Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
00001.tif
(USC Thesis Other)
00001.tif
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type o f computer printer.
The quality o f this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
THE IMPACTS OF REFORMED COUNTY POLITICAL STRUCTURES
ON COUNTY POLICY OUTCOMES
by
Yongman Park
A Dissertation Presented to the
FACULTY OF THE SCHOOL OF PUBLIC ADMINISTRATION
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PUBLIC ADMINISTRATION
December 1995
Copyright 1995 Yongman Park
UMI Number: 9617133
UMI Microform 9617133
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY OF SOUTHERN CALIFORNIA
SCHOOL OF PUBLIC ADMINISTRATION
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089
This dissertation, written by
Yongm an, # P a r f t.................
under the direction o f h i. a... Dissertation
Committee, and approved by ad its mem
b e r s , has b e e n presented to and accepted by
the faculty o f the School o f Public
Administration, in partial fulfillment o f
requirements of the d e g r e e o f
DOCTOR OF
PUBUC ADM INISTRATION
(Date. . Q < ? . t.( ?ke . r . .? A*. A? . 9 5
DEDICATION
This dissertation is dedicated to my parents who educated me to be independent, and
to my wife, Inae, who advises and advocates everything I do.
ACKNOWLEDGMENTS
Four years ago when I began graduate studies at the School of Public
Administration, the prospect of accomplishing the work was uncertain. The gradual
progress of my knowledge culminating in this dissertation could never have been
completed without the mindful support of many persons.
I first wish to express my appreciation to three dissertation committee
members. Professor Ross Clayton, chairperson of my committee, has been the
single most supportive person, supervising the whole process from my coursework
to this dissertation. His generosity, guidance, and intellectual and moral
encouragement were ever-present and unfaltering whenever I plunged into trouble
during the preparation of this dissertation. I will never forget all he showed me.
Professor Chester A. Newland taught me by practice what moral standard a human
being should have. His warmth, dignity, and insight about human civilization will
remain as an example for me to emulate for the rest of my life. Professor Jeffrey I.
Chapman stimulated and sharpened my academic interest and world perspective.
His sincerity as a teacher and researcher will guide my later intellectual
development.
I want to thank the Inter-university Consortium for Political and Social
Research (ICPSR) for the provision of county financial census data files at a
minimal price. Linda Lopez, a research assistant in the computer lab of the Political
Science department, helped me to place the order from ICPSR and upload the data
files onto the university mainframe computer. I acknowledge her assistance. I also
want to thank Jennifer Jung for her careful reading and grammatical correction of
the dissertation draft. The staff of the Korean Cultural Center in Los Angeles,
where I worked before beginning my studies, also deserves my deep recognition.
Korean doctoral students in the school, particularly Professor Choong Kyong Huh,
must be thanked for their stimulation of my academic interests in various subjects.
Finally, I wish to specially acknowledge all my family members who have
provided me with comfort and love. I hope my children, Sunmin and John,
understand why Dad had no time to play with them. I am sincerely grateful to my
mother-in-law for her financial support in the last year of my study.
V
TABLE OF CONTENTS
Page
DEDICATION.......................................................................................................... ii
ACKNOWLEDGMENTS........................................................................................ iii
LIST OF TABLES................................................................................................... viii
LIST OF FIGURES.................................................................................................. x
CHAPTER I: INTRODUCTION........................................................................... 1
1.1 Introduction........................................................................................... 1
1.2 Background on the Research Problem.............................................. 2
1.3 Purpose and Significance of the Dissertation................................... 5
1.4 Scope and Methodology of the Dissertation.................................... 6
1.5 Overview of the Dissertation.............................................................. 8
Notes............................................................................................................ 10
CHAPTER II: COUNTIES IN TRANSITION: SUBURBANIZATION,
FUNCTIONS, AND GOVERNING STRUCTURES 1 1
2.1 Introduction.......................................................................................... 1 1
2.2 History and Nature of County Government..................................... 14
2.3 Suburbanization................................................................................... 19
2.3.1 Evidence........................................................................................ 21
2.3.2 How Suburbanization Affects County Government................ 26
2.4 County Functions................................................................................. 32
2.5 Forms of County Government........................................................... 37
2.6 Summary............................................................................................... 47
Notes............................................................................................................ 49
CHAPTER III: RESEARCH ON THE IMPACT OF LOCAL POLITICAL
STRUCTURES............................................................................ 51
3.1 Introduction.......................................................................................... 51
3.2 The Legacy of the Reform Age.......................................................... 52
3.2.1 Ethos Theory............................................................................. 52
vi
3.2.2 Reformed Political Structures.................................................... 54
3.3 Research on City Reformed Structures............................................. 56
3.3.1 Development of Research........................................................... 56
3.3.2 Two Methodological Issues........................................................ 61
3.4 Research on County Governments.................................................... 63
3.4 1 Recent County Studies............................................................... 63
3.4.2 Variables Unique to County Studies........................................ 65
3.5 Summary............................................................................................... 69
Notes............................................................................................................ 71
CHAPTER IV: RESEARCH DESIGN, DATA SOURCES, AND
VARIABLES................................................................................. 73
4.1 Introduction........................................................................................... 73
4.2 Research Design.................................................................................. 74
4.2.1 Cross-Sectional Design.............................................................. 74
4.2.2 Model of the Study..................................................................... 75
4.3 Data Sources......................................................................................... 77
4.3.1 List of Data Sources.................................................................... 77
4.3.2 Data Characteristics.................................................................... 78
4.4 Policy Outcome Variables-Dependent Variables............................. 80
4.4.1 Total Expenditures Per Capita................................................... 80
4.4.2 Total Tax Revenues Per Capita................................................. 81
4.4.3 Total Expenditures Per Capita for Urban Functions............... 81
4.4.4 Summary of Dependent Variables............................................. 83
4.5 Reformed Political Structures-Main Independent Variables 84
4.5.1 County Structural Reform Index................................................ 84
4.5.2 Separate Applications of County Reformed Items................. 87
4.6 Determinant Variables........................................................................ 87
4.6.1 Demographic and Socioeconomic Variables............................ 87
4.6.2 County Functional Scope........................................................... 89
4.6.3 State Mandates............................................................................ 89
4.6.4 Intergovernmental Revenues..................................................... 91
4.6.5 Region........................................................................................... 92
4.6.6 Age................................................................................................ 92
4.6.7 Other Determinant Variables...................................................... 93
4.7 Data Analysis Techniques.................................................................. 93
4.8 Research Hypotheses Statements....................................................... 94
Notes............................................................................................................ 97
vii
CHAPTER V: RESEARCH RESULTS AND DISCUSSIONS......................... 98
5.1 Introduction........................................................................................... 98
5.2 Results and Discussions for All Counties...................................... 99
5.3 Results for MSA Counties.................................................................. Ill
5.4 Results for Non-MSA Counties.......................................................... 114
5.5 Long-term Impact of Reformed County Structures.......................... 116
5.6 Summary of Major Research Findings............................................. 118
CHAPTER VI: CONCLUSIONS........................................................................... 122
6.1 Introduction........................................................................................... 122
6.2 Summary of the Previous Chapters.................................................. 122
6.3 Implications of This Dissertation....................................................... 125
6.4 Recommendations for Future Study................................................... 127
BIBLIOGRAPHY..................................................................................................... 130
APPENDIX............................................................................................................... 147
viii
LIST OF TABLES
2.1: Urban and Rural Population Change (Percentage) by Urban Area
Size................................................................................................................................ 22
2.2: County Population Change (Average) by Region and Size....................... 25
2.3: Percentage of Selected Services Contracted Out by Cities......................... 28
2.4: Correlation Coefficient between County Population Change and
Unincorporated Ratio Controlling for Population....................................... 31
2.5: Selected Functions Performed by Metropolitan County Governments,
1971, 1975, and 1989..................................................................................... 34
2.6: Percent Distribution o f Local Government Expenditures by Type of
Local Government and by Function, 1966-1967, 1976-1977, and 1986-
1987.................................................................................................................... 35
2.7: Forms o f County Government, 1971, 1978, and 1989.............................. 39
2.8: Change of County Forms Between 1978-1989........................................... 40
2.9: County Forms in Home Rule States in 1989................................................ 43
2.10: Correlation Coefficients Between Row Officers Index and County
Home Rule Status, Executive Form, and Executive/Administrator
Form, Controlling for Population.............................................................. 45
4.1: County Distribution by Population Category.............................................. 79
4.2: Comparison Between the Cases Included in the Study and the Universe
of County Data................................................................................................ 83
4.3: Mean Values of Per Capita Total Expenditures, Tax Revenues, and
Urban Expenditures by Regions.................................................................... 83
4.4: County Reform Items and Coding Scheme................................................. 84
4.5: Varimax Rotated Factor Matrix Among Reform Items............................. 86
ix
4.6: Distribution of Counties by Reform Index.................................................. 87
4.7: Expected Directions Between Dependent Variables and Determinant
Variables.......................................................................................................... 95
5.1: Regression Results for All Counties........................................................... 100
5.2: Mean Test for Three Dependent Variables by County Home Rule
Status................................................................................................................ 108
5.3: Correlation Coefficients Between Row Officers Index and Three
Dependent Variables...................................................................................... 109
5.4: Correlation Coefficients Between County Row Officers Index and
Selected County Characteristic Variables.................................................... 109
5.5: Regression Results for MSA Counties........................................................ 111
5.6: Regression Results for Non-MSA Counties................................................ 115
5.7: Regression Results for the Counties With No Formal Change During
the Period of 1978-1989................................................................................. 117
X
LIST OF FIGURES
2.1: Urbanization Chart......................................................................................... 23
2.2: Forms of County Government....................................................................... 39
3.1: Focal Research Variables and Ethos Theory................................................ 56
3.2: Enlarged Relationships Among Policy Outcomes and Their Determinant
Variables............................................................................................................ 70
Yongman Park Ross Clayton
The Impacts of Reformed County Political Structures
On County Policy Outcomes
Review of previous research on city and county reformism led to formulation
of the following research hypothesis: reformed county governments render more
efficient policy outcomes. County executive and administrator forms and at-large
elections are utilized as the reformed structure variables. County policy outcomes
are measured by three per capita fiscal performance variables: total expenditures;
total tax revenues; and total urban expenditures.
The literature review also identified several other determinant (or control)
variables. They are: (1) demographic factors-population, density, population
change; (2) socioeconomic factors-per capita personal income; percentage of
persons below the poverty level; (3) functional scope; (4) state mandates-functional
mandates, property tax and debt limits; (5) five regional divisions-New England,
Mid-Atlantic, North Central, South, West; (6) county age measured by number of
decades; (7) home rule status; (8) elected row officers index; and (9) number of
special districts.
The number of counties included in the ordinary least squares regression
models is 2,246, which represents 74 percent of all 3,042 counties. Data sources
are: (1) Annual Survey o f Governments 1990: Finance Statistics; (2) Census of
Government 1987: Governmental Organization File; (3) USA Counties 1994: A
Statistical Abstract Supplement; (4) Population of Counties by Decennial Census:
1900 to 1990; (5) County Government Structure: A State by State Report (NACO,
1989); (6) State Laws Governing Local Government Structure and Administration
(ACIR, 1993); and (7) State Mandating o f Local Expenditure (ACIR, 1978).
Research results show that more reformed county governments render higher
levels of policy outcomes, rejecting the efficiency hypothesis of this dissertation.
These positive relations are very clear regardless of whether these analyses are
applied separately to metro or non-metro areas. The long-term effects of reformed
county structures also show the same outcomes. In addition, other research
variables unique to county governments such as home rule status, row officers
index, and county unincorporated ratio are found to have significant and individual
effects in determining county policy outcomes. These facts demonstrates the
research potential of these unique variables for future county studies.
1
CHAPTER I
INTRODUCTION
1.1 Introduction
Do reformed county political structures impact policy outcomes? This is a main
question this dissertation attempts to answer. The impact of structure has long been
a topic for political scientists (for example, parliamentary vs presidential systems).
Structural reform has been a tenet of sorts in guiding the county modernization
movement ever since the progressive era, and particularly in the early 1970s.
Efforts to improve the performance of county governments, however, have been less
noticed than those focused on municipal governments.
Recent arguments about the embedded and long-lasting influence of
institutions and culture hint that changes in behavioral rules and social constraints,
rather than structural reorganization, play a more fundamental role in enhancing the
long-term performance of an economic or political system (North, 1994, 1990;
Ostrom, 1986). Nonetheless, much research, mainly on municipal governments,
shows that the political structure o f a governing body has, at the least, a mediating
effect on the outcomes of city politics.
Factors determining political structure were the primary focus o f students of
municipal governments in earlier years. But that focus changed because of the
difficulty in isolating community factors which historically affected adoption of
current governing structures many years (or decades) earlier. Instead, current
studies emphasize the impact of political structure on policy outcomes of city
politics, and focus on the determinants of such outcomes. These two subjects as
applied to county governments are the main focus of the inquiry of this dissertation.
1.2 Background on the Research Problem
Since counties were first established by the colonial Virginia government in 1634,
the nature of county governments has remained unchanged. Counties have been the
administrative arms of the state for about three centuries. Their autonomous
functional responsibilities for local affairs were secondary to city governments, and
their scope was also very limited. This situation continued even throughout the first
urbanization process which took place at the turn of this century. While much
attention from both reformers and researchers was drawn to the municipalities in the
progressive era, the county system was ridiculed by the term, “The Dark Continent
of American Politics” (Gilbertson, 1917).
Like other social systems, in the American county system, “form follows
function” (Cassella, 1971). In the absence of growth in their functional
responsibility, county governing structures had nothing to do with change. These
two aspects of counties-functional growth and formal change— move together. Until
recently, the county has been called the “forgotten government” in research
literature (Schneider & Park, 1989; Marando & Thomas, 1977). Almost all
arguments for structural reform have focused on the problems of municipal
governments, not county governments.
This situation is changing fast. In the early 1970s, an article titled, “3,049
Labs for Local Government Testing” recognized county changes and the need for
county research (McArthur, 1971). NACo’s four year-books published during
1975-1978 and ACIR’s report in 1972 are the most notable information sources
about expanding county functions and structural change in the 1970s. In 1977,
Marando and Thomas’s research on the decision-making role of county
commissioners in Florida marked the beginning of empirical county studies.
Nonetheless, serious academic interest in county governments was not aroused until
the 1980s.
Since then, much research on county governments has been done in various
areas such as: governance structures (Cigler, 1995; Lewis & Taylor, 1994; DeSantis
& Renner, 1994, 1993; Jeffery et al., 1989; DeSantis, 1989; Schneider & Park,
1989); functions (Park, 1994; Benton & Menzel, 1993, 1991; Benton & Rigos,
1985); home rule (Martin & Nyhan, 1994; Salant & Martin, 1993; Salant, 1988;
Berman et al., 1985); fiscal aspects (Duncombe et al., 1992; Ebel, 1991; Downing,
1991; Thomas, 1991a; MacManus & Pammer, 1990; Greene, 1987); rural county
problems (Cigler, 1993; Braaten, 1991); and state-county relations (Salant, 1993;
Waugh & Streib, 1993; Streib & Waugh, 1991a, 1991b; Waugh, 1988; Berman &
Martin, 1988).
The burst of research interest in county governments has not been accidental.
Counties had drawn no interest from researchers studying city government reform in
the progressive era. Environmental forces toward county change now have those
researchers paying attention to county governments. These environmental
imperatives are known as suburbanization and development of polycentric suburban
communities (Downs, 1994; Hughes, 1993; Harthshorn & Muller, 1992; Stanback,
1991; Frisbie & Kasarda, 1988; Richardson, 1988), leading to the evolution of the
megacounty (Fosler, 1991; Herson & Bolland, 1990; Fishman, 1990) or full-service
county governments (Benton & Menzel, 1993). Chapter two discusses in detail how
these environmental phenomena affect county governments.
