perception s of cultural, criminal, and economic fear, disconnected from homelessness, that pushes cities to pass anti homeless legislation. This study cannot explain all the variance of anti homeless laws within cities T here remains a great deal of local dynamics t hat go es unexamined in this study. Regional differences may change the strength or weakness of each finding. Unfortunately, only 3 7.9 % of the total variability was accounted for in this study. This study does not disprove or prove any theory completely. Ho wever, e ven if reverse causality were present for one or multiple variables, the small effects that are likely to result would not be significant enough to account for the majority of the variance that goes undiscovered or change the ultimate conclusions o f this research T he results are not what were expected F uture researchers studying anti homeless laws and why cities implement them should be leery of making firm assertions that crime, homeless populations, and revanchist neoliberalism are the cause of anti homeless laws The results of this study do not exclude theories, but point toward a synthesis of all theories T his study reveal s that there are some commonalities across the coun try that hel p explain why anti homeless la ws continue to grow. Frey ( 2011, 2012) and Lee, Iceland, Sharp (2012) anticipate that diversity and populations will continue to grow in urban and suburban areas changing the political and urban landscape s of this country. Globalization and neoliberal growth dynamics in cities are also expected to continue and will predictably destabilize the local landscape For such reasons, t here is no evidence to suggest that the rate in which cities are criminalizing homelessness will slow.

PAGE 115

! 106 The goal of this research was to develop a better u nderstanding of why cities are increasingly implementing laws that criminalize survival behaviors of those who do not live in housing. In the face of the UN, USICH and homeless advocate s across the country pleading with cities to address the systemic caus es of homelessness instead of punishing those without housing, cities continue to implement anti homeless laws. The goal of this thesis to understand such dynamics has been reached in part. However, a large percentage of the answer remains elusive The results of this study indicate that cultural shifts may have the most significant role to play in anti homeless law growth over the past ten years as cultural threat influences perceptions of crime Further research is needed to clarify several key questio ns and unexplained issues created by this study. First, further research is needed on how one s perception of crime influences policy outputs. Second, further research is needed to better measure particular parts of neoliberalism such as foreign direct in vestment and how it might influence local policing practices and anti homeless laws Neolib e ralism is very difficult to quantify and more research is needed to prove ( or disprove ) its role in driving the growth of anti homeless laws. Third, further rese arch is needed to explain why a decrease in homeless rates could correspond with an ant i homeless law increase. Although an attempt was made to explain this result, without further research there are concerns about the validity of that particular claim an d the presence of reverse causality Lastly, f urther research is needed to test the revised hypotheses ( Table 9 ). First the revised threat hypothesis predicts that the threat of homeless people is influenced by property crime and reinforce d by a n overall culture of fear growing in urban areas The neoliberal threat predicts that growing economic insecurity resulting from stagnant home

PAGE 116

! 107 values and increased globalization ; spec ifically volatile job markets support s the overall culture of fear T he new cultur e shift hypothesis predicts growing unease and disrupted feelings of security associated with the instability associated with constant ly increasing minority populations, new urban professionals and a rapidly changing urban landscape. Combined with fear gen erated by property crime and economic insecurity, cultural threat emerges and pushes cities to implement laws to restrict behaviors in public space while grasping for control and social order. It is important to note that the results of this study should n ot be over stated. While many relationships support a theory of change, many relationships are contradictory An additional limitation of this study pertains to the unexamined institutional and political variables. The institutions and political environmen ts and histories of cities have not been examined during this study. Further investigation of the impact of various plans to end homelessness produced by cities may be important in the future. For example, Denver Colorado explicitly states in their Ten Y ear Plan that the city will implement measures to restrict panhandling and loitering in particular places while the Chicago Illinois plan does not (Denver's Road Home, 2009 p.21 ; City of Chicago 2005) Two possible confounding issues went unexamined. The first was the possibility that reverse causality is present. Without further study there is no way to know for sure whether a reverse causal effect between violent crime, homeless rates, and anti homeless laws exist s A longitudinal comparative analysis w ou ld go a long way in answering this question as well as many others.

PAGE 117

! 108 The second possible confounding factor is the potential that anti homeless laws are a result of compassionate "tough love." Many public officials in Denver Colorado, as in London Engl and, cited beliefs that homeless people need municipal governments to require the modification of their behavior in an attempt to make them less deviant (Whelley, 2013; Johnsen and Fitzpatrick 2010). More quantitative and qualitative as sessments of public discourse are needed to differentiate between support for anti homeless laws originating from revanchist motivations and motivations originating from compassionate medicalized understandings of homelessness Today our nation and the cities within it are rapidly diversifying and urbanizing T h e United States and its localities need to have a conversation about how we are to learn to live in closer proximity to one another and amid the unease of disorder associated with new global economic dynamics Th is conversation needs to include the homeless, the poor the voices of the civil rights era that remain, the voices of the new generations, and voices of all races This conversation needs to allow people to release their fear of the "other" in order to e mbrace a more socially j ust and collaborative way forward There is little doubt throughout the scientific community and homeless advocacy community that homeless ness and the barriers to becoming housed are predominantly a result of factors beyond the con trol of the individual homeless person as evide nt by research conducted on mental illness, housing market competition substance abuse, employment insecurity, low wages and a lack of social capital personal networks, and related support systems I argue that as fear of the other increases within a city, the amount of social capital in that same city will dwind le Through experiencing the other ...the uncertain becomes fixed and known and thus less threatening But more than

PAGE 118

! 109 this, we lear n how to inte ract with otherness to secure a sense of wellbeing as we move through urban space (Bannister, Fyfe, and Kearns, 2006 p.932) It is our innate social connectedness tha t fear provokes us to overlook and it is that fear that might explain the rise in laws tha t criminalize the very existence of a class of people in our cities.

PAGE 119

! 110 REFERENCES Adolino, J., & Blake C (2011). Comparing Public Policies Issues and Choices in Industrialized Countries. 2cd Ed. Washington D.C.: CQ Press. American University (2013) 2000 election data: All states all counties. R etrieved September, 2013, from http://www.american.edu/spa/ccps/data sets.cfm Amster, R. (2003). Patterns of exclusion: Sanitizing space, criminalizing homelessness. Social Justice 30 (1 (91)), 195 221. Axe lson, L. J., & Dail, P. W. (1988). The changing character of homelessness in the United States. Family Relations 463 469. Bannister, J., Fyfe, N., & Kearns, A. (2006). Respectable or respectful?(In) civility and the city. Urban Studies 43 (5 6), 919 937. Barak, G. (1991). Gimme shelter: A social history of homelessness in contemporary America New York: Praeger. Boushey, G., & Luedtke, A. (2011). Immigrants across the US Federal Laboratory Explaining State Level Innovation in Immigration Policy. State P olitics & Policy Quarterly 11 (4), 390 414. Beckett, K., & Herbert, S. (2008). Dealing with disorder : Social control in the post industrial city. Theoretical Criminology 12 (1), 5 30. Brenner, N., & Theodore, N. (2002). Cities and the geographies of "act ually existing neoliberalism". Antipode 34 (3), 349 379. Corman, H. & Mocan, N. (2005). Carrots, sticks, and broken bindows Journal of Law and Economics 48 (1), 235 266. City of Denver: The Denver Commission to End Homelessness (DCEH). (2005) Ten year plan to end h omelessness: A report to the citizens of Denver. Retrieved February 14, 2012, from http://www.denversroadhome.org City of Chicago (2005 ). Implementation schedule for Chicago's 10 year plan to end h ome lessness: Getting housed, staying h oused. Retrieved February 23, 2014 from http://www.homebaseccc.org/PDFs/TenYearPlannng/NAEH%20Chicago%20Handou t.pdf Denver Mayor's Office (2012 April 24) Mayor Council meeting. Denver Colorado. Retrieved February 5, 2013, from http://denver.granicus.com/ViewP ublisher.php?view_id=16