Keeping pace with the growing interest in county governments is a gradual
increase in empirical research on reformed county structures. DeSantis and Renner
(1994) measured the effects of the county administrator and executive forms,
controlling for region, population size, and per capita income, etc. Updating his
previous study with Schneider (1989), Park (1994) explored the complementary
nature of local governments’ services and expenditures among counties, central
cities, and suburban cities in metropolitan areas. Park incorporated political
structures and state mandate variables into his interactive expenditure determinant
model. Related closely to the reformed structure of county governments, Salant and
Martin (1993) examined the determinants of the number of county elected row
officer positions mandated by state laws.1
Another home rule study by Martin and Nyhan (1994) identified a state’s
growth orientation, innovativeness, and number of municipalities as significant
predictors for whether a state grants home rule power to its counties. Benton and
Menzel (1991) investigated the determinants of the county functional scope with
such variables as county administrator form, home rule charter, number of
municipalities in a county area, and other demographic and socioeconomic factors.
These studies about county governing forms are very important to this
dissertation, but they are not sufficient to permit sole reliance on them. The
theoretical underpinnings for the impacts of reformed governing structures have
been developed from studies of city politics. These empirical research results cannot
be neglected.
1.3 Purpose and Significance of the Dissertation
The main purposes of this dissertation are twofold. The first purpose is to verify
whether the reformed political structure of county governments has an impact on
their fiscal performance.
A group of county scholars recently recognized the significance of the study
of impacts of reformed county structures (Menzel et al., 1992). The researchers
classified county research agenda for the 1990s into eight topic sets. In their first
agenda set titled, “Structure, Reform, and Performance,” the impacts of reformed
county structure on county performance were ranked first, with the impact of home
rule status ranking second. In Cigler’s research list (1994), these research items
were also placed on top. This high priority placed by impact studies upon reformed
county structure indicates that county scholars consider it as the most important
research topic these days.
The second purpose of this dissertation is to examine the determinants of
county policy outcomes. This dissertation attempts to investigate the impacts of
some “unique variables for county studies,” for example, county home rule status,
number of row officer positions, and county unincorporated ratio. The empirical
assessment about these variables can contribute to current knowledge about county
governments and suggest ideas for future county studies because they have not been
examined in detail by previous research.
1.4 Scope and Methodology of the Dissertation
Many perspectives exist for viewing the political system of local government
differently as it processes citizens’ demands and produces final policy outcomes in
response to these demands (Clark, 1994; Miranda & Walzer, 1994; Ross et al.,
1991, chapter 3; Kantor, 1988; Wong, 1988; Sanders & Stone, 1987; Peterson,
1987 & 1981). In urban studies, the question is which perspective is more accurate
in explaining the variance in how policy outcomes are decided.
Broadly, two perspectives compete with each other in predicting policy
outcomes of city politics. One perspective emphasizes internal, political, cultural,
and socioeconomic factors; while the other perspective focuses on external,
economic, and environmental imperatives. Wong (1988) named the first as the
“political factor model,” and the second as the “economic constraint model.” This
dichotomy is also suggestive of appropriate research methods. Case or field surveys
are usually utilized in the political model, but secondary statistical data analysis is
common in the economic model.
It may be irrelevant at this point to directly apply an economic model to the
study of county governments without further theoretical elaboration about the
differences of internal and external constraints which cities and counties face.
Therefore, to raise research questions about reformed county structures, this study
reviews in chapter three the main content of ethos theory, a political factor model.
But the research questions focused on in this dissertation are very similar to the
typical determinant models of local government expenditures. In defining and
selecting policy outcome and control variables for the determinant model, an
economic constraint model is partially adopted. For example, rather than the
internal socioeconomic factors like racial, ethnic, and religious variables utilized
intensely in ethos theory, external determinant variables and those variables unique
to county governments are chosen for the model used in this dissertation.
Three types of county political structures are examined: the commission,
county administrator, and county executive forms. These structural types combined
with other county reform items comprise a county structural reform scale, which is
different from the city reformism index (details in chapter four). A separate
application of structural and electoral variables is also formed in chapter four. The
data for these variables are obtained from the Government Organization File 1987
by the Bureau of the Census (1990a) and NACo’s two county structural surveys
(1978 and 1989).
To control variations in county expenditure decisions, several additional
variables are incorporated as independent variables. Basically a multiple regression
technique (ordinary least squares) is used to isolate the significant variables, and
bivariate and partial correlations are utilized to explain the degree of correlation
among the focal research variables. County financial data such as expenditures, tax
revenues, and functions are obtained from the Annual Survey of Governments, 1990:
Finance Statistics [Computer File] by Bureau of the Census (1992). USA Counties
1994: A Statistical Abstract Supplement [CD-ROM Version! by the Bureau of the
Census (1994b) is used as the source of county demographic and socioeconomic
data.
1.5 Overview of the Dissertation
This dissertation is divided into six chapters. The second chapter begins with a
description of contextual factors like suburbanization which led county governments
to adopt a reformed structure and provide broader urban functions. The main focus
is placed on expansion of counties’ urban and area-wide functions and their
consequent structural changes. This chapter also attempts to analyze population
growth in county unincorporated areas by constructing a county unincorporated ratio
and to assess relationships between county governing forms, home status, and the
number of row officer positions.
The third chapter is devoted entirely to a literature review on the ethos
theory, the reformed political structures of city governments, and other expenditure-
determinant variables. Results from recent research on county governments are also
explained in this chapter.
Chapter four introduces the research design, data sources, statistical
techniques, and the operationalization of the research variables. Research
hypotheses are formulated at the end of this chapter. Chapter five presents
regression results and discusses the findings.
Finally, chapter six summarizes the main contents of this dissertation, draws
out its practical implications, and provides some recommendations for future study.
10
Notes
1. The term “row” officers refers to the independent county officials whose popular
elections are mandated by state laws or county home rule charters. Their titles,
responsibilities, numbers, and types vary across states. This dissertation utilizes
fourteen “row” officers positions for formulating a row officers index (see chapter
two for details).
These elected county officials are called “row” officers as a whole due to
“their appearance in a row on organizational charts or election ballots and the
relative autonomy of each office from the central board” (Jeffery, et al., 1989).
The term “row” officer is used by a standard textbook (Duncombe, 1977), a
doctoral dissertation (Salant, 1988), academic journal articles (Cigler, 1995; Salant,
1991), and a comprehensive NACo survey (Jeffery, et al., 1989). But a recent
article written by Salant and Martin (1993) uses “county constitutional officers”
instead of row officers; thus these two terms are interchangeable.
11
CHAPTER II
COUNTIES IN TRANSITION:
SUBURBANIZATION, FUNCTIONS, AND GOVERNING STRUCTURES
2.1 Introduction
The aim of this chapter is to provide evidence on how county governments have
changed in their nature, functional scope, and governing structures since World War
II. It is hardly possible to pinpoint a specific event as the cause of county change,
but suburbanization is usually identified as a fundamental cause (Marando & Baker,
1993; Thomas 1991b). A contextual adjustment argument is not new from the point
of view of contingency theories in the organizational behavior field (Scott, 1992).
No system could survive unless it could effectively sense rapid environmental
changes, analyze them, and adapt itself appropriately. Previous administrative
reform efforts and the current reinvention movement are examples of adaptation
processes to environmental changes (Gore, 1994; Dilulio et al., 1993).
Suburbanization simply refers to the concentration of people in suburbia, not
simply in central cities. The degree of suburbanization intensified after World War
II, and suburbanites have strongly urged modernization of county governments to
accommodate growing public service demands. The county modernization
movement basically pursued improvement of county problem-solving capacity.
Duncombe (1977) identified six main trends in the county modernization movement;
they are: (1) changes in county services; (2) the home rule movement which alters
relationships of state and county governments; (3) adoption of newer forms of
county organization; (4) growth of cooperative intergovernmental linkages; (5)
growth in county revenues; and (6) an increase in political accountability brought
about by reapportionment. Among these elements, county service expansion and
adoption of newer forms of county governments relate directly to the purposes of
this dissertation. This chapter focuses on these two trends in detail. County home
rule status is discussed in a later section on county forms which addresses
relationships among different types of county governing forms.
The next section begins with a brief history of the American county system
to show how the county has evolved. Then, evidence of suburbanization is
presented as one cause of county change. How and in what degree counties’
functional scope and structural forms have changed conclude this section.
Two new measurements utilized for the first time in this chapter are worthy
of a brief introduction here. The first is the “unincorporated ratio,” and the second
is the “row officers index.” Using the unincorporated ratio of a county is significant
in that until now much literature has taken for granted that population growth in a
county’s unincorporated areas is directly related to that county’s growth in functions
and structure. But researchers have not analyzed empirically the question of
whether population growth during last decade actually took place in a county’s
unincorporated areas more than in its incorporated areas (central cities and suburban
cities). The absence of empirical evidence in these studies is, because with the
exception of the Census o f Governments surveyed in years ending in 2 and 7 (or
every five years), population data for a county and all the municipalities in its
jurisdiction are not available. The county unincorporated ratio for 1987 used in this
chapter is calculated using the Census o f Governments 1987 which was published in
1990. Using this ratio, a positive relation, though weak, is revealed between
population change in counties during 1980-1990 and those counties’ unincorporated
ratio in 1987, which confirms arguments about population growth in a county’s
unincorporated areas. This ratio is included in regression models dealing with the
impact of county reformed structures on policy outcomes in a later chapter.
The “row officers index” is closely related to discussions about adoption of
the county executive and administrator forms of governance and the impact of
county home rule status. A recent empirical study by Salant and Martin (1993)
utilized a similar index and explored determinant factors for the number of county
constitutional officer positions (row officers). These authors counted the number of
elected row officer positions mandated by state constitutions or general laws. But
Salant and Martin’s study is not written at the individual county level but at the state
level. Therefore, with their data it is not possible to relate the number of row
officer positions of a county to the home rule status and governing forms of that
county. The row officers index in this chapter, however, is calculated at the
individual county level so that it is possible to examine whether home rule counties
have a smaller number of elected row officer positions than non home rule counties,
1 4
and whether executive form counties and administrator counties tend to have fewer
of these positions than commission form counties. This index is calculated from the
Government Organization File [Computer File] in the Census o f Governments 1987.
This file includes the information about popularly elected officials for all local
governments which is collected every ten years (Bureau o f the Census, 1988: v).
From this information, 14 elected county row officer positions are classified and
counted at the county level. More details about both the unincorporated ratio and
row officers index are presented in a later section.
2.2 History and Nature of County Government
Counties and cities, though both creatures of state government, are very different in
many aspects. A “city” is not only the political entity for a population-dense area,
but a center of culture, containing theaters, museums, libraries, colleges,
symphonies, etc.1 The city is also basically the nucleus of economic activities in a
given region. From the early colonial era, the British Crown in fostering
mercantilism, encouraged populated settlements to form their own local governments
for the purpose of regulating trade.2 Ironically, those towns and villages soon
turned out to be “the matrix of self-sustaining cultural life, business ambition, and
political training in the colonies, all of which culminated in their secession from the
British Empire” (Glaab & Brown, 1967: 2). The fact that, even before the state’s
existence, these towns were active as self-sustaining entities partly explains why
cities possess a broader, more privileged status than counties. Thus, the power to
decide which services to provide and what governing structures are needed basically
belongs to cities themselves. State constitutions later endorsed the cities’ broad
discretionary authority, but the authority was by and large inherently given to cities
from the beginning.
In contrast, the county system was modeled after the English shire as the
vehicle for locally performing a state’s affairs.3 Counties were chosen and instituted
by colonists to take care of remaining local functions requiring an area-wide scale
and local equality; these were functions which had proved to be inappropriate for
cities. The counties’ responsibility for local affairs was secondary to that of cities so
that the number of county functions was limited. County officials in charge of such
functions were appointed by the colonial governor; counties could do nothing
without a legal endorsement by the colonial assembly. A county was simply “the
unit for representation in the colonial assembly, and the unit of military, judicial,
highway, and fiscal administration” (Fairlie, 1906; Gilbertson, 1917).
The county system, however, varied regionally from the outset due to
different settlement patterns and regional situations. New York counties featured a
board of supervisors consisting of the towns’ elected supervisors, whereas
Pennsylvania elected county commissioners at large. Southern states adopted
Virginia’s model of the county so that counties in these states took primary
responsibility for local affairs. This was because the counties in those states came
16
into existence before the towns and because their settlement patterns differed from
the New England states where towns were the primary unit of local government.
These regional variations affected later dispersion of county systems in new states.
For example, the Pennsylvania form was most influential for counties in the Western
states (Salant, 1991).
The English shire and colonial settlement patterns are not the only factors
affecting the formation of the current American county system. Martin (1993)
elucidated four other influences; they are: Jacksonian democracy; state court
decisions (for example, Judge Dillon's Rule); reform movements in the progressive
era; and suburbanization after World War II. The Jacksonian idea coupled with the
frontier spirit changed the method of selecting county officials from appointments by
state governments to popular elections. This change represented significant progress
in that, by electing their officials popularly, counties came closer to having the
attributes of a unit of local government.
The dual nature of the county— administrative arm of state and unit of local
government-has always confused those interested in county government. Why did
state governments establish and define counties as local governments, not as their
administrative districts? Porter (1922) said that the answer was to be found in the
desire for self-government rooted deeply in the Anglo-Saxon temperament and the
doctrine of administrative decentralization in the United States. By the desire for
self-government, Porter meant that the American people are willing to pay the price
for enjoying broader local democracy (self-government). By price, he meant gross
inefficiency, lack of uniformity, failure to attain proper standards, and worst of all,
gross wastefulness, all of which were expected to be avoided if state affairs were
managed by state districts.4 This self-government explanation is congruent with the
observation indicating that the English shire was also a de facto unit of local
government though its legal status was as the Crown’s administrative arm (Martin,
1993; Fairlie, 1906). This English tradition with counties has been expanded further
by Jacksonian democracy.
Some state court rulings made in the mid 1800s were contrary to the
expansion of county autonomy by Jacksonian democracy. Among these court
decisions, Judge Dillon’s ruling in 1868 is the most famous. In short, his decision
declared that a municipality could only possess and exercise: (1) powers granted in
express words; (2) powers implied or incident to the power granted; and (3) powers
indispensable to the objects and purposes of incorporation. His decision has been
considered to, on one hand, clarify the legal status of municipalities, but on the
other hand, to restrict the inherent autonomous power of city governments (see
Martin, 1990). But, recent academic interpretations are a bit different. The legal
doctrine supporting Dillon’s rule, that is subordination of local governments to the
state, pre-existed in earlier case law. And emphasis should be placed on the latter
part of his ruling where he called for “fair or reasonable construction of grants of
power to localities” (ACIR, 1993c). While Dillon’s rule was applied fully to cities
18
established for special purposes, court rulings regarding the legal status of the
county were more restrictive than Dillon’s rule for cities. Basically, they defined
the county as a quasi-corporation and a mere administrative arm of the state. The
supreme court of Ohio in 1857 stated:
A municipal corporation proper is created mainly for the
interest, advantage, and convenience of the locality and its
people; a county organization is created almost exclusively
with a view to the policy of the state at large, (cited from
Kneier, 1930b: 142)
This court decision about county status contributed to “stultifying the powers of
county government throughout the 1800s, and has hindered counties in providing
municipal services in the 1900s” (Duncombe, 1977: 23).
Urbanization and the reform movement at the turn of this century have
affected central city governments significantly, but county governments only
minorly. For example, since 1911 when California first allowed home rule charters
for counties, only twenty-four states have given their counties the right to have a
home rule charter. Twenty-one states, or about 90 percent, granted their counties
the chance after 1945, whereas merely three states followed California’s example
before World War II (Salant, 1991). In short, the first aspect of the two natures of
county— the administrative arm of the state-had remained stronger until
suburbanization began to proceed extensively after World War II.