PAGE 120

! 111 Denver City Council : Full Council (2012 a April 30 ), Denver City Council: Denver Urban Camping Ba n. Retrieved February 5, 2013, from http://denver.granicus.com/ViewPublisher.php?view_id=16 Denver City Council: Full Council (2012 b May 14 ), Denver City Council: Denver Urban Camping Ban Retrieved February 5, 2013, from http://denver.granicus.com/ViewPublisher.php?view_id=16 Denver City Council (2013, May 14). Denver City Council proceedings. Denver Colorado. Retrieved on February 5, 2013, from http://denver.granicus.com/ViewPublisher.php?view_id=21 Denver's Road Home. (2009 ) Ten year plan to end h omelessness. Retrieved on February 23, 2014 from http://www.denversroadhome.org/js/tiny_mce/plugins/filemanager/files/DRH_JulyUp date_vF _LR.pdf Florida, R. L. (2012). The Rise of the Creative Class: Revisited New York: Basic books. Ferrell, J. (2001). Tearing down the streets: Adventures in urban anarchy New York: Palgrave Macmillan. Finegold, K. (2004). Block grants: Historical overv iew and lessons learned. Retrieved on November 5, 2013 from http://www.urban.org/publications/310991.html Frey, W. H. (2011). Melting pot cities and suburbs: Racial and ethnic change in metro America in the 2000s Brookings Institution, Metropolitan Polic y Program. Frey, W. (2012). Population growth in metro America since 1980: Putting the volatile 2000s in perspective. Metropolitan Policy Program Papers. Washington DC: Brookings Institute Gans, H. J. (1995). The war against the poor: The underclass and antipoverty policy New York: Basic Books. Guardian News and Media Limited, Inc. (2013). Full US 2012 election county level results to download. Retrieved September, 2013, from http://www.theguardian.com/news/datablog/2012/nov/07/us 2012 election county results download Harcourt, B. E., & Ludwig, J. (2006). Broken windows: New evidence from New York City and a five city social experiment. The University of Chicago Law Review 73(1), 271 320. Harvey, D. (2012). Rebel cities: from the right to the city to the urban revolution New York: Verso Books.

PAGE 121

! 112 Henry, N.W., 2001 (last updated 2004) "R square and Standardization in Regression" Virginia Commonwealth University. Retrieved on January 1, 2014 from http://www.people.vcu.edu/~nhenry/Rsq.htm Hinkle J. C. ( 2014). Broken windows t hesis. In Bruinsma G & Weisburd D. (2014) Encycloped ia of criminology and criminal j ustice (pp. 213 223). New York: Springer Hogan, M. J., Chiricos, T., & Gertz, M. (2005). Economic insecurity, blame, and punitive attitudes. Just ice Quarterly 22 (3), 392 412. Hooks, G. & Lobao L. (2010). Space and Politics. I n: K.T. Leicht and J.C Jenkins (eds.), Handbook of politics: state and society in global perspective, handbooks of sociology and social r esearch (367 384) New York NY: Sprin ger Science + Business Media, LLC. Hopkins, D. J. (2011). The limited local impacts of ethnic and racial diversity. American Politics Research 39 (2), 344 379. Kahler, M. (2004). Economic security in an era of globalization: definition and provision. The Pacific Review 17 (4), 485 502. Kam, C. D., & Nam, Y. (2008). Reaching out or pulling back: Macroeconomic conditions and public support for social welfare spending. Political Behavior 30 (2), 223 258. Kawash, S. (1998). The homeless body. Public Culture 10 (2), 319 339. Kluegel, J. R. (1987). Macro economic problems, beliefs about the poor and attitudes toward welfare spending. Soc iety for the Study of Social Problems 34 82. Kriznik, B. (2011). Selling global Seoul: Competitive urban policy and symbol ic reconstruction of cities. Revija za sociologiju 41( 3 ) 291 313. Retrieved May 22, 2013 from hrcak.srce.hr/file/112207 Kelling, G. L., & Coles, C. (1996). Fixing broken windows: Restoring order and reducing crime in American cities. New York: Simon and Schuster Inc. Johnsen, S., & Fitzpatrick, S. (2010). Revanchist sanitisation or coercive care? The use of enforcement to combat begging, street drinking and rough sleeping in England. Urban Studies 47 (8), 1703 1723. Johnson, D. (2001). Punitive attitude s on crime: Economic insecurity, racial prejudice, or both?. Sociological Focus 34 (1), 33 54.

PAGE 122

! 113 Lee, B. A., Iceland, J., & Sharp, G. (2012). Communities over three decades. r etrieved on January 1, 2014 from http://diversity.ucsc.edu/news events/news/images /US2010.pdf. Levitt, S. D., & Dubner, S. J. (2006). Freakonomics Rev ised Ed ition LP: A rogue economist explores the hidden side of e verything HarperCollins. Lyon Callo, V. (2000). Medicalizing homelessness: The production of s elf blame and s elf governin g within homeless s helters. Medical Anthropology Quarterly 14 (3), 328 345. Meyer, J. (2011, October 21). Denver Mayor Michael Hancock supports effort to outlaw sleeping on 16 th street mal. Denver Post Retrieved May 22, 2013, from www.denverpost.com/reco mmended/ci_19165276. Meyer, J. (2013, August 25). Denver hopes to plant seeds of c hange, transform "Triangle Park" area. Denver Post. Retrieved March 26, 2014, from http://www.denverpost.com/politics/ci_23935631/denver hopes plant seeds change transform m aestas park Milken Institute (2013). Best performing cities: Large cities 2008 2013. Retrieved November 1, 2013, from http://best cities.org/bestcities.taf?rankyear=2013&type=large cities rankings Mitchell, D. (1997). The annihilation of space by law: T h e roots and implications of anti homeless laws in the United States. Antipode 29 (3), 303 335. Mitchell, D. (1998). Anti ho meless laws and public space: Begging and the First Amendment. Urban Geography 19 (1), 6 11. Mitchell, D. (2003). The right to the city: Social j ustice and the fight for public space New York, NY: The Gulford Press. Mitchell, Don (2011). Homelessness, American s tyle. Urban Geography 32(7) 933 957 Mughan, A. (2007). Economic insecurity and welfare preferences: A micro level analys is. Comparative Politics 39(3), 293 310. Murphy, J., & Tobin, K. (2011). Homelessness comes to school London: S age National Alliance to End Homelessness (NAEH) (2012) Denver s h elter a ssessment, Retrieved June 24, 2013, from http://wwwdenversroadhome .org/jiles/Denver%20Shelter%20Assesment_2012.pdf

PAGE 123

! 114 National Law Center on Hom eless & Poverty (NLCHP). 2011. Criminalizing Crisis: The c riminalization of homelessness in U.S. c ities. Retrieved. February, 2013, from http://www.nlchp.org/content/pubs/11.14.11% 20Criminalization%20Report%20&%2 0Advocacy%20Manual,%20FINAL1.pdf Neale, J. (1997). Homelessness and theory reconsidered. Housing Studies 12 (1), 47 61. Nichols, A. et al. (2011). Economic insecurity and the great r ecession. Retrieved of January 1, 2014 from http://www.urban.org/publications/1001570.html. Peck, J., & Tickell, A. (2002). Neoliberalizing space. Antipode 34 (3), 380 404. Politico, Inc (2013) 2012 Presidential e lection. Retrieved September, 2013, from http://www.politico.com/2012 el ection/map/#/President/2012/ Pruijt, H. (2013). Culture wars, revanchism, moral panics and the creative city. A reconstruction of a decline of tolerant public policy: the case of Dutch anti squatting legislation. Urban studies 50 (6), 1114 1129. Quigley, J. M., Raphael, S., & Smolensky, E. (2001). Homeless in America, homeless in California. Review of Economics and Statistics 83 (1), 37 51. Quigley, R. J., & Taylor, L. C. (2004). Evaluating health impact assessment. Public Health 118 (8), 544 552. Raph ael, S. (2010). Homelessness and housing market regulation. How to house the homeless. New York: Russell Sage Robinson, T., (2013). The Denver camping ban: A report from the street. Retrieved May 22, 2013 from http://www.denverhomelessoutloud.org Ryan, W. (1976). Blaming the victim (Vol. 226). New York: Random House Digital, Inc. Sawhill, I. V. (2003). The Behavioral Aspects of Poverty. Public Interest 153 79 93. Scheve, K., & Slaughter, M. J. (2004). Economic insecurity and the globalization of pro duction. American Journal of Political Science 48 (4), 662 674. Sennett, R (1970) The uses of disorder: P ersonal identity and city life. New York: W.W. Norton & Company, Inc Selbin, J., Co oter, J., Meanor, E., & Soli, E. (2012). Does Sit Lie Work: Will B erkeley's' "measure s" increase economic activity and improve services to homeless p eople?. Available at SSRN 2165490