2.3 Suburbanization
In the development of urban America, urbanization preceded suburbanization.
Along with the progress of the industrial revolution during the years of 1860-1920,
massive migration took place from domestic rural areas and overseas into central
cities. This social change gave rise to various social adjustments. Government
reform movements took place locally during the last 40 years of the period. The
next chapter presents the main ideas and the legacy of the reform movement.
Traditionally, a suburb is defined as “a densely settled nonagricultural
community adjacent to and dependent upon a large central city” (Dye, 1965). By
dependent, it meant a one-way flow of manpower from suburbia to the central city
and the movement of products in the opposite direction. This conceptualization
developed by the human ecology school emphasizes intraurban spatial processes
(Frisbie & Kasarda, 1988). Following biotic determinism, this school focuses on
the determining force to equilibrate urban land uses. It identified that force as “the
free competition for space and the price system in the property market.”5 As free
competition for urban space continues, a hierarchical and unequal economic order
develops between the central business district (CBD) and its periphery (Logan &
Molotch, 1987). Thus, economically speaking, firms and people decide to remain
in suburban areas by balancing the costs and benefits associated with leaving the
central business district and staying in suburbia. Thus, those who decide to live in
the suburbs are affluent middle-class white families (politically Republican) who can
2 0
afford the larger residential space and long commuting distance to the central
business district. These characteristics of traditional suburban communities usually
became the subject of criticism on various grounds (Baldassare, 1992; Hawkins &
Percy, 1991; Neiman, 1991; Logan, 1991).
Development of suburbia was also accelerated by federal and local
governments’ pro-suburban policies (O’Connell, 1989; Fava, 1989), and by real
estate developers’ aggressive involvement in suburban development (Keating, 1989).
Other than these reasons, technical advancements in transportation are also identified
as having contributed to this development (Banfield, 1974).
Recently, however, the dependent nature of suburbs has been challenged by
the notion of coexistence of suburban downtowns with central business districts in
urban metro areas (Muller, 1989). The term “suburban downtown” is exchangeable
with many other terms implying the same phenomenon, for example “edge city,”
“urban core,” “suburban business district (SBD)” (Harthorn & Muller, 1992;
Fishman, 1990; Garreau, 1988; Richardson, 1988).
This polycentric nature of metro areas is said to develop when the
agglomeration benefits of the central business district reached a ceiling given the
growing commuting cost to the central business district. Therefore, firms in
suburban centers would be “equally well-off” with the employment of lower cost
labor near subcenters (Ladd & Wheaton, 1991). Labor is already abundant in
suburban areas as suburbanization has continued for several decades. Under these
economic circumstances, offices, retail stores, and other urban amenities followed
manufacturing firms and people moving into suburbia. This trend is basically
recognized by a great increase in intra-suburban or inter-suburban commuting
(Stanback, 1991). Using commuting patterns and 1980 population census data,
Hughes (1993) identified Anaheim/Santa Ana/Garden Grove SMSA and three other
SMSAs as multi-centered metropolitan areas. Giuliano and Small (1991) also
identified 32 subcenters within the Los Angeles region in 1980. These new
developments made counties in suburban areas into so-called “megacounties,”
implying that these counties become as important as central cities had been in the
previous urbanization period (Herson & Bolland, 1990).
In this section, I first present demographic evidence of urbanization and
suburbanization and next a discussion on how suburbanization affects counties.
2.3.1 Evidence
Table 2.1 shows trend data on how urban America has grown. In rural areas, the
rate of population growth continued to decline until 1970, while urban areas have
grown at a far greater rate than the total population growth rate. This has led to an
almost 10 percent increase in people living in urban areas, for every twenty years,
except the 1970-1990 period. As of 1990, the urban population reached 75.2
percent nationwide. Thus, 396 urbanized areas are home to 158.3 million people
(Bureau of the Census, 1993).
2 2
In particular, during the sixty years between 1850-1910, the population
growth rate in places with over 500,000 people (mainly in the central cities)
recorded three-digit numbers every twenty years. This period also is marked by
high population growth rates in all urban territories (see the left rectangular area).
It is not only immigration from the rural hinterland but also from foreign countries
to central cities that contributed to this phenomenon. During the thirty years from
1881 to 1910, almost 15 million foreigners came to America. The magnitude of
urbanization in this period varied regionally: Mid-Atlantic and North Central cities
grew the most from this migration (Mohl, 1985). Since the aftermath of this
period, the population growth rate in urban areas and the entire country dropped
gradually. But the growth rate of urban areas still remains sizable, and it even
increased up to 54.8 percent during 1950-1970. These trends also can be identified
visually with figure 2.1.
Table 2.1: Urban and Rural Population Change (Percentage) by Urban Area Size
Change
Size
1830-
1850
1850-
1870
1870-
1890
1890-
1910
1910-
1930
1930-
1950
1950-
1970
1970-
1990a
Total Pop. 80.2 66.2 63.2 46.1 33.5 22.7 34.8 22.4
Urban Areas 214.5 179.4 123.3 90.0 64.2 39.9 54.8 25.3
over 1,000,000 - - - -
77.2 15.5 7.8 6.3
500,000-999,999 - 91.4 59.4 41.2 -22.1
over 500,000 - 213.2 176.5 157.6 80.9 27.6 19.4 -5.3
100,000-499,999 224.6 281.5 108.0 68.1 76.3 14.4 39.5 37.9
25,000-99,999 173.7 89.7 153.1 90.9 57.5 37.3 94.9 43.84
5,000-24,999 145.7 158.3 95.3 67.4 53.5 33.4 71.6 32.91
Other 150.8 243.7 109.7 63.7 26.6 205.5 66.2 14.14
2 3
Change 1830- 1850- 1870- 1890- 1910- 1930- 1950- 1970-
Size 1850 1870 1890 1910 1930 1950 1970 1990“
Rural Areas 67.4 45.9 42.5 22.4 7.7 0.8 -0.6 14.42
Urban Percent6 15.3 25.7 35.1 45.7 56.2 64.0 73.5 75.2
a. Population change rates during 1830-1970 are calculated based on the data from The Historical
Statistics o f the United States, Colonial Times to 1972. Part I. pp. 11-12 (Bureau of the Census,
1976), and the data of 1970-1990 from The Statistical Abstract of the United States: 34 (Bureau of
the Census, 1993).
b. These rates are calculated as of the end years 1990, 1970, 1950, etc.
Figure 2.1: Urbanization Chart
Population Change
250
urban area population change rate
200
rural area population change rate
150
urbanization percentage
to o
18)0-
li 50
-50
1850-
1870
1870-
1890
1890-
1910
1910-
1930
1930-
1950
1950-
1970
1970-
1990
Year Periods
Table 2.1 also clearly shows the suburbanization phenomenon (see the right
rectangular area). Since 1950, the growth of mid- and small-size urban areas is
remarkable in comparison with central cities which marked a far lower growth rate
2 4
than the national average. In particular, during the years 1970-1990, the cities with
a population of 5,000-499,999 grew most, whereas the big cities of more than
500,000 in population recorded a negative growth rate (-5.3 percent). This, coupled
with the nominal growth of central cities throughout the four decades, seems to
confirm that since 1950, people have moved from big American cities to smaller
suburban areas. This internal population movement in urban areas is one of the
main characteristics of suburbanization which led to the development o f suburban
centers in the 1970s and 1980s.
Suburbanization like urbanization varied by region. Table 2.2 provides a
more detailed picture of population change of counties of various sizes and different
regions. Among counties which have populations over 1,000,000, the counties in
the Southern and Western states continue to grow far faster than the national
average, whereas New England, Mid-Atlantic, and North Central counties are
decreasing in population or growing at a far slower pace than the national average.
During 1970-1980, the decrease in population in those three regions was salient not
only in counties with over 1,000,000 in population but also in counties between
500,000-999,999. In particular, during the years 1980-1990, in almost all
population sizes, counties in the Mid-Atlantic and North Central regions recorded
the lowest growth rates, confirming the usual arguments about population migration
from frost-belt states to sun-belt states.
25
Table 2.2: County Population Changes (Average) by Region and Size
Region/Size 1950-1960 1960-1970 1970-1980 1980-1990
US total 6.93 6.06 16.68 4.28
Over 1,000,000 counties 11.81 8.85 6.26 10.17
New England 16.36 12.80 -2.16 2.30
Mid-Atlantic -3.07 0.18 -5.39 -2.34
North Central 13.95 3.32 -9.52 -2.79
South - 40.12 27.94 19.61
West 45.45 23.95 18.99 24.02
500,000-999,999 Counties 23.59 18.90 4.81 11.69
New England 1.30 8.25 -2.67 4.81
Mid-Atlantic 30.32 17.01 -4.09 1.57
North Central 18.23 12.77 -2.32 3.10
South 28.11 15.97 10.94 14.75
West 32.81 34.55 16.31 26.66
100,000-499,999 Counties 34.19 19.40 16.23 12.52
New England 15.93 10.58 14.11 11.73
Mid-Atlantic 22.35 14.19 6.59 3.88
North Central 29.86 15.51 7.61 4.39
South 39.89 21.11 20.28 19.04
West 53.83 33.17 37.14 22.20
25,000-99,999 Counties 13.49 10.65 20.02 7.84
New England 8.37 11.59 15.94 10.71
Mid-Atlantic 11.60 8.87 10.33 3.99
North Central 15.20 10.67 11.79 2.14
South 9.39 9.41 24.19 9.80
West 31.22 17.07 33.96 17.33
5,000-24,999 Counties 0.88 2.91 16.35 1.12
New England -0.36 6.59 24.42 12.98
Mid-Atlantic 7.53 5.46 14.72 5.91
North Central -1.88 -2.28 7.37 -4.34
South 0.75 5.94 18.91 3.21
West 9.89 5.52 32.36 7.89
2 6
Region/Size 1950-1960 1960-1970 1970-1980 1980-1990
Below 4,999 Counties 1.15 -3.10 11.62 -3.55
New England -5.96 14.07 31.93 15.28
Mid-Atlantic -2.67 10.15 4.88 -25.56
North Central -3.43 -8.65 -1.87 -10.65
South 0.62 -5.81 12.73 0.73
West 5.50 3.64 26.15 2.45
a. Source: Population o f Counties by Decennial Census: 1900 to 1990 (Bureau of the Census,
1994c). The Bureau's Population Division compiled this data in October, 1994; it was obtained
from the Census Bureau FTP Site on the Internet.
b. The regional division used here follows the Census Bureau, except for the New England states.
They are separated from the North Eastern region, and North Eastern is named as Mid-Atlantic
here. This exception is meaningful because the county functions in New England are very
limited in comparison to other North Eastern counties (Schneider & Park, 1989). These regional
divisions are used throughout the dissertation unless otherwise are indicated.
2 .3 .2 How S u b u rb an izatio n Affects C ounty G overnm ent
The population increase in suburban areas instead of central cities rapidly changed
the nature of county governments. Faced with this increase, a suburban area can
incorporate itself into a municipal government or remain in the county as an
unincorporated area. But the population increase in unincorporated county areas and
in suburban cities as well accounts for the expansion of urban-type public service
provision by county governments. The increase in the population of suburban cities
indirectly contributes to a county’s functional expansion because cities frequently
contract out the production of some public services to county governments.
Recently, there have been many discussions about the advantages and disadvantages
of privatization of public services by local governments (Hirsch, 1995; Prager,
1994; Ferris & Graddy, 1991; ACIR, 1987). However, active usage of service
2 7
privatization by suburban cities occurs mainly because they can save on production
costs (Reason Foundation, 1994; 1CMA, 1992).
The easy availability of a service contract arrangement at the local level, like
the Lakewood plan in Los Angeles County, is another reason for the popularity of
privatization (Ostrom et al., 1988; Miller, 1981).
Table 2.3 show how many cities of various sizes deliver their selected
services by contract. In general, the smaller the city, the more contracting-out they
use. Differences in the contracting-out percentages are also evident among services.
Hospitals, landfills, and nursing homes have a much higher percentage of
contracting-out than libraries and fire protection services. Cities seem to provide
libraries and fire protection service by themselves. This may be because these are
among the basic charter-mandated services in cities. In contrast, a majority of cities
tend to provide hospitals and nursing home services for their citizens by contract. It
seems that because commonly, the county is the main provider of these functions in
urban areas, cities contract for services with the county rather than providing these
services directly. Regarding the high contract percentage for landfills, this is
because landfill operation has large scale economies, so contracting-out rather than
direct provision is a cost-saving option for smaller cities.
2 8
Table 2.3: Percentage of Selected Services Contracted Out by Cities
Size
Service (A) (B) (C) (D) (E)
(F)
Hospitals 46% 52% 97% 49% 73%
Landfills 26% 38% 37% 41% 42% 47%
Libraries 2% 11% 13% 10% 13% 16%
Nursing Homes 54% 73% 56% 55%
Fire Protection 1% 3% 4% 6% 10% 18%
a. Source: Census o f Governments 1987: Government Organization File Computer File] (Bureau o
the Census, 1990a). In this file, other types of services are also surveyed such as airports,
electric utilities, gas utilities, public transits, sewage, stadiums, and water supplies. Widi this
file, however, information about to whom the cities contracted out their services is not available.
b. The percentage is calculated by the number of cities reporting contracting-out divided by the
number of cities reporting the operation (directly or by contract) of that service.
c. Column titles
(A) 100,000-499,999 population (number of cities: 158)
(B) 25,000-99,999 population (845)
(C) 10,000-24999 population (1,297)
(D) 5,000-9,999 population (1,488)
(E) 1,000-4,999 population (5,599)
(F) 0-999 population (7,987)
Besides services being contracted out, as the population concentrates in a
given area, a regional government of sorts is necessary to handle new area-wide
problems crossing over the boundaries of small local jurisdictions. Central cities
already experienced this phenomenon when immense immigration from domestic
rural areas and overseas took place during the industrial age of the last century.
They tried to solve these area-wide problems by aggressively annexing coterminous
urbanizing areas (Rusk, 1993) and by improving their problem-solving capacity, that
is by implementing the progressive reform agenda explained in the previous chapter.
2 9
But the second urbanization period in suburbia after World War II is taking
place outside central cities. Thus, these problems have to be dealt with jointly by
suburban cities and the county government. They can jointly establish a new
regional government to handle the problems, but local citizens’ support for
establishment of new regional government is weak and geographically varied
(Baldassare, 1994). As a result, many scholars recognize the county government as
an important regional entity. County government is advantaged as an
intergovernmental actor in metro areas basically because of its broader geographic
coverage (Berman, 1993a; ACIR, 1992b; Parks, 1991; Marando & Reeves, 1991a;
Thomas, 1991a; ACIR, 1982). Each suburban city is too small to efficiently handle
many “capital-intensive” urban functions in metro areas like air quality, mass
transit, water, gas, and solid waste.6 County involvement in these area-wide
functions is argued to have increased the fastest in recent times (Salant, 1991).
Dodge (1990: 358) even asserts, “If counties had not existed at the beginning of the
decade, something like them would probably have been invented by the end of it to
deliver sub-regional and regional services.” These services are provided through
bargaining, negotiation, agreement, and signing of contracts between the county
government and suburban municipalities. Hence, counties, especially metro
counties, turn out to be prominent public service providers to suburban areas.
Another argument relates to the greater population increase in county
unincorporated areas than in incorporated areas during the 1980-1990 period (Benton
& Menzel, 1993; Thomas, 1991b; Marando & Reeves, 1991b). This assertion is
significant in that the county is directly responsible for the unincorporated areas.