PAGE 124

! 115 Selbin, J., Cooter, J., Meanor, E., & Soli, E. (2012). Does Sit Lie Work: Will Berkeley's" Measure S" Increase Economic Activity and Imp rove Services to Homeless People?. Siegel, F. (1995). Reclaiming our public spaces. Metropolis. New York Siegel, F. (1997 ). The future once happened here: New York, DC, LA, and the fate of America's big cities San Francisco: Encounter Books. Smith, N. (1996). The new urban frontier: Gentrification and the revanchist city New York: Routledge. Turner, R. S. (2002). The politics of design and development in the postmodern downtown. Journal of Urban Affairs 24 (5), 533 548. United States Interagency Cou ncil on Homelessness (USICH) (2012). Searching out solutions: Constructive a lternatives to criminalization of homelessness [Electronic version] Retrieved May 22, 2013, form http://www.usich.gov/issue/alternatives_to_criminalization United Nations: Offi ce of the High Commissioner for Human Rights (2012, April 23). Retrieved May 22, 2013, from http://www.ohchr.org/EN/NewsEvents/Pages /DisplayNews.aspx?NewsID=12079&LangID=E United Nations: Population Fund (UNFPA) (2013). Linking popul ation, poverty and development. Retrieved on June 1, 2013 from http://www.unfpa.org/pds/urbanization.htm United States Census Bureau (1990) 1990 Census of population and housing public law 44 171 data (official) age by race and Hispanic origin. Retrieved July, 14, 2013, fro m http://censtats.census.gov/pl94/pl94.shtml. United States Census Bureau (2000) American fact finder: 2000 census. Retrieved July, 14, 2013, from http://factfinder2.census.gov/faces/nav/jsf/pages/index.xhtml. United States Census Bureau (2011) American fact finder: American community survey. Retrieved July, 14, 2013, from http://factfinder2.census.gov/faces/nav/jsf/pages/index.xhtml. United States Department of Justice (DOJ): Federal Bureau of Investigation (2010) "Uniform Crime Reporting Statistics (U CR)" Database. Last updated 2010. Retrieved on July 15, 2013 from http://www.bjs.gov/ucrdata/. United States Housing and Urban Development (HUDP) (2013). PIC and HIC data since 2007. Retrieved March 1, 2013 from https://www.onecpd.info/resource/3031/pit and hic data since 2007

PAGE 125

! 116 Van Eijk, G. (2010). Exclusionary policies are not jus t about the neoliberal city': A critique of theories of urban revanchism and the case of Rotterdam. International Journal of Urban and Regional Research 34 (4), 820 834. War d, K. (2010). Towards a relational comparative approach to the study of cities. Progress in Human G eography 34 (4), 471 487. Wasserman, J., A & Clair, J., M. (2011) At home on the street: People poverty & a hidden culture of homelessness. Boulder CO : Lyn ne Rienner Pub Whitaker, I., P. and Time, V. (2001 ). Devolution and welfare: The social and legal implications of state inequalities for welfare r eform in the United States Social Justice San Francisco 28(1), 76 90. Wilson, J., Q. and Kelling, G., L. (1982). Broken Windows [Electronic version] Retrieved on April 1, 2013 from, http://www.theatlantic.com/doc/print/198203/broken windows Wright, T. (1997). Out of place: Homeless mobilizations, subcities, and contested landscapes New York: SUNY Press. Wyly, E., & Hammel, D. (2005). Mapping neo liberal American urbanism. Gentrification in a global context: The new urban colonialism 18 38.

PAGE 126

! 117 APPENDIX A Results Summary Test Independent Variable Significant Relationship ? Direction of Relationship Assessm ent of the Strength of the Relationship Threat Average Homeless Population (2005 2012) Yes + Weak Percent Change in Homeless Population (2005 2012) Yes Weak Average Violent Crime Rate (per 100,000; 2000 2010) No N/A N/A Percent Change in Violent C rime Rate (per 100,000; 2000 2010) Yes M oderate Average Property Crime Rate (per 100,000 2000 2010) No N/A N/A Percent Change Property Crime rate (per 100,000; 2000 2010) Yes + Strong Neoliberal Percent Change in Housing Units (2000 2011) No N/A N/A Percent Change in Median Owner/Occupied Housing Value (2000 2011) Yes Strong Moderate Growth in Aggregate Property Value (2000 2011) Yes Weak Growth of Location Quotient of High Tech Sector relative to national mean. (2008 2012) Yes + Moderate Weak Culture Shift Average Political Voting Gap (+ = Democrat; 2000 2012) No N/A N/A Percent Change in Political Voting Gap (+ = Democrat; 2000 2012) Yes + Strong est Average Black/ White + Average Hispanic/ White + Average Asian/ White Dissimilarity Index (1990 2000 2010) Yes Weak Percent of Population White only (2011) Yes Strong est Percent Change in Median Income (2000 2010) No N/A N/A