Furthermore, if state laws are not favorable to incorporation and annexation
attempts, the increase of public services demanded by the residents in
unincorporated areas is directly related to the expansion of the county’s urban
functions and the state of county finances. There are many indications that
incorporation and annexation are more restrictive in some states. For example, a
1990 survey by ACIR (1993b) reports on many states’ procedural hurdles for
annexation, such as property owners’ petitions (34 states), city ordinance or
resolutions (32 states), public hearings (27 states), referendum and majority approval
required in a city (14 states), referendum and majority approval required in the area
to be annexed (19 states), and approval of the county governing authority (11
states). More important is that the number of states requiring public hearings,
referendum and majority approval in a city, and approval of county authority
increased from 20 to 27, from 10 to 14, and from 5 to 11 states, respectively,
comparing to the 1978 survey by Hill (1978). Rusk (1993) asserts that a recent
variation of economic growth among central cities mainly stems from annexation
difficulties caused by state limitations or nonexistence of an unincorporated area
around central cities. The ACIR survey (1993) also reports various limits imposed
on incorporation of new municipalities, for example minimum population (36
states), minimum area (17 states), minimum distance (16 states), and minimum ad
valorem tax base (6 states).
Table 2.4 shows that the county unincorporated ratio was positively related to
population change during 1970-1980 and 1980-1990. Coefficients for these periods
are about .225 and .220 and both are statistically significant at the .01 level.
Though likely, it is premature to conclude that unincorporated areas have grown
more than incorporated areas during the last two decades. Nonetheless, this result
implies that the higher the unincorporated ratio of a county as of 1987, the higher
the county’s population growth rate has been during the last two decades.
Table 2.4: Correlation Coefficient Between County Population Change and
Unincorporated Ratio Controlling for Population
Population Change Population Change
1980-1990 1970-1980
Unincorporated Ratio of .2201* .2252*
1987
* Significant at the .01 level (2-tailed).
a. Source: Census o f Governments 1987: Government Organization File IComputer File] (Bureau of
the Census, 1990a). Population o f Counties by Decennial Census: 1900 to 1990 (Bureau of the
Census, 1994).
b. Total county number included in the table is 3,042. Eight counties are given zero unincorporated
ratio because their ratios became a negative value.
c. Population controlled are 1980 population for 1980-1990 period, and 1970 population for 1970-
1980 period.
d. Unincorporated ratio of 1987 is calculated by ( a / b ) x 100 where A is the number of people in a
county’s unincorporated area and B is the total county population. To obtain A, the population of
municipalities located in the county is subtracted from the total county population.
3 2
2.4 County Functions
What counties really have done and what they are supposed to do are not always
congruent because many of the current county functions are required by federal or
state laws. Up to 60% of county expenditures are used in implementing these
mandated functions (Berman, 1993b), and counties like Orange or San Bernadino
County in California operate more than a thousand mandated programs (Salant,
1988). These mandated functions of counties mainly depict what American county
governments do, but do not show what important roles counties have come to
assume since suburbanization transformed the nature of counties. Currently counties
are expected to be an interlocal leader (Cigler, 1995; Klinger, 1991; Parks, 1991;
Todd, 1991), regional problem solver (Berman, 1993a; Salant, 1991; ACIR, 1982;
Peterson, 1977), and local financing agent (Thomas, 1991a). Verifying the extent
of these newly developed county roles with survey data is not easy because surveys
are scant and do not categorize their question items by such roles. Therefore, an
attempt is made to introduce a few previous survey results about county functions,
which provide a glimpse into such changes.
Sorting the multiplicity of county functions in a simple way is hardly possible
because functions have expanded over time and were transferred vertically or
horizontally between counties and states, or between counties and other local
governments. Nonetheless, it is not difficult to identify a prototype of county
functions from old textbooks like Porter’s (1922). Porter enumerated twelve
3 3
traditional functions which are said to have been conducted by counties from the
outset for the purpose of local self-government. Those functions include: (1)
maintenance of peace; (2) administration of justice; (3) administration of probate,
and other specialized judicial work, and keeping of vital statistics; (4) poor relief;
(5) maintenance of schools; (6) care of highways; (7) administration of tax
machinery; (8) administration of election process; (9) recording of land titles; (10)
militia organization; (11) serving as an administrative district for purposes of the
state government; and (12) miscellaneous minor functions.
As to a last category of a county’s minor functions, Porter recognized those
functions as playgrounds, hospitals, parks, and libraries. But these functions are
neither miscellaneous nor minor to counties these days. Rather they are considered
to be newer areas which county governments have undertaken since the turn of the
century. Besides these new county functions, Kneier (1930a) additionally reported
broad county assumptions of other new functions such as airports, regional plans,
new methods of county aid to agricultural interests like county fairs, county nurses,
full-time county health department, and higher education.
Table 2.5 which shows NACo’s 1971 and 1975 surveys demonstrates a
percentage increase in selected functions performed by metro counties, and these are
mostly urban functions.7 ICMA’s 1988 survey does not report the functions
performed by metro counties, but by counties with over 25,000 in population.
Thus, the data are not directly comparable to the 1971 and 1975 surveys.
3 4
Nevertheless, the survey results present the type and degree of urban services
provided by counties in urban areas. Between 1971 and 1975, metro counties
reported that they became more rapidly involved in area-wide and developmental
functions such as industrial development (28%), subdivision control (27%), solid
waste collection (22%) and disposal (28%), mass transit (23%), and airports (18%).
More important is the small differences in percentages reported between metro and
nonmetro counties in many urban functions (Lawrence & DeGrove, 1976).
Table 2.5: Selected Functions Performed by Metropolitan County
Governments, 1971, 1975 and 1988
Functions
1971
% of
No. total
1975
% o f
No. total
in
crease
m g
% of
No. total
Responding Counties No. 150 100 291 100 - 370 100
Fire Protection 47 31 139 48 17
_
50
Mental Health 104 69 240 82 13
-
65
Animal Control 75 51 204 70 19 - 71
Hospitals 61 41 137 47 6 - 24
Mass Transit 7 5 81 28 23 - 22
Airports 36 24 121 42 18
- 37
W ater Supply 31 21 90 31 10
-
28
Solid W aste Collection 31 21 124 43 22
-
23
Solid W aste Disposal 55 37 190 65 28
-
64
W ater Pollution Control 45 30 131 45 15
- -
Air Pollution Control 55 37 115 40 3
- -
Subdivision Control 77 51 226 78 27
- -
Industrial Development 32 21 143 49 28
- -
Museums 25 17 75 26 9
-
30
Libraries 86 57 216 74 17 " 63
a. Source: Lawrence and DeGrove (1976) and ICMA Special Report (1989).
b. The 1988 survey by ICMA reports the percentage of functions which counties with over 25,000
population perform.
c. - indicates that survey data is not available or is incomparable to the former surveys.
3 5
Table 2.6 below presents the percentage change of total local expenditures
spent by counties, municipalities, and special districts respectively over a twenty-
year period between 1967-1987. Generally, the importance of municipalities as
measured by the amount of expenditures is gradually decreasing in almost every
function except highways, air transportation, water transportation, and natural
resources. On the other hand, counties have increased their proportion of
expenditures in almost every function; the fastest growing functions between 1967
and 1987 were health (21%), other sanitation (solid waste disposal etc., 17%),
correction (8%), police (7%), housing renewal (7% ), water transportation (7% ),
higher education (6%), fire protection (6%), public welfare (5%), sewerage (5% ),
and libraries (5%).
Table 2.6: Percent Distribution of Local Government Expenditures by Type of
Local Government and by Function, 1966-1967, 1976-1977, and
1986-1987
1966-1967 197A-1977 1986-1987
Functions A B C A B C A B C
Local Schools 6 10 0 7 10 0 7 8 0
Higher Education 10 20 0 14 15 0 16 4 3
Highways 43 45 2 41 46 2 39 48 0
Public Welfare 67 31 0 61 38 0 72 27 0
Hospitals 44 38 18 47 28 26 43 25 32
Health 51 44 3 65 31 2 72 25 3
Police 17 78 0 22 72 0 25 70 0
Fire Protection 4 87 4 8 80 6 10 75 8
Sewerage 7 66 21 14 56 25 12 59 24
O ther Sanitation 4 89 1 12 80 1 21 68 3
Parks/Recreation 16 70 10 21 64 9 19 65 12
Housing/Renewal 0 56 44 1 55 43 7 53 39
3 6
Functions
1966-1967
A B C
1976-1977
A B C
1986-1987
A B C
Air Transport 24 50 25 21 51 27 22 50 28
Water Transport 0 30 69 3 31 65 7 36 57
Parking Facilities 0 90 7 3 88 7 4 86 9
Correction 70 30 0 77 23 0 78 22 0
Natural Resources 51 0 49 56 0 44 46 3 50
Libraries 21 64 8 28 55 1 1 26 53 14
Public Buildings 48 47 0 51 45 0 50 44 0
a. Source: Data for 1986-1987 is calculated from Census of Governments 1987: Compendium of
Government Finances. Volume 4 Number 5. Table 10 (Bureau of the Census, 1990c: 12). For
1976-1977 and 1966-1967, State and Local Roles in the Federal System (ACIR, 1982; 10-11).
ACIR’s date is also calculated from the same census survey, the Census o f Government 1967 and
1977.
b. Column titles
A. Counties, B. Municipalities, and C. Special Districts.
c. Township governments and school districts are omitted in the table because the expenditure
portion of townships is small and not covered by all states, and the expenditures of school
districts are concentrated in the first two education functions only. Therefore, the rows which do
not come up to 100 percent should be attributed to the expenditures spent by both townships and
school district governments.
This trend is congruent with the commonly stated arguments explained
earlier, confirming that counties have grown most in urban and area-wide regional
services. More important is the continuous dominance of counties in traditional
functions like corrections and public welfare. Therefore, the fact that counties gain
importance in one area (new urban functions), and sustain their dominance in
another area (traditional functions) accounts for the growth of counties after World
War II.
3 7
2.5 F orm s o f C ounty G overnm ent
Forms of county government commonly refer to three basic forms of county
governing bodies, which are called the county board of commissioners, board of
supervisors, county court, etc.8 The three basic forms are commission, county
administrator (or manager), and county executive forms. Classifying a county’s
governing structure into one of these three forms is not easy due to the variations of
the county administrator form (Cigler, 1995; Jeffery et al., 1989). In addition,
some states mandate one specific form, and other states provide optional forms
which a general law county can adopt. For example, counties in Arkansas,
Kentucky, and Tennessee must have the county executive form (Cigler, 1995).
Those counties in Alabama, Arizona, Maine, Mississippi, Nebraska, Nevada, New
Hampshire, New Mexico, North Carolina, North Dakota, and South Carolina are
allowed to have commission or county administrator forms (DeSantis, 1989). To
avoid the difficulty in classifying the county administrator form, this dissertation
follows NACo’s survey data from 1978 and 1989 for county forms.
The characteristics of the three forms are almost equivalent to those discussed
in the three municipal governing forms (Renner & DeSantis, 1993). Thus, the
commission form features plural functions as both an administrative and legislative
body with a number of elected independent officials, which are commonly called
“row officers.” This plural aspect of the commission form, though democratic, has
been criticized as “antiquated, headless, inefficient, and collegial.” In contrast, the
3 8
county executive (in a city, the mayoral position) form is said to provide “the most
visual policy making, the needed strong political leadership, and the best system of
checks and balances for the community. ” The most prominent benefit of the county
administrator form is the enhancement of professionalism, complementing the
probable weakness resulting from either the commission or executive forms. The
advantages and disadvantages of each form are discussed in detail by Koehler
(1983), Duncombe (1977), and Zeller (1975). Nonetheless, the main thesis in those
discussing county forms is that as county areas are urbanized, change from the
commission form to the county administrator form, then to a county executive form
is desirable.
Though the county reform agenda of the progressive era included changing
governing forms from commission to the county executive form, according to Zeller
(1975), “in 1960, with the exception of a few urban counties, there is no chief
executive officer for the county.” In a 1971 survey by ICMA, 80%, or 793
counties of a total 993 counties responding to the survey, had the commission form;
18%, or 184 counties, had the county administrator form; and only 2%, or 16
counties, had the county executive plan (ACIR, 1972). In NACo’s 1978 and 1989
surveys, the commission form decreased to 78.8 percent in 1978, and to 62.6
percent in 1989, whereas the administrator plan increased to 25.5 percent in 1989,
and the executive plan increased most, to 11.9 percent in 1989 (see table 2.7).
3 9
Figure 2.2 shows graphically the percentages of county governing forms in 1971,
1978, and 1989.
Table 2 .7 : Form s o f C ounty G overnm ent, 1971, 1978, a n d 1989
1971 1978 1989
% o f % o f % of
County Forms No. total No. total No. total
Total Number of Counties 993 100 3,039 100 3,042 100
Commission Form 793 80 2,398 78.8 1,902 62.6
County Administrator Form 184 18 519 17.2 776 25.5
County Executive Form 16 2 122 4.0 363 11.9
a. Source: ACIR (1972), NACo (1978), and Jeffery et al. (NACo, 1989).
b. The difference in total number of counties is because the 1971 survey was sampled, and 1978 and
1989 surveys were based on the universe of counties. However, the total number of counties in
diis table, consequently the number of each of the three forms, is slightly different from the two
NACo 1978 and 1989 reports because the NACo reports include county equivalents such as city-
county consolidated counties and independent cities which are commonly treated as municipal
governments by Bureau of die Census. In this dissertation, all county data is presented widi
“county governments” only, defined as of 1987 by Bureau of the Census unless inclusion of the
“county equivalents” is clearly indicated. Three new counties were created during the 1978-1989
period: two in Alaska, and one in New Mexico.
Figure 2 .2 : Form s o f C ounty G overnm ent
■ Commission
Form
■ County
Adminisrator
Form
■ County Executive
Form
Percentage
80 78.8
1871 1978 1989
4 0
Table 2.8 below demonstrates the change in county forms during 1978 and
1989. Among 2,398 counties with the commission form in 1978, 330 changed to
the county administrator form, and 221 to the county executive form, while 1,848
counties remained the same. The 221 counties which changed to the county
executive form include all 119 Kentucky counties and 87 Tennessee counties.
Interestingly, 54 out of 519 administrator counties in 1978 came back to the
commission form in 1989, while 19 counties moved to the executive form.
Structural change of executive counties occurred least. Only one county replaced its
form with the commission plan; the others kept the same executive form as in 1978.
Table 2.8: Change of County Forms Between 1978 and 1989
County Forms in 1978
Total
Number
Countv Forms in 1989
To CO To CA To CE
Commission Form 2,398 1,847 330 221
County Administrator Form 519 54 446 19
County Executive Form 122 1 0 121
Total County Number 3,039 1,902 776 361
a. Source: NACo (1978) and Jeffery et al. (1989).
b. This table does not provide information about a double change in county form. For example, if a
county change its form from commission to administrator and finally to executive form, it is
reported only as changing from commission to executive.
Home rule status of a county has been discussed closely with structural
changes of county forms since the progressive era. It has been argued that with
4 1
home rule, a county is able to “secure the identity and to organize and manage its
local affairs with full ability, which have been conspicuously mishandled either by
the state authorities or by the smaller local units” (Gilbertson, 1917: 145). Snider
(1952) in his mid-century review on county governments also asserted the necessity
of county home rule to overcome “a major obstacle in the path of county progress,
that is, county dependence on the state for approval of improvement measures.”
The effectiveness of county home rule is still a top priority on the county
research subject list (Menzel et al., 1992), but the actual adoption record for home
rule by American counties is far less impressive to researchers (see Salant, 1991,
1988; Jeffery, 1989; Duncombe, 1977). Though 38 states provide some type of
home rule options for their counties (Cigler, 1995), only 24 states grant structural
home rule authority to their counties, as compared to the 40 states which grant it to
their cities (ACIR, 1993b). Furthermore, among the 1,307 eligible counties, only
117 counties have home rule (Salant, 1991).
Why the counties, though granted the option, are reluctant to adopt home
rule is vividly explored by Salant’s dissertation (1988). According to her
explanation, counties are still subject to legal state controls stipulated in their state
constitution and general laws even after the charter is successfully passed. Court
interpretations are favorable to the state authority based on Dillon’s rule. State-wide
initiatives and expanding state and federal mandates make the local charters
meaningless. Furthermore, the process to adopt and amend the charter is costly and
4 2
time-consuming. Thus, it is natural for counties which are granted the charter
adoption authority to decide to remain as general law counties if the costs by far
outweigh benefits expected from home rule. County incompetence, not only around
the home rule issue but in other areas, in relation to state authority is identified by
recent surveys (Waugh & Streib, 1993; Salant, 1993).
Relating home rule status to county forms, two common questions are
answered in this section with statistical data. One question is whether home rule
counties adopted county executive or administrator forms significantly more than
non-home rule counties. The other question is whether home rule counties have
fewer number of row officer positions than non-home rule counties. The first
question is explicitly supported by scholars who argue that home rule counties are
able to organize and manage their concerns more efficiently. The reform movement
encouraging home rule charter-adoption “has traditionally called for the appointment
rather than election of county constitutional officers” (Salant & Martin, 1993). The
second question is an auxiliary argument of the first. Koehler (1983), for example,
implies that for efficiency and control purposes, executive and administrator form
counties tend to have fewer row officers.
Table 2.9 answers the first question. There is a significant association
between the home rule status of counties and their county forms in states in which
home rule adoption is allowed (Pearson’s x 1 =215.26, df=2, p < .01). The
strength in this association is not small (Spearman’s r=.3663, p< .01). Home rule
4 3
counties are more likely to have the county executive (54%) or county administrator
forms (41%) than commission form (only 5%). Contrastingly, non-home rule
counties are far less likely to have the executive plan (only 9%) in comparison with
the commission (61 %) and administrator plan (30%). These results indicate that
with home rule status, counties tend to have the executive form significantly more
often than in general law counties. Thus, it can be inferred that home rule adoption
induces counties to have more modernized (reformed) governing structures.
Table 2.9: County Forms in Home Rule States in 1989
Home Rule Status in 1989 CE
Countv Forms in 1989
CA CO
Home Rule Counties (102)
Non Home Rule Counties (1,180)
55
101
54%
9%
42 41%
354 30%
5 5%
725 61%
* Pearson’s X % =215.26, df=2, Spearman’s r=.3663, both are significant at the .01 level.
a. Source: Jeffery etal. (NACo, 1989).
b. The total number of counties in home rule states, 1,282 in this table, is different from Salant’s
1,307 eligible counties because county equivalents such as consolidated counties with cities and
independent cities, are excluded in this analysis.
Table 2.10 below answers the second question, showing that home rule
counties are more likely to have fewer row officers than non-home rule counties. In
this analysis, correlation coefficients are adjusted by controlling for population. The
correlation coefficient between the two variables is -.2063 (p<.01). In addition,
executive/administrator counties are also negatively related to the row officers
index, indicating that these counties in all states and in home rule states have fewer
4 4
row officer positions than in commission form counties. Their correlation
coefficients are -.2784 in home rule states (p< .01) and -.2666 in all states (p < .01).
These data suggest that executive/administrator counties in home rule states have a
slightly stronger probability of having fewer elected row officer positions than in
executive/administrator counties of all states.
Contrary to traditional arguments, however, executive counties in home rule
states have no relationship to the number of elected row officer positions (r=-.0478,
not significant), or have more row officer positions in all states (r=.0500, p c.O l).
70.9% of all counties and 78.9% of counties in home rule states have a moderate
number of elected row officer positions, ranging from four to eight (see the
frequency table below). The mean value of county elected row officer positions in
all states and in home rule states is 6.56 and 6.06, respectively. Regionally, New
England and Mid-Atlantic counties tend to have a smaller number of elected row
officer positions than the counties in the North Central or Western states. But the
regional mean value of elected row officer positions between the counties in all
states and in home rule states appears to be indifferent (see the mean value table).
4 5
T able 2 .1 0 : C o rrelatio n C oefficients B etw een Row officers Index a n d C ounty
H om e R ule S tatu s, E xecutive F o rm , an d E xecutive/A dm inistrator
F o rm , C ontrolling fo r P opulation
HRS EXDI1 EXDI2 EXAD1 EXAD2
Row Officers Index in 1987 -.2063* -.0478 .0500* -.2784* -.2666*
* Significant at .01 level (2-tailed).
a. Source: Row officers index in 1987 is from the Census of Governments 1987: Government
Organization File [Computer File) (Bureau of the Census, 1990a). Other column variables are
from Jeffery et al. (NACo, 1989).
b. Row officers index is the aggregated incidence, in which 1 is given to an elected row officer
position in a county, and 0 is given to a row officer position which is not elected or nonexistent.
Thus, this index number is different from the actual number of elected row officers in a county.
Included are 14 county row officers: assessor, attorney, auditor, county clerk, clerk of court,
county or probate judge, coroner, sheriff, recorder, collector, treasurer, superintendent of
schools, and surveyor.
1) Frequency of row officers index
All counties Counties in home rule states
Value Frequency Percent Value Frequency Percent
0 17 .6 0 16 1.2
1 58 1.9 1 5 .4
2 80 2.6 2 37 2.9
3 88 2.9 3 55 4.3
4 267 8.8 4 172 13.4
5 639 21.0 5 283 22.1
6 611 20.1 6 207 16.1
7 476 15.6 7 166 12.9
8 433 14.2 8 176 13.7
9 251 8.3 9 79 6.2
10 85 2.8 10 54 4.2
1 1 25 .8 1 1 21 1.6
12 11 .4 12 10 .8
13 1 .0 13 1 .1
Total 3042 100.0 Total 1282 100.0
4 6
2) Means an d standard deviations of row officers index by region
All counties Counties in home rule states
Region Mean StdDev Region Mean StdDev
U.S. Total 6.159 2.040 U.S. Total 6.064 2.188
New England 3.326 .964 New England 3.447 .828
Mid-Atlantic 4.986 1.884 Mid-Atlantic 4.986 1.884
North Central 6.858 1.805 North Central 7.103 1.959
South 5.808 2.072 South 5.153 1.862
West 6.311 1.873 West 6.226 2.196
c. First row titles:
• HRS: home rule status (home rule counties = 1; non home rule counties=0).
• EXDI: executive form (executive counties = 1: others=0).
• EXAD: executive or administrator forms (both form counties = 1; others=0).
• 1 and 2 attached to EXDI and EXAD define the variables as categorized in the home rule states
and in the all states, respectively.
The results presented here are contrary to the recent empirical study by
Salant and Martin (1993). In their study, the expected negative relationship between
four reform variables and the number of county elected row officers is not
statistically significant, though the direction of the relationships between them is
negative. This result may have occurred because their research hypotheses cannot
be manifested by the state level data used. Their research question is whether
reform items (at-large elections, home rule charter, administrator form, and reform
index adding up the previous three items) have a negative relationship with the
number of row officers i f those items are adopted by the counties. But whether a
county adopts these reform items or not is different from whether the state legally
mandates its counties to adopt. This argument is clear, considering that home rule
4 7
counties account for only 117 out of the eligible 1,307 counties. Therefore, to
verify the negative relationship between county reform items and the number of
elected row officers, the elected row officers data of an individual county are
necessary. In this sense, the significant correlation results and the row officers
index employed in this section offer considerable potential for future research.
2.6 S um m ary
The aim of this chapter is to investigate how much American county governments
have changed in their functional scope and governing structures since World War II.
Until the progressive era, county governments remained almost the same as in the
colonial period with traditional functions and the commission form. Consequently,
the county’s role as an arm of the state was distinct, though the county’s other role
as a unit of local government seemed miscellaneous and minor in those years.
Rapid suburbanization has led county governments to emerge as significant
area-wide public service providers and regional problem solvers at the local level.
During 1967-1987, counties increasingly diverted a larger proportion of total local
expenditures to area-wide urban functions like fire protection, police, housing
renewal, libraries, and higher education, while municipalities spent less on those
functions. Yet counties still remain strong, or have grown in traditional functional
areas like corrections and public welfare. In particular, the metro or urban counties
increasingly have undertaken a larger number of urban functions since the 1970s.
County governing forms, judged by their functional growth, have
modernized, but at a slower pace. Still about 63% of counties operate with the
commission form, while executive form counties comprise only about 12%.
Counties are very reluctant to adopt home rule charters, which are a major vehicle
for accelerating county modernization. Nonetheless, the data indicate that home rule
counties are more likely to have reformed governing forms (executive and
administrator forms) and a smaller number of elected county row officer positions.
4 9
Notes
1. The city is defined by the Bureau of the Census (1994a) as “the political
subdivision within which a municipal corporation has been established to provide
general local government for a specific population concentration in a defined area. ”
Boroughs (classified as county governments in Alaska), towns, villages, and cities
are classified as municipal governments. Towns in the New England states,
Minnesota, New York, and Wisconsin are classified as township governments
because they are not closely related to a specific population concentration in a
defined area.
2. Mercantilism is hard to define, but its main notion is that the public interest can
be interpreted and secured best by state regulation. Early colonial life in towns and
villages- not only economic life, but political and social life- were broadly restricted
by the mercantilistic idea and strict Protestant tenets. The decline of mercantilism
was due to labor shortages and subsequent immigration from rural areas caused by
commercial growth (Glaab & Brown, 1967). The concept of mercantilism has
proven to well befit the description of the third world’s governmental regulations in
formal economic and political sectors. These regulations in the formal sector
establish entry barriers and set forth a peculiar situation, that is, the informal sector
generates a greater portion of the gross national product than the formal sector does.
For more explanation about mercantilism and its application to Peru, see de Soto’s
The Other Path (1989).
3. Virginia adopted the English shire system first. According to Fairlie (1906: 19),
“In 1634 Virginia was divided into eight shires, and new shires or counties were
gradually organized.” This account indicates that at the outset, the shire and county
were interchangeable and coexistent, but the county soon replaced the shire. The
English shire system and the adoption processes are well described in Fairlie’s book.
A shorter version appears in the Cleveland Foundation’s report (1980), and an
overview with concise interpretation is found in Martin’s article (1993).
4. Since the reform age, the streamlined bureaucracy with its centralized decision
making structure regardless of the organization size was accepted by many people as
efficient, professional, and equitable. According to research on the effects of
fragmented local structures, this statement is not correct at this time. To see both
views, see the three ACIR reports (1993a, 1992, 1987), Dolan (1990), Ostrom, et
al. (1988), Bish & Ostrom (1975), Ostrom (1972), and CED (1970, 1966).
5 0
5. The assumption of free competition for land is unique in America in that many
other countries implement a very restrictive and centralized land use policy.
Subsequently, greater latitude of local officials in decisions about land use is
allowed. The location decision by a private sector organization is closely related to
its future return from utilization of the land (Logan & Molotch, 1987).
6. Classification of urban functions btsed on the capital-intensive and labor-
intensive aspects became commonplace after urban economists sought to find the
optimal production size, that is scale economies, of particular urban function since
the early 1960s. Conclusions of this study are: (1) capital-intensive functions like
public utilities have more extended scale economies than labor-intensive functions
like police patrol, and (2) scale economies also vary by the particular attributes of
any single public service. For example, police radio communication function has
larger scale economies than police patrol. For an introduction to various studies in
this area, see Fox et al. (1980). The three ACIR reports provide related theories
and corollaries (1987), as applied to Allegheny County (1992) and St. Louis County
(1988).
7. Urban functions refer to those functions which are “furnished to meet the urban
needs of densely populated areas rather than general needs of both urban and rural
areas” (Lawrence & DeGrove, 1976). Duncombe (1977) provided a list of urban
services commonly introduced: parking, public housing, air and water pollution
control, enforcement of building and housing codes, water supply, sewage disposal,
mass transit, the provision of metropolitan airports and park systems, and the
operation of convention centers, stadiums, and museums.
8. Titles of the elected county officers in the county governing body differ in some
states. In Arkansas and Kentucky, they are called justices of peace; in Louisiana,
police jury; and in New Jersey, freeholders. For other states, see Jeffery et al.
(1989) and Salant (1991).
C H A P T E R ffl
R E S E A R C H O N T H E IM P A C T O F L O C A L P O L IT IC A L ST R U C T U R E S
3.1 In tro d u c tio n
As stated in chapter one, research on the impact of reformed political structures is
substantial for municipal governments, but is scarce for county governments. This
situation, however, is rapidly transforming because the nature of county
governments has changed from being the administrative arms of states to being units
of local government due to an intensive suburbanization process. As explained in
chapter two, county governments have taken on broader functional responsibilities
for local affairs and adopted reformed governing structures.
Interest in the impact of reformed city political structures dates back to the
progressive era. Serious research, however, did not begin coincidentally with that
practical interest. In the early 1960s, the middle-class ethos theory advocated by
Banfield and Wilson (1963) produced much empirical research on city reformism.
That research attempted to find an answer to several related questions: (1) why do
some cities adopt reformed political structures, when other cities do not? (2) what
socioeconomic factors of a city contribute to the adoption of reformed political
structures? (3) do reformed political structures render efficient policy outcomes?
and (4) what factors other than political structures determine city policy outcomes?
This chapter attempts to answer the latter two questions because the main purpose of
5 2
this study is to examine the impact of reformed structure on policy outcomes and to
identify the determinants of policy outcomes.
This chapter first introduces the original locus of research problem by
explaining the main ideas of the ethos theory. Next, it discusses empirical studies
about the impact of reformed political structures on city governments. Then, recent
study findings about county governments are presented. Drawing hypotheses for
this dissertation is deferred to the next chapter on methodology because it is more
appropriate to present them after operationalization of research variables is
completed. However, literary arguments for unique variables in county studies are
brought forth into this chapter to make subsequent hypotheses statements more
relevant.
3.2 T he Legacy o f th e R eform A ge
3.2.1 E th o s Theory
Since the progressive era, the belief that the reformed structures of a city
government positively relate to the performance of the government has very much
influenced political activists, academic researchers, and concerned citizens. It has
been asserted that the reformed political structures correspond to the “ethos” of the
public-minded middle-class. Therefore, the subsequent policy outcomes of
municipal governments result in a lower level of budget expenditures. In other
words, a reformed structure renders lower per capita taxes and expenditures by
5 3
reducing the sensitivity of city politics to the social cleavages of a particular
community. The latter phenomenon was observed to demand a relatively high city
budget (Banfield & Wilson, 1963).
This ethos, or social cleavages theory, presupposes that among urban masses,
there are two clearly defined and different mental attitudes and political styles
regarding the public interest— what it is and how it should be treated. The first
attitude is that of the native Anglo-Saxon Protestant middle class. They have
accepted the existence of a disinterested “public interest” guiding the city as a
whole, believed in the separation of public and individual life, and emphasized civic
virtues in public life.
The other attitude is sustained by immigrants mainly from Europe where the
political culture was strongly influenced by the rule of authoritarian kings and
hierarchy. Therefore, they tend to “interpret political and civic relations chiefly in
terms of personal obligations, and place strong personal loyalties above allegiance to
abstract codes of law or morals” (Hofstadter, 1955: 9). These two distinct political
attitudes are named after Weber’s famous terminology as “legal-rational” cultures
and “traditional” political cultures, respectively (Schiesl, 1977).
The separation of class1 in a community as an independent variable has been
incorporated into many studies about city politics and the determinants of local
public expenditures. It is, however, not the same as a debate over the Anglo-Saxon
Protestant middle class vs the other European, predominantly Catholic, groups
5 4
because such a distinction has already been diluted through generational change.
Alternative variables utilized in urban studies are, for example, the percent of non
white, Catholic, or foreign-born population.