PAGE 127

! 118 APPENDIX B Regression Summery Test Independent Variable Regression Coefficient P Test Null rejected? Th reat Average Homeless Population (2005 2012) 0.00018 0.00264 Yes Percent Change in Homeless Population (2005 2012) 0.37769 0.02282 Yes Average Violent Crime Rate (per 100,000; 2000 2010) 0.00019 0.83014 No Percent Change in Violent Crime Rate (per 100,000; 2000 2010) 1.70303 0.03784 Yes Average Property Crime Rate (per 100,000 2000 2010) 0.00026 0.12519 No Percent Change Property Crime rate (per 100,000; 2000 2010) 3.50985 0.01977 Yes Neoliberal Percent Change in Housing Units (2000 2011) 2.75 001 0.20062 No Percent Change in Median Owner/Occupied Housing Value (2000 2011) 1.94764 0.0181 Yes Growth in Aggregate Property Value (2000 2011) 2.7860E 11 .048 Yes Growth of Location Quotient of High Tech Sector relative to national mean. (2008 2012) 0.46465 0.0043 Yes Culture Shift Average Political Voting Gap (+ = Democrat; 2000 2012) 0.35506 0.76451 No Percent Change in Political Voting Gap (+ = Democrat; 2000 2012) 7.3437 0.00438 Yes Average Black/ White + Average Hispanic/ White + Ave rage Asian/ White Dissimilarity Index (1990 2000 2010) 0.02843 0.02023 Yes Percent of Population White only (2011) 4.7586 0.01149 Yes Percent Change in Median Income (2000 2010) 2.75859 0.33269 No

PAGE 128

! 119 APPENDIX C List of Independent Variables: Threa t % Unsheltered Homeless ( 2012 HUD) T otal Chronically Homeless (2012 HUD) Chronic ally Homeless Unsheltered (2012 HUD) % Chronically homeless ( 2005 2012 HUD) % Chronically Homeless Unsheltered ( 2005 2012 HUD) % P opulation Unsheltered Chronic ( 2005 2 012 HUD) % Change in Homeless (2005 2012 HUD) % Change State Homeless ( 2005 2012 HUD) % P opulation Homeless (HUD) % P opulation unsheltered (HUD) Change in Violent Crime Rate (1995 2011 FBI) Change in Violent Crime Rate (20 00 20 11 FBI) % Change VCR ( 19 95 2011 FBI) % Change VCR (2000 2011 FBI) Change in Property Crime Rate (19 95 2011 FBI) Change in Property Crime Rate (20 00 2011 FBI) % Change Property Crime Rate (19 95 2011 FBI) % Change Property Crime Rate (2000 2011 FBI) Average PCR (FBI) Avera ge VCR (FBI) Average Homeless population ( 2005 2012 HUD) Neoliberal Growth Creativity Rank (Florida 2011) Creativity Index (Florida 2011) Tech Rank (Florida 2011) Talent Rank (Florida 2011) Tolerance Rank (Florida 2011) %Creative Class employmen t (Florida 2011) % Working Class Employment (Florida 2011) % Service Class Employment (Florida 2011) FDI State in Millions (2012) (www.fdiintelligence.com) # of FDI contracts per city (limited data) (www.fdiintelligence.com) # of FDI contracts per St ate (www.fdiintelligence.com) Creative class Employment Pop (Florida 2011) Average EDU population W/ Bachelors and above ( Census 2000 2011) Growth EDU population W/ Bachelors and above (Census 2000 2011) % Change EDU Population W/ Bachelors and above (Census 2000 2011)