3.2 .2 R eform ed P olitical S tru ctu res
A representative structural reform package for municipal administration was
completed by the National Municipal Review when it finished its revised Model City
Charter in 1919 by incorporating the council-manager plan. Along with the council-
manager form, the Charter included clauses such as those for a merit system in civil
service, extensive home rule, and at-large election of aldermen (Mohl, 1985).2
Recently, Renner and DeSantis (1993) summarized and presented a nine-item core
agenda of the municipal reform movement. The agenda are: (1) civil service, or
merit, system; (2) competitive bidding for government contracts; (3) fair election
practices; (4) secret ballots; (5) nonpartisan elections; (6) short ballots; (7) at-large
elections; (8) council-manager plan; and (9) initiative, referendum, and recall
measures.3
From these reform items, researchers usually utilize three structural aspects:
council-manager vs other forms (mainly mayor-council forms), nonpartisan vs
partisan elections, and at-large vs district elections. The council-manager form is
believed to affect the level of internal efficiency of municipal governments through
business-like and professional management. In contrast, nonpartisan and at-large
5 5
elections are believed to directly influence the level of policy outcomes by
modifying the way to aggregate both politically and geographically divided local
interests in a community. Local nonpartisan elections reduce the sensitivity of city
politics by isolating local politics from the influence of national parties and by
protecting city interests as a whole from those of local political interest groups. At-
large representation is to reduce the influence of geographically divided local
politics, inducing city councils to be able to focus more on the city interests as a
whole.
Therefore, reformed political structure as a research variable is utilized to
describe either one-dimension governing body forms (DeSantis & Renner, 1994;
Parks, 1994; Schneider & Teske, 1993; Schneider & Park, 1989; Knoke, 1982;
Liebert, 1976; Dye & Garcia, 1978; Dye & MacManus, 1976; Alford & Scoble,
1965), or as a numerically calculated index, the “scale of reformism,” a
combination of two or three dimensions (Welch & Bledsoe, 1988; Sharp, 1986;
Farnham & Bryant, 1985; Morgan & Pelissero, 1980; Lyons, 1978, 1977; Karnig,
1975; Cole, 1971; Clark, 1968; Lineberry & Fowler, 1967; Wolfinger & Field,
1966).
5 6
3.3 R esearch on C ity R eform ed S tru ctu res
3.3.1 D evelopm ent o f R esearch
At the outset of the research on the impact of reformed structures, the main
objective was to study what social characteristics of a community lead a municipal
government to adopt a reformed structure. In other words, why does one city come
to have a reformed political structure, while another does not? An auxiliary aim was
to see if the adopted reformed structure (mainly the council-manager form with
nonpartisan and at-large elections) results in lower public expenditures by city
governments. This research, which was done widely in the 1960s and early 1970s,
was known as the study of the middle-class ethos, or social cleavages theory.4 In
this theory, community characteristics are regarded as the cause, with reformed
political structure as an intervening variable, and policy outcomes as the
consequences. The diagram below presents the supposed directional relationship.
Figure 3 .1 : Focal R esearch V ariables an d E th o s Theory
C ause Intervening V ariable C onsequence
| Community Characteristics | »
Since the very beginning o f empirical studies about ethos theory, the research
results have been a source of controversy. On the one hand, the hypothetical link
5 7
from social cleavages in a community to reformed structures and then to the policy
outcomes was verified (Lyons, 1978; Clark, 1968; Lineberry and Fowler, 1967;
Alford & Scoble, 1965). On the other hand, it was argued that such a relationship
was spurious because when city location (geographical region) was incorporated as a
control variable, the association between social cleavages and reformed structure
disappeared (Wolfinger & Field, 1966), or became very weak (Cole, 1971). In a
broadened model including region, reformed structure, and socioeconomic variables,
Dye and MacManus (1976) offered a compromise by showing that socioeconomic
variables predicted the form of government better within specific regions than
nationwide.
R egion
Why does geographic region rather than community characteristics better explain the
variance of adopted political forms? Knoke (1982) interestingly interpreted the
importance of regions on the adoption of reformed structure as due to neighborhood
effects of imitation and contagion. He wrote:
Among the theories of reform discussed at the beginning, the
spatial diffusion model held up well. The present study
specifies these regional differences as arising not from social
compositional differences of regions’ cities but from some
type o f imitation or contagion effect as represented by the
level o f neighboring regional cities previously adopting reform
government. (Knoke, 1982)
5 8
This explanation is very similar to Shannon’s argument (1991) that the “pacesetter
phenomenon” and “catch-up imperatives” are reasons why state or local
governments cannot help but compete in adopting public policy innovations.
After applying an event-history method with city data to the adoption and
change of the commission and council-manager forms during 1900-1942, Knoke
disproved the culture and conflict clash theories (similar to ethos theory). Rather
than middle class cities, “the least affluent communities were the most rapid
acceptors of both commission and managerial government.” The reason, he said, is
because communities with limited resources were more likely to solve their urban
problems through such a reformed structure.
Functional Scope
Another study of the functional scope of city government also contributed to our
understanding of ethos theory and the factors affecting local public taxation and
spending decisions (Park, 94; Schneider & Park, 1989; Schneider, 1989; Stein,
1982; Benton & Rigos, 1985; Sharp, 1986; Dye & Garcia, 1978; Liebert, 1974,
1976). According to Liebert (1974, 1976), functional inclusiveness reflects
embedded legal responsibilities, historical background settings, past political actions
of structural reform, and even ideologies and institutionalized local political culture.
In short, the functional scope of a government is a fundamental indicator of the past
5 9
and present nature o f a city in terms of determining policy outcomes. Inferring
from the important nature of functional scope, he said:
The significance of the functional scope of a government may
well be greater than that which has been assigned to more
traditional notions of political structure. When, for example,
we considered the impact of inclusiveness on municipal
expenditures, the importance of both municipal reformism and
decentralization among reputed elites dropped to a minimum.
(Liebert, 1974)
initially, Liebert’s functional indicators of inclusiveness were created for four city
functional areas: schools, welfare, hospitals, and highways. Variations of city
expenditures in these four functions are most salient among city governments.
Liebert (1974) found that the presence of the city manager form and at-large
constituency are negatively correlated to his functional indicators of schools,
welfare, and hospitals, respectively. This finding demonstrates that reformed cities
are more likely than unreformed cities to have a narrow functional scope.
Liebert’s indicators of inclusiveness were modified later by Dye and Garcia
(1978). Since then, this modified Liebert’s functional index has been utilized most
commonly in city research to control for variation of functional responsibilities
among municipal governments. The functional index of each city is usually
operationalized by the number o f functions for which more than a token amount is
spent (over one dollar per capita). A city with education, welfare, and hospitals is
coded as a “functionally comprehensive city,” while a city without these three
6 0
functions is coded as a “functionally specialized city” (Dye & Garcia, 1978). Dye
and Garcia reported that when socioeconomic and demographic variables are
combined with the functional index variable, they better explain the level of cities’
taxes and expenditures. Meanwhile, another study by Clark, Ferguson, and Shapiro
(1982) attempts to develop another index of functional performance for the purpose
of isolating legal and fiscal differences among local governments. Their functional
performance index has not been used much by later researchers who have opted to
use Liebert’s index.
Age
Like geographical region and functional scope, city age is also regarded as a historic
and contextual variable that influences the adoption of reformed political structures
and the level of policy outcomes. Though both age and region act as contextual
variables, age somewhat explains the different nature of city governments than does
the region. Liebert (1976: 51) said, “the latter (region) includes the broader
contextual effects of a regional tradition, whereas city age more narrowly identifies
traditions dominant when the city was organized.” Age of central cities or
metropolitan areas is found to be related to the functional index and total
expenditures of city and county governments (Park, 1994; Schneider & Park, 1989;
Benton & Rigos, 1985; Dye & Garcia, 1978). Age is also found to differentiate
various functional scopes of city governments (Stein, 1989), and to discriminate
61
adoption decisions between the mayor and manager forms (Dye & MacManus,
1978).
Besides these studies about reformism, a number of public finance
economists have investigated the determinants of state and local public expenditures
since Fabricant (1952) first reported on the public service expenditure function using
combined expenditures for all state and local governments. Since then, analysis of
the local public expenditure function has been expanded and made more
sophisticated by incorporating broader variables which are believed to affect median
voter preferences and to be associated with the level of local public expenditures and
tax decisions. The variables utilized in these studies are demographic, economic,
and fiscal. Well-written reviews of previous research on the determinants of public
expenditure are by Chicoine & Walzer (1985), Bahl et al. (1980), and Inman
(1979). Another study by urban economists concerns the size of governments. This
study focuses on why governments are growing over time and the determinants of
government. Various perspectives on the causes of government growth and research
results are reviewed by Mueller (1989).
3.3.2 Tw o M ethodological Issues
During verification of ethos theory and its extension to identify contextual variables,
two methodological problems have drawn much attention. One is about the
difficulty in identifying the impact of community characteristics on political
6 2
structure, mainly due to different time frames. The other is about the nature of the
reformism index.
D ifferent T im e F ram es
The methodology utilized to assess a causal relationship between socioeconomic
factors and adoption of reformed forms has been criticized because of the difficulty
in matching current political forms with the community characteristics which
probably influenced the process of adoption of the forms many years earlier
(Morgan & Pelissero, 1980; Wilson & Banfield, 1971). A survey of the form of
local governments is relatively easy, but collecting data about community
characteristics for the period of adoption is hardly possible. This is because the
nature of a community usually has changed substantially since the adoption of the
reformed structures. Relating current political forms to current socioeconomic
factors is believed to be unreliable in verifying the causal relationship between the
two variables. Thus, as shown in figure 3.2, each socioeconomic factor is treated
by recent research as an independent variable affecting policy outcomes.
R eform ism Index
The opposing results of early research on the relationships between reformed
political structures and policy outcomes aroused questions about the nature of the
reformism index. Is the reformism scale additive or not? Summarizing problems in
6 3
the early studies with the reform scale, Bryant (1976) identified two different
dimensions of the reformism scale by using factor analysis. He stated that the first
factor is clustered exclusively around the allocation of executive power (the
executive dimension-governing form, mayor’s veto power, and method of mayoral
election). The second factor is clustered around council modification (legislative
dimension-number of councilmen, council-election method, and the length o f term
for councilmen). In other words, the first factor is mainly related to the form of the
governing body (mayor-council, council-manager, or commission), and the second
factor is related to the representation method (at-large vs district, and partisan vs
nonpartisan elections). Applying these two different reformism factors, Farnham
and Bryant (1985) concluded that: (1) the full (or original) reform index is more
strongly associated with region and population size than with community
socioeconomic characteristics; and (2) separating the index into executive and
legislative dimensions provides more significant independent variables than the full
reform index.
3.4 Research on County Governments
3.4.1 Recent County Studies
One thing to be clarified is that the county executive form is regarded as a
“reformed form” in county studies. This different treatment of the county executive
form makes sense, considering that what was most criticized in the county reform
6 4
movement was the headless aspect of the commission form. Nonetheless, the county
executive form poses difficulties in formulating the “reformism index” in county
studies. In the next chapter on methodology, both the county executive and
administrator forms are operationalized for purpose of formulating a county
structural reform index and applied separately to the cause-effect regression models.
Scant empirical research has been done on the impact of county governing
forms (DeSantis & Renner, 1994; Park, 1994; Duncombe etal., 1992; Schneider &
Park, 1989; Benton & Rigos, 1985). A consistent research result is that both the
county executive and administrator forms are positively related to higher levels of
total county expenditures (DeSantis & Renner, 1994; Park, 1994; Schneider & Park,
1989). For individual county functions the administrator form is found to produce
higher per capita expenditures for county welfare, solid waste, and police functions,
and the county executive form is found to render higher expenditures for solid waste
and police (Park, 1994). Another study found that the presence of a county
administrator increases county functional scope as measured by a number of reported
county functions (Benton & Menzel, 1991). These research results seem contrary to
the common notion of city reformism. County researchers interpret county
structural reform as adopted “ in order to more effectively control and manage their
expanding service delivery role” (DeSantis & Renner, 1994; Schneider & Park,
1989). This interpretation confirms the notion that “form follows function” in the
American county system (Cassella, 1971).
6 5
3.4.2 Variables Unique to County Studies
Several variables are unique and influential to county governments, and it is argued
that they must be included in county research. They are: (1) unincorporated ratio;
(2) county home rule status; (3) number of row officers; (4) metro and nonmetro (or
urban and rural) status; (5) number of municipalities and special districts; (6)
presence of central cities; and (7) state mandates for various county functional
responsibility and finances. Items one to four are explained in chapter two, and the
other three are the focus of this section.
State Mandates
State mandates about functional responsibilities and finances also affect city
expenditures, but they may have stronger effects on county expenditures than on city
expenditures. This is because: (1) a state’s legal distribution of local public
functions is different between the county and city governments within that state; and
(2) county governments are the primary local units for implementing state affairs
equally at the local level. Therefore, the more states mandate a specific county
function, the more county governments have to increase their spending for it. Park
(1994) found that state mandates have a significant effect on county spendings for
welfare, highways, solid waste, and education. A problem in using the state
functional mandate variable is in obtaining the number of state mandates; this survey
6 6
was done only once by ACIR (1978). Though the data of ACIR’s survey is a little
outdated, it is included as the measure of state functional mandates for this study.
Unlike functional mandates, state fiscal mandates are easily available from a
recent ACIR report (1993b). The effects of state fiscal mandates on local
expenditures are self-evident. The more a state issues tax and debt limitations, the
less resources the local government has to spend. This fiscal state mandate variable
is used by some previous researchers on city and county governments, and their
results show significant negative effects on county expenditures (Park, 1994) and tax
revenue of cities (Sharp, 1986). Some other results show either positive effects on
suburban city expenditures (Schneider, 1989), or no relationship with central city
expenditures (Park, 1994). Different effects of state fiscal limitations on county and
city expenditures indicate that cities are more flexible than counties in compensating
for their restricted tax resources with other revenue resources like user charges.
Number of Municipalities
The number of municipalities is utilized in some empirical research mainly to
explain the impact of governmental fragmentation on the size of governments. Since
the progressive era, governmental fragmentation in local areas has been criticized as
the cause of various local problems. Public choice theorists, however, see local
fragmentation as an appropriate arrangement by people’s structural choices for
enhancing political responsiveness and economic efficiency. Thus, debate over the
6 7
advantages of the two options-fragmentation and consolidation-has taken place for
a long time (ACIR, 1993a, 1992b, 1988, 1987; Rusk, 1993; Lyons et al., 1992;
Edwards & Bolland, 1991; Dolan, 1990; Phares, 1989; Zax, 1989; Benton &
Gamble, 1984; Marando, 1979; Gustely, 1977; Bish & Ostrom, 1975; Ostrom,
1972; CED, 1970; Ostrom et al., 1961). To summarize previous empirical
research, consolidation does not result in lower expenditures and less taxation
(Benton & Gamble, 1984; Gustely, 1977), and the number of governments per
1,000 capita decreases the county’s share of own source revenues (Zax, 1989).
But according to Dolan (1990), the impact of fragmentation varies by how
fragmentation is measured. Fragmentation measured by the number of governments
(absolute fragmentation) or by governments per 1,000 capita (relative fragmentation)
decreases the cost of local governments, but fragmentation measured by fiscal
dispersion (fiscal dispersion fragmentation) increases the cost of local governments.
These results show that fragmentation is related to the level of local fiscal outcomes.
The next chapter incorporates one measurement of fragmentation— the number of
municipal governments in a county area-to examine the impact of fragmentation on
county policy outcomes. Dolan’s fiscal dispersion fragmentation measurement is not
applicable because counties are the unit of analysis. Therefore dispersion cannot be
calculated for a single unit in any one case. Dolan obtained fiscal dispersion among
all local governments in a county area.
6 8
Presence of Central Cities
The presence of central cities in a county area is also considered as a significant
variable for county studies because local governments cooperate with each other on
the one hand, but on the other hand, they compete with each other (Kenyon &
Kincaid, 1991). On the local cooperative side, county governments will not take
functional responsibility for some functions which central cities already offer. But
on the competitive side, county governments will provide some functions like
economic developmental functions, even though central cities also provide them.
According to Park’s research (1994) on central cities and counties in metropolitan
areas, expenditures of central cities increase county expenditures. For individual
functions, expenditures of central cities increase county expenditures in housing,
solid waste, police, and education. These research results indicate that the
competition effect between central cities and county governments is stronger than the
cooperation effect in metropolitan areas.
Special D istricts
Special districts are usually created for three reasons: (1) to secure broader economic
scales, (2) to take politics out of governments, and (3) to avoid state tax and debt
limitations on local governments (Herson & Bolland, 1990; Chicoine & Walzer,
1985; ACIR, 1982). Rules for creation of a special district vary by each state legal
system. However, once a special district is established in a county area, it either
6 9
takes responsibility for newly arising service demands or for some local services
which existing local governments have provided. Therefore, it can be said
tentatively that the number of special districts in a county area is negatively related
to the level of county expenditures. It is hard to identify which special districts are
particularly related to county governments because their functions and geographical
coverage are diversified. The next chapter utilizes as an independent variable the
number of special districts in a county area subtracted by the number of special
districts covering the area as exactly as the boundaries of municipalities and
township governments. This measure is not perfect, but may exclude special
districts which only function within these sub-county jurisdictions.
3.5 Summary
This chapter explained the focal research variables related to the impact of political
structure by reviewing literature on ethos theory, determinants of local public
expenditures, and recent county studies. Figure 3.2 below, albeit not complete,
presents a graphical model to show how different the causal path of current studies is
from that of ethos theory (figure 3.1). In this figure, socioeconomic factors are
treated as independent variables which determine the level of policy outcomes
because of the difficulty in matching time frames.
7 0
Figure 3.2: Enlarged Relationships among Policy Outcomes and Their
Determinants
■ " " » Causal Relations
« ► Associational Relations
Political Structure
Functional Scope
Functional Scope
Unincorporated Area
Demographic Variables
Number of Governments
Home Rule
(Row Officers)
Socioeconomic Variables
State Mandates
(Fiscal)
Fiscal Variables
(Grant-in-aid)
Policy Outcome
(Level of Taxation/Expenditure)
7 1
Notes
1. The separation of classes based on economic status does not seem to be well
accepted or articulated in American society and its social theories. Rather, culture
(ethnic) and religion-oriented division has been the more standard criterion
(Banfield, 1974, chapter 2). Recently, Murray (1993) described behavioral
characteristics of the “urban underclass.” The urban underclass, he stated, are the
people who are trapped in a cycle of crime, drugs, poverty, illiteracy, and welfare
or homelessness. He stressed, as important, the “white underclass” which has
begun to arise.
2. The council-manager plan was not supported by the National Municipal Review
from the outset. In its first Model City Charter written in 1898, a strong mayor-
council plan was recommended. This was because the council-manager plan was not
yet devised (it was first implemented in 1913), and the framers of the Charter
thought it most urgent to reduce the influence of political machines by centralizing
and strengthening municipal executive authority. During the brief transition from a
strong mayor-council form to the council-manager plan, the commission form of
municipal government became popular (for details, see Mohl, 1985, chapter 6).
However, no single form of city governing body has dominated city
structural reorganization plans. Structures are adopted based on size of population,
geographic region, or metro status, for example. To identify such a diversity, the
“Municipal Form of Government Survey” by ICMA in 1991 is a good source. Its
survey results are reported and analyzed by Renner and DeSantis (1993) in The
Municipal Year Book 1993.
3. According to Martin (1993), the reform agenda for county government in the
progressive era was somewhat different from that recommended for city
governments. County reforms called for: (1) the abolition of the fee system; (2) a
return to appointment, rather than election, of most county officials; (3) the
professionalization of county government; and 4) the granting of county home rule.
4. Regarding the middle-class ethos theory, Wilson and Banfield (1964; 1971)
argued that the public-regarding middle-class did not always prefer lower taxation
and expenditures to higher ones. Rather they tended to cast a favorable vote for
public projects and necessary financial tools (high taxes and new debt issues) if it
benefits the community as a whole. Because this observation is far different from
the original argument of middle-class ethos theory, these authors suggested some
modifications in their 1971 article, saying that the “middle-class” Anglo-Saxon
Protestants in City Politics would have been better described as “upper-class” or
“upper-middle-class. ”
7 2
The opposite effects of middle-class ethos on local policy decisions (mainly
financial decisions) made two researchers declare later, “thus, the issue for local
government remains unsettled” (Morgan & Pelissero, 1980). Nonetheless, most
studies are based on the hypothesis that a reformed structure is negatively related to
the level of local public expenditure.
7 3
CHAPTER IV
RESEARCH DESIGN AND DATA SOURCES
4.1 Introduction
This chapter explains the research design, data sources, operationalization of
research variables, and statistical techniques. After the operationalization
discussion, research hypotheses are formulated only for reformed structural variables
and dependent variables. Then, a summary table for expected directions among
research variables is presented.
This study basically employs a cross-sectional and cause-effect design. Data
are obtained from computer files and a CD-ROM provided by the Bureau of the
Census and surveys by the National Association of Counties (NACo) and the
Advisory Commission on Intergovernmental Relations (ACIR). Policy outcome
variables (dependent variables), financial, functional, and socioeconomic variables
(independent variables) are measured by continuous bases; while county political
structures and state mandates variables (also independent variables) are scaled, or
coded as dummy variables. Three dependent variables are examined separately with
independent variable sets. With this data, regression techniques are utilized to
examine cause-effect relations between various independent variables and dependent
policy outcome variables.
7 4
4.2 Research Design
4.2.1 Cross-Sectional Design
Cross-sectional design is advantageous in gathering a large number of cases and
related research variables at a single point of time. This type of design is
appropriate to either survey, secondary data analysis, or both. It is also said to be
desirable for the researcher to attempt to procure universe data (Miller, 1991).
Once data is obtained, this design can analyze it for numerous important
relationships among variables. In spite of these advantages, expenses are very large
if a survey has to be conducted. For secondary data, if data sources are diversified,
it is hard to obtain universe data and all related variables because the time difference
among data sources makes them unmatchable.
Universe county data is not used in this study due to this problem. To
examine causal relations, independent variables should precede the dependent
variable or be surveyed in the same time frame. The Census o f Governments 1987
was surveyed earlier than NACo’s structural survey in 1989 and ACIR’s state
mandates survey in 1990. Thus, the Annual Survey o f Governments 1990 is
employed instead of the 1987 universe county survey. Implications of not using
universe county data are explained in a later section by comparing the data set used
with universe county data.
7 5
4.2.2 Model of the Study
The model employed in this chapter is composed of policy outcome variables and
their determinants. Thus, simply speaking, this study aims to establish and verify
the determinant model of county expenditures. Determinants refer to the research
variables, which have been identified by previous research introduced in prior
chapters.
There are many competing perspectives about what causes a social action,
but social scientists generally explain an individual’s or an organization’s actions by
the operational mix of its desire and opportunity set (Elster, 1989). Desire refers to
the objective of a social actor, which guides him in the undertaking of an action in a
certain way. The opportunity set is the boundary, or zone within which an actor can
make a feasible decision. This set is known to be defined by formal and informal
rules derived from physical, social, and legal environments. Social and legal rules
are called “humanly devised constraints,” that is, social institutions (North, 1994,
1990; Ferris & Tang, 1993).
Whether human beings and organizations have an ultimate and dominant
desire (preference or objective), and whether they act rationally with the guidance of
the dominant desire are very controversial in the social sciences (North, 1994, 1990;
March, 1992; Ostrom, 1991; Simon, 1985; Pfeffer, 1982; Ramos, 1981; Burrell &
Mogan, 1979; Bernstein, 1978). These arguments are important and valuable to the
progress of the social sciences. But it is not necessary to begin with these arguments
7 6
because the nature of county governments can be narrowed down easily. A county
government is basically a local government with general purposes, operating within
the American federal system. It is not an individual or private organization, and
more importantly it is not a national or state government.
Thus, the determinant model in this study assumes two things: (1) county
governments try to achieve efficiency as the primary objective in accomplishing
county public functions; and (2) various conditions from social, legal, and fiscal
situations constrain the full attainment of efficiency in policy outcomes. Presented
below is an equation of the model.
^
Yj = Policy Outcome Variables
a, = Intercepts
Pi = Regression Coefficients of Determinant Variables
Xj = Determinant Variables
et = Unexplained Error Terms
The first assumption is advantageous in overcoming the difficulty of defining
a public service demand function for “a representative person in a community.”
The determinant model utilized by urban economists usually assumes the median
voter as the representative person in a community, city or other local government.
However, this median voter model is known to have serious problems regarding an
appropriate indicator of the median voter preference (median income or other? ) and
multiple policy ballot measures (Mueller, 1989; Fisher, 1988; Inman, 1979). The
second assumption is also useful for circumventing the problem of identifying
7 7
internal political factors of counties. The model in this chapter with these two
assumptions may fit with the “statistical” approaches categorized by Bahl et al.
(1980).1 Bahl et al. (1980) say about these approaches:
Statistical approaches identify “plausible” determinants of
government expenditure variations....they then attempt to
verify these plausible relations with statistical significance.
These studies are formulated on an a priori basis without
explicit discussion of a utility function, budget constraint,
production function, or decision-making mechanism. (71)
4.3 Data Sources
4.3.1 List of Data Sources
Data sources of the study are listed as follows:
(1) Bureau of the Census, 1992. Annual Survey o f Governments 1990: Finance
Statistics [Computer File].
(2) Bureau of the Census, 1990. Census o f Governments 1987: Government
Organization File [Computer File).
(3) Bureau of the Census, 1994. USA Counties 1994: A Statistical Abstract
Supplement [CD-Rom Version].
(4) Bureau of the Census, 1994. Population o f Counties by Decennial Census: 1900
to 1990. Compiled by The Bureau’s Population Division.
(5) National Association of Counties, 1989. County Government Structure: A State
by State Report. Surveyed by Blake R. Jeffery, Tanis J. Salant, and Alan L.
Boroshok.
(6) Advisory Commission on Intergovernmental Relations, 1993. State Laws
Governing Local Government Structure and Administration.
(7) Advisory Commission on Intergovernmental Relations, 1978. State Mandating of
Local Expenditure.
7 8
4.3.2 Data Characteristics
This study cannot utilize universe county data surveyed by the Bureau of the Census
in 1987 because of the time difference among research data sources. Therefore, as
listed above, the main data set is from the Annual Survey o f Governments 1990
(Bureau of the Census, 1992). The number of counties included in this set is 2,246,
which represents 74 percent of all 3,042 counties. The Bureau of the Census stated
why it selected these counties as follows:
All county governments with a 1986 population of 75,000 or
more; other counties are included with certainty in the sample
because of their importance in special categories, or because
of the relative magnitude of their financial activity. Units not
included with certainty are sampled with varying probabilities
within an area, type of government, and size ordering, (from
technical documentation supplement of the 1990 annual
computer file)
This description clearly shows that the 2,246 counties are not randomly
sampled. It may cast a difficulty in generalizing research results to the county
universe. But further exploration about the data set found that data
representativeness is still well secured except for less-populated counties, which
have been dropped from the universe data. Nonetheless, the less-populated
counties, although excluded more than highly-populated counties, remain substantial
in number in this data set, so it may be safe to say that research results of this
dissertation can be generalized as being about universal counties.
7 9
Table 4.1: County Distribution by Population Category
Population
2,246 Counties
Number Percent
3,042 Counties
Number Percent
below 5,000 134 6.0% 274 9.0%
5,000-19,999 720 32.1% 1,136 37.3%
20,000-49,999 634 28.2% 827 27.8%
50,000-99,999 360 16.0% 377 12.7%
100,000-499,999 323 14.4% 323 10.6%
over 500,000 75 3.3% 75 2.5%
a. Sources: The Census o f Governments 1987 (Bureau of the Census, 1990b) and The Annual Survey
of Governments 1990 (Bureau of the Census, 1992).
b. Population category is based on the 1987 population.
Table 4.2 shows the number and percentage of counties by region, governing
forms, home rule, and metropolitan status. All regions lost some number of
counties, but New England, Mid-Atlantic, and Western regions slightly gained in
percentages of number distribution. Nonetheless their percentage inflation does not
appear serious. Commission form and non metro counties are reduced the most.
Table 4.2: Comparison Between the Cases Included in the Study and the
Universe of County Data
Categories
2,246 Counties
Number Percent
3,042 Counties
Number Percent
Region
New England 43 1.9% 52 1.7%
Mid-Atlantic 124 5.5% 144 4.7%
North Central 756 33.7% 1,051 34.5%
South 989 44.0% 1,375 45.2%
West 334 14.9% 420 13.8%
Governing Forms
Commission 1,266 56.4% 1,902 62.5%
8 0
Categories
2,246 Counties
Number Percent
3,042 Counties
Number Percent
Administrator 701 31.2% 111 25.5%
Executive 279 12.4% 363 11.9%
Home Rule Status
Home Rule County 100 4.5% 102 3.4%
Non Home Rule County 2,146 95.5% 2,940 96.6%
Metropolitan Status
Metro County 676 30.1% 735 24.2%
Non Metro County 1,570 69.9% 2,307 75.8%
Sources: same as table 4.1.
4.4 Policy Outcome Variables-Dependent Variables
Using expenditure data as the proxy o f policy outcomes is not ideal because of the
difficulty in conceptualizing and quantifying actual public sector outcomes (Chicoine
& Walzer, 1985). Though some researchers attempt to use other data as a
measurement of public policy outcomes, for example public employment data (Bahl
etal., 1980; Cole, 1971; Wolfinger & Field, 1966), expenditures are widely used as
the closest measurement of public policy outcomes. In this section, three policy
outcome variables are calculated from the county fiscal survey data.
4.4.1 Total Expenditures Per Capita
County expenditures are aggregated into various meaningful categories. The most
used-category is total expenditure per capita because it presents overall involvement
of local governments (DeSantis & Renner, 1994; Park, 1994; Stein, 1989;
Schneider & Park, 1989; Mogan & Pelissero, 1980). Reformed counties are
expected to make an effort to lower total expenditure per capita.
4.4.2 Total Tax Revenues Per Capita
Revenue-side variables as a measurement of policy outcome are employed without
consistency. Research using city data employed total taxes per capita (Sharp, 1986;
Dye & Garcia, 1978), total revenue per capita (Mogan & Pelissero, 1980),
tax/income ratio (Cole, 1972; Lineberry & Fowler, 1967), tax bill computed by
multiplying the effective tax rate by the median home value (Schneider, 1989), or
per capita taxes as a percentage of per capita personal income (Stein, 1989). Total
tax revenue per capita is utilized here because reformed counties are expected to
attempt to lower this measure.
4.4.3 Total Expenditures Per Capita for Urban Functions
Division into traditional and urban functions is meaningful at the county level
because the county structural reform movement has been brought about as county
functions, particularly urban functions, have expanded at a faster pace (see chapter
two). Though expansion of county urban functions has influenced the reform of
county governments, the reformed political structure of county governments, once
adopted, is expected to implement these functions efficiently. Traditional functions
8 2
are commonly mandated by states for county governments to provide them, while
urban functions are asserted to be optional rather than mandatory to county
governments (Duncombe, 1977). Thus, it is hypothesized that the impact of
reformed county political structures is greater on urban functions than on traditional
functions.
Only urban county functions are aggregated for the dependent policy
outcome variable in this dissertation. Practically, it is not easy to isolate county
urban functions from the survey data on county finances. As introduced in chapter
two, Duncombe (1977) listed twenty county urban functions. Based on his list, the
following eleven county functions are aggregated on the basis of per capita
expenditures, which are surveyed by the Bureau of the Census. They are: air
transportation; higher education; housing and community development; libraries;
parking; parks and recreation; sewerage; sanitation other than sewerage; transit
subsidies; utilities; and protective inspection and regulation. Protective inspection
and regulation is included because it aggregates county expenditures for regulation
of private enterprise for the protection of the public and inspection of hazardous
activities.