PAGE 129

! 120 Average Median Rent 2007 2011 Growth of Median rent (Census 2000 2011) % Change In Median Rent (Census 2000 2011) % Change Aggregate Property Value (C ensus 2000 2011) 2011 A ggregate home Value (Census 2011 ) Growth Agg regate Value ( Census 2000 2011) Change in Median Owner Occupied Value (Census 2000 2011) % C hang e Median Owner Occupied Value (C ensus 2000 2011) Housing Unit Growth ( Census 2000 2011) % Change in Housing Units ( Census 2000 2011) Change in Rental Vacancy ( Census 200 0 2011) 2011 Rental Vacancy ( Census 2000 2011) 1yr J ob Growth 08 (best cities.org) 5y Wage G rowth 08 (best cities.org) 1yr W age Growth 08 (best cities.org) % J ob Growth 08 (best cities.org) 5yr HT GDP 08 (best cities.org) 1yr Relative H igh T ech GDP 08 (best cities.org) H igh T ech L ocation Q uotient (HT LQ) 2007 20 08 (best cities.org) # Of HT GDP >1 08 (best cities.org) Metro POP 08 (best cities.org) 2008 Index (best cities.org) State (best cities.org) 2012 Rank (best cities.org) 5yr Job Growth ( best cities.org) % 5yr wage growth (best cities.org) 2 y ea r Employ change (best cities.org) 5 Y ea r HT GDP Growth (best cities.org) 2011 HT GDP LQ (best cities.org) # Of HT LQ over 1 (1 equals national average) (best cities.org) 2012 Metro Population (best cities.org) Rebased Index (best cities.org) % Metro Pop Change (best cities.org) Change in Index (best cities.org) AVE index (best cities.org) AVE Job Growth (best cities.org) % AVE Job Growth (best cities.org) AVE #of HT LQ (best cities.org) Growth of # of HT LQ over 1 (best cities.org) Change in % MGR/Income (census 20 07 20 11) Average MGR/I (census 20 07 20 11)

PAGE 130

! 121 Culture Shift 1990 Pop ulation ( 1990 Census) 2000 Population ( 2000 Census) 2011 P op ulation ( 2011 Census) Population Growth ( 2000 2011 Census) Pop ulation Growth ( 1990 2011 Census) % Pop Growth ( 2000 2011 Census) % Pop ulation Growth ( 1990 2011 Census) Average Population ( 1990 2011 & 2000 2011 Census) 2011 State Pop ulation ( 2011 Census) Growth of White pop ulation (2000 2011 C ensus) 2011 % White only population ( 2011 Census) Change in % White only population ( 2000 2011 Census) % Change Black Population ( 2000 2011 Census) % Change Hispanic Pop ulation ( 2000 2011 Census) 2011 % Black Population ( 2011 Census) 2011 % Hispanic Population ( 2000 2011 Census) % Change Poverty ( 2000 2011 Census) % Population in Poverty ( 2011 Census) % Populations Age 0 17 ( 2011 Census) % Population Age 18 64 ( 2011 Census) % Population Age 65 over ( 2011 Census) % Populations 6 5 and over Poverty ( 2011 Census) White/Black Dissimilarity index 1990 (http://www.s4.brown.edu/) White/Black Dissimilarity index 2000 (http://www.s4.brown.edu/) White/Black Dissimilarity index 2010 (http://www.s4.brown.edu/) Average W hite/Black Dissimilarity index 1990 2 010 (http://www.s4.brown.edu/) % Change White/Black Dissimilarity index 20yrs (http://www.s4.brown.edu/) % Change White/Black Dissimilarity index 10yrs (http://www.s4.brown.edu/) White/Hispanic Dissimilarity index 1990 (http://www.s4.brown.edu/) White/ Hispanic Dissimilarity index 2000 (http://www.s4.brown.edu/) White/Hispanic Dissimilarity index 2010 (http://www.s4.brown.edu/) Average White/Hispanic Dissimilarity index 1990 2010 (http://www.s4.brown.edu/) % Change White/Hispanic Dissimilarity index 2 0yrs (http://www.s4.brown.edu/) % Change White/Hispanic Dissimilarity index 10yrs (http://www.s4.brown.edu/) White/Black Exposure index 1990 (http://www.s4.brown.edu/) White/Black Exposure index 2000 (http://www.s4.brown.edu/) White/Black Exposure inde x 2010 (http://www.s4.brown.edu/) Ave White/Black Exposure index (http://www.s4.brown.edu/) % Change White/Black Exposure index 20yrs (http://www.s4.brown.edu/) %Change White/Black Exposure index 10 yrs (http://www.s4.brown.edu/)

PAGE 131

! 122 Total Average White/ Black+Hispanic+Asian Dissimilarity Index 1990 2010 (http://www.s4.brown.edu/) Average annual temperature Average Jan Temp 2013 State Governor's party 2012 Obama % vote 2012 Romney % vote 2012 voting Gap (+=Dem) 2000 Gore % vote 2000 Bush % vote 20 00 Voting Gap (+=Dem) Average Voting Gap (2000 2012) 2008 DEM Rep Gap 2008 DEM Rep Gap [Absolute Value] Average voting gap [Absolute Value] Voting Gap 2012 [Absolute Value] % Political Gap Shift (D=+) (2000 2012) Change in Gini Index (2007 2011 Cen sus ) A v e change in Gini Index (2007 2011 Census ) % Change in Per Capita Income (2000 2011 Census) 2011 Median Income (2000 2011 Census ) Growth of Median Income (2000 2011 Census ) Average Median I ncome (2000 2011 Census) % Change Median Income (2000 2 011 Census) % Change White/Asian Dissimilarity 20 years (http://www.s4.brown.edu/) %Change White/ Asian Dissimilarity 10 years (http://www.s4.brown.edu/) % Change White/Asian Exposure 10 years (http://www.s4.brown.edu/)

PAGE 132

! 123 A PPENDIX D Standardized Results : Test Independent Variable Standardized Correlation Coefficient P Test Null rejected? Threat Average Homeless Population (2005 2012) 0.54607 0.00264 Yes Percent Change in Homeless Population (2005 2012) 0.19295 0.02282 Yes Average Violent C rime Rate (per 100,000; 2000 2010) 0.02923 0.83014 No Percent Change in Violent Crime Rate (per 100,000; 2000 2010) .21458 0.03784 Yes Average Property Crime Rate (per 100,000 2000 2010) 0.16228 0.12519 No Percent Change Property Crime rate (per 10 0,000; 2000 2010) .23792 0.01977 Yes Neoliberal Percent Change in Housing Units (2000 2011) 0.12066 0.20062 No Percent Change in Median Owner/Occupied Housing Value (2000 2011) 0.25882 0.0181 Yes Growth in Aggregate Property Value (2000 2011) 0.3567 6 .048 Yes Growth of Location Quotient of High Tech Sector relative to national mean. (2008 2012) 0.24883 0.0043 Yes Culture Shift Average Political Voting Gap (+ = Democrat; 2000 2012) 0.03316 0.76451 No Percent Change in Political Voting Gap (+ = Democrat; 2000 2012) 0.26923 0.00438 Yes Average Black/ White + Average Hispanic/ White + Average Asian/ White Dissimilarity Index (1990 2000 2010) 0.26079 0.02023 Yes Percent of Population White only (2011) 0.3029 0.01149 Yes Percent Change in Med ian Income (2000 2010) 0.10028 0.33269 No

PAGE 133

! 124 APPENDIX E Pearson's Correlation M atrix : 1: Total Homeless Laws 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2: Average Homeless Pop .13 1.0 3: % Change Homeless .23 .06 1.0 4: % C hange VCR .24 .19 .00 1.0 5: % Change PCR .09 .09 .00 .53 1.0 6: Ave PCR .12 .26 .07 .1 .22 1.0 7: Ave VCR .09 .08 .07 .14 .13 .48 1.0 8: % Pol Gap Change .43 .00 .09 .21 .08 .06 .07 1.0 9: Ave Pol Gap .07 29 .02 .07 .06 .08 .40 .22 1.0 10: Total Ave Diss Index .12 34 .06 .09 .08 .08 .40 .12 .38 1 .0 11: % White only .19 .17 .09 .13 .09 .18 .62 .1 .46 .53 1 .0 12: % Change Median O/O Value .18 .25 .09 09 .21 .06 .09 .07 25 .03 .12 1 .0 13: % Change in Housing Units .05 .05 .06 .18 .18 .02 .11 .11 .30 .35 .27 .11 1 .0 14: Growth of LQ HT .12 .02 24 .14 .12 .11 .09 .04 .14 .1 .01 .09 .09 1 .0 15: % Change Median Income .08 .19 .08 .1 .12 .17 .11 .05 .17 .04 .05 .61 .05 .05 1 .0 16: Agg Growth Value .03 .88 .03 .16 .08 .28 .03 .02 3 .31 14 .26 .03 .11 22 1 .0 (Note: Bold numbers are statistically significant (> .05) and numbers correspond to variables.)