8 3
4.4.4 Summary of Dependent Variables
Table 4.3 below shows mean values of policy outcome variables tabulated by
region. As explained in chapter two, counties in New England states are functioning
minimally, while counties in Southern and Western states take much broader
responsibilities. Interestingly, counties in the Mid-Atlantic states spend the most for
urban functions. Mid-Atlantic and North Central counties spend almost the same
amount in total per capita total expenditures, but their difference in per capita
expenditures for urban functions is great. This may be because the Mid-Atlantic
region is far more urbanized than the North Central region, or because states in this
region mandate these functions more than the states in the latter region.
Table 4.3: Mean Values of Per Capita Total Expenditure, Tax Revenues, and
Urban Expenditures by Region
Region Expenditures Tax Revenues Urban Expenditures
All Counties 649.3 202.6 43.2
New England 94.3 21.8 3.5
Mid-Atlantic 578.5 154.4 69.9
North Central 524.2 148.7 16.6
South 704.1 234.2 56.7
West 874.1 274.2 58.9
a. Sources: The Annual Survey o f Governments 1990 (Bureau of the Census, 1992).
b. Nine Alaska boroughs are excluded in this calculation because they inflate the mean values of
Western region. Their level of three measures is extremely high due to huge revenues from the
industries of natural resources. Regression models utilized in the next chapter also tend to exclude
them during the procedures of outliers deletion.
8 4
4.5 Reformed Political Structures-Main Independent Variables
Two measurements about county reformed political structures are presented in this
section: one is the county structural reform index formulated with the county
executive/administrator form, county home rule status, and the number of county
row officers position. The other is a separate application o f three county governing
forms and district/at-large election o f county supervisors.
4.5.1 County Structural Reform Index
Initially, this chapter attempted to establish a single reformism index as employed by
research on municipalities (Sharp, 1986; Lyon, 1978; Clark, 1968; Lineberry &
Fowler, 1967;). However, the city reformism index is criticized as not having an
additive nature (Bryant, 1976), and information about county nonpartisan practice is
not available from existing county surveys.2 Thus, a new county structural reform
index is formulated instead. For county reform items, five aspects of county reform
are chosen as listed in table 4.4.
Table 4.4: County Reform Items and Coding Scheme
Reform Items Unreformed (Code=0) Reformed (Code=1)
Form commission executive/administrator
Election district at-large
Home Rule non home rule home rule
No. of Supervisors large small
No. of Elected Row Officer large small
Positions
8 5
a. Sources: The Census o f Governments 1987 (Bureau of the Census, 1990b) and County Government
Structure: A State by State Report (NACo, 1989).
b. Mean values of home rule counties are used as the base to dichotomize the number of supervisors
(nine persons) and elected row officer positions (four positions). Home rule counties are said to
achieve any reform item they want, thus they can be the reference for classification of reform
items of other counties. Also for the election method, the figure of 42 percent is used to divide
counties into reformed and unreformed structures because home rule counties elect on average 42
percent of their supervisors by an at-large constituency.
With these five county reform items, this study employs principal
components factor analysis and Cronbach’s scale reliability test to identify similar
items and to check their internal consistency. Table 4.5 shows that two different
factors exist among the five items. Form, home rule, and number of elected row
officer positions are clustered at factor 1, and number of supervisors and election
items are clustered at factor 2. This result is very similar to Bryant’s analysis (1976)
in that in the city reformism index, executive and legislative aspects are distinct
enough not to use them together (see chapter three). Only reform items clustered at
factor 1 are used for formulation of the county structural reform index because the
reliability coefficient of factor 2 items is far smaller than for factor 1 items
(Cronbach a = .35 for factor 2 items and Cronbach a =.52 for factor 1 items). Even
the alpha=.52 level of factor 1 items is not good from the standpoint of a strict scale
reliability standard. This low rate among county reform items indicates that county
governments do not adopt these reform agenda as a single package. This means that
county reform adoption is still at the state’s discretion, not at the counties’ will.
8 6
City data show better than a .80 consistency rate among city reform items (Bryant,
1976).
Table 4.5: Varimax Rotated Factor Matrix Among Reform Items
Reform Items Factor I (Structural) Factor (Electoral)
Form .740 -.215
Home Rule .664 -.009
Row Officer .732
,332
Election -.119 .750
Supervisors .018 .765
Source: same as table 4.4.
The county reform index is formulated by adding scores of whether each
factor 1 item exists. For example, counties with the commission form, non home
rule, and a large number of elected row officer positions get zero points, while
counties with an executive or administrator form, home rule, and a small number of
elected row officer positions get three points. Further, counties with any one of the
reform items receive one point, and counties with any two items have two points.
Distribution of counties by the reform index is shown in table 4.6 below. According
to the table, the majority of counties (51 percent) are still among “the least reformed
counties.” There are only 66 among the most reformed counties which get three
points, comprising less than three percent of the total.
Table 4.6: Distribution of Counties by Reform Index
8 7
Reform Index Counties Percent
0 1,146 51.0%
1 721 32.2%
2 313 13.9%
3 66 2.9%
source: same as the table 4.4.
4.5.2 Separate Applications of County Reform Items
County governing forms and election types are coded as dummy variables as
follows. For county forms, the commission form is the reference for this coding.
(1) county executive form is coded 1, and other forms are coded 0.
(2) county administrator form is coded 1, and others are coded 0.
(3) at-large method is coded 1, and district election is coded 0.
4.6 Determinant Variables
4.6.1 Demographic and Socioeconomic Variables
Socioeconomic variables employed in previous city research are too diversified to
list them all consistently. The studies to verify the ethos theory and identify
discriminant factors for the adoption of city reformed political structures have
incorporated typically racial, ethnic, religious, and economic variables (Farnham &
Bryant, 1985; Dye & Garcia, 1978; Lyon, 1978; Lineberry & Fowler, 1967).
This study, however, has chosen an eight-variable set and argues that they
“measure the social condition and range of change of counties” (Marando & Reeves,
1991a). These variables are not theoretically consistent, but they demonstrate “three
important dimensions of counties”: (1) urbanization; (2) population change and
service demand confronting a county; and (3) wealth. Among the eight variables,
county land size is deleted because Marando and Reeves (1991a) found no
significant correlation with the county problems reported by county officials, and it
has not been used in other research. The other two measures of urbanization,
population and population density, still remain in the set because they are frequently
incorporated in other county research (DeSantis & Renner, 1994; Benton & Menzel,
1991). These seven variables are obtained from USA Counties 1994 (Bureau of the
Census, 1994b), and explained below:
Urbanization
• population size, 1990
• population density, 1990- population per square mile
Population change and service demand
• population growth- percent change of county population, 1980 to 1990
• housing unit change- percent change of total housing unit, 1980 to 1990
Wealth
• per capita personal income, 1990
• median home value, 1990
• percent persons below poverty level, 1989
8 9
4.6.2 County Functional Scope
An adjusted Liebert’s functional scope (see chapter three) is applied to the twenty-
one county functions surveyed by the Bureau of the Census (1992b). A county
which spends over one dollar per capita for a function scores one point. The score
for all functions are aggregated at the individual county level. Included in this
calculation are twenty-one county functions: air transportation; correction;
education; financial administration; fire protection; health; highways; hospital;
housing and community development; judicial and legal; libraries; natural resources;
parking facilities; parks and recreation; police protection; protective inspection and
regulation; public welfare; sanitation other than sewerage; sewerage; transit
subsidies; and utilities.
4.6.3 State Mandates
Two state mandate measures are incorporated in this study. The first is state
functional mandates which counties have to provide. The other is state fiscal
mandates which limit counties’ full capacity to raise resources.
State Functional Mandates
Previous studies employed a number of state program mandates surveyed by ACIR
in 1978, titled State Mandating of Local Expenditure (Park, 1994; Stein, 1982).
This report identifies 77 specific local programs mandated by state governments.
9 0
Aside from this survey, no other survey on state mandates has yet been found.
Thus, the number of state program mandates reported by ACIR’s survey is used in
this study. One thing to be remembered is that the explanation about the results
should be taken cautiously because changes or additions to state mandates after the
survey are unknown. Nonetheless, the results may be useful unless one can be sure
that state policy on local governments has not been consistent during that period.
State Fiscal Mandates
ACIR’s 1990 survey (1993b) on state laws regulating local governments structure
and administration is used for this study as two state fiscal mandate variables to
county governments. The ACIR report is a follow-up of Hill’s 1978 survey on the
same aspects of state laws. Previous studies employed either Hill’s survey results
(Schneider, 1989; Stein, 1982) or individual state fiscal limitation measures such as
debt limit (Sharp 1986) and property tax limit (Benton & Rigos, 1985).
This study has attempted to utilize Stein’s elaboration (1982) of five scale
scores for the state fiscal mandate variable. Stein added scores of the five fiscal
items which are coded as 1 and 0. The five state fiscal items are:
(1) Referendum for bond issues required
(2) Maximum life of bond issues regulated by state
(3) Interest ceiling on bond issues regulated by state
(4) Property tax limits imposed by state
(5) Short-term borrowing allowed.
But the reliability test for Stein’s fiscal mandate scale is very discouraging
(Cronbach’s Apha=.18). Thus, only state property tax and debt limitations are
separately included as state mandate variables.
4.6.4 Intergovernmental Revenues
The availability of intergovernmental aid is positively related to the level o f county
expenditures. Many functions provided by county governments are dependent on
state financial aid in the areas of redistributive social services, judicial systems and
corrections, and public safety. County governments received 36.4 percent o f their
revenues from intergovernmental aid in 1987-1988, while municipalities received
28.3 percent from intergovernmental aid in the same year period (Ebel, 1991).
State aid to county governments was 32.4 percent, whereas federal aid to them was
only 2.5 percent.
Though state and federal aids are surely important to the level o f county
expenditures and revenues, they are omitted in later regression models because of
two reasons. First, these variables, particularly state aid, tend to make the models
curvilinear, not linear. Data transformation does not help much in solving this
violation of the linearity assumption. Second, they tend to render many other
control variables statistically insignificant. This study not only focuses on the
impact of reformed structures, but aims to examine the influence of other control
variables on county policy outcomes. This phenomenon occurs because the
9 2
intergovernmental aid variable is too influential on county expenditures and
revenues. For example, county tax revenues per capita are correlated at .986 with
state aid per capita.
4.6.5 Region
Five regional divisions are incorporated in this study (see chapter two). The West is
the reference region for the following coding scheme.
(1) New England is 1, and other regions are 0.
(2) Mid-Atlantic is 1, and other regions are 0.
(3) North Central is 1, and other regions are 0.
(4) South is 1, and other regions are 0.
4.6.6 Age
County age is calculated as the number of decades after 1990 in which a county has
gained a population of 50,000 or more. Stein (1989) also calculated the age of a
metro area as the number of decades since it has achieved metro status, and
Schneider and Park (1989) employed the age of a SMSA area as the years counted
after the area passed the 50,000 population. The reason for counting decades after
1900 is because county population data is only available between 1900-1990 from
the Bureau of the Census.
9 3
4.6.7 Other Determinant Variables
Countv home rule status
• home rule county= 1; non home rule county=0
Countv unincorporated ratio
• see chapter two
Countv row officers index
• see chapter two
Number of municipalities
• see chapter three
Number of special districts
• see chapter three
Presence of central cities
• number of cities with population of 50,000 or more
4.7 Data Analysis Techniques
To investigate cause-effect relations, an ordinary least squares regression technique
is used. To use this technique, four assumptions should be checked such as
normality, linearity, homoscedasticity, and no multicollinearity (Norusis, 1994;
Gujarati, 1988). Examining the data for these assumptions produces the following
conclusions.
(1) Policy outcome variables are transformed to the logarithm form with base 10 to
make the models linear.
9 4
(2) Housing unit change is deleted because it is highly correlated with population
change. Their correlation coefficient is .93.
(3) The number of central cities is deleted because it is highly correlated with
population (r=.92).
(4) The number of municipalities is also deleted because it is highly correlated with
the number of special districts (r=.75).
(5) Intergovernmental aid is deleted due to the two reasons stated above.
(6) The number of cases included in each regression model slightly varies from that
of the original data set because of the deletion of outliers which locate outside of
the 3 absolute values of the studentized residuals.
As presented, factor analysis and scale reliability tests (Cronbach Alpha test) are also
utilized to formulated county structural reform index.
4.8 Research Hypothesis Statements
Research hypotheses are stated only for the relationships between policy outcome
variables and county reformed political structures. Relations between policy
outcome variables and control variables are presented in the next summary table
with the expected directions.
Even though recent county research consistently found positive relationships
between county reformed forms and county expenditures (see chapter three),
hypotheses stated here are based on the traditional assertions about normative and
empirical relations between two variables. Thus, if this study finds positive
9 5
relationships again, future research on the relationships about county reformism and
policy outcomes may state its hypotheses in the opposite direction.
Hypothesis one:
• The higher the county structural reform index, the lower the per capita county
total expenditures, per capita county total tax revenues, and per capita county
expenditures for urban functions.
Hypothesis two:
• The more reformed the county political structures are, that is county executive or
county administrator form and at-large election method, the lower the per capita
county total expenditures, per capita county total tax revenues, and per capita
county expenditures for urban functions.
Hypothesis three:
• The relations stated by hypotheses one and two are shown more clearly for
metropolitan counties than for nonmetropolitan counties.
Table 4.7: Expected Directions Between Dependent Variables and Other
Determinant Variables
Variable Expenditures Taxes Urban Exp.
Demographics
Population 1990 +- +- +-
Density + + +
Population change 1980-1990 + - +
Socioeconomic Factors
Personal Per Capita Income + + +
Percent People Below Poverty Level + - +-
Functional Scope + + +
96
Variable Expenditures Taxes Urban Exp.
State M andates
Functional Mandates + + +
Property Tax Limit - - -
Debt Limit - + -
Region
New England - - -
Mid-Atlantic - -
+ -
North Central - - -
South +- +- +-
Age + +- +
Home Rule Status - - -
Unincorporated Ratio + +- +
Row Officers Index + + +
Special Districts - - "
9 7
Notes
1. Bahl et al. (1980) identified six expenditure decision models and provided main
elements of each model: (1) constrained utility maximization model; (2) median
voter model; (3) budget maximization or bureaucratic model; (4) public choice
model; (S) organizational theory model; and (6) statistical approaches.
2. DeSantis (1989) surveyed this item, but the number of counties responding to his
survey is 1,295. Thus, this information is not relevant to this dissertation because of
the many missing cases.
CHAPTER V
9 8
RESEARCH RESULTS AND DISCUSSIONS
5.1 Introduction
This chapter presents research results according to the order of research hypotheses
stated in chapter four. Findings for all counties, MSA counties, and non-MSA
counties are separately discussed in three sections. In each section, findings of other
independent variables associated with the county reform index are not presented
because their results are almost equivalent to those rendered by regression models of
individual reform items. Thus, beta coefficients of the county reform index are
placed at the top of the result table, and bold typed. The other numeric values
below the reform index represent only regression results for the models with
individual reform items and other independent variables.
Other findings are presented at the end of the three sections for the counties
which have not changed their governing forms during the 1978-1989 period. This
analysis is meaningful for examination of the long-term impact of the county reform
items. The county structural reform index is not calculated for this analysis because
of the lack of other county reform items utilized for the index formation in 1978.
Regression coefficients are presented as standardized format (Beta weights) in
result tables because the primary purpose of this research is to examine the direction
and relative importance of the independent variables. For the relative importance of
9 9
a variable, standardized beta is more appropriate than the usual regression
coefficients. Though it is argued that beta weights are affected by correlations of
the independent variables (Norusis, 1994), they still reflect the comparability of
importance among independent variables and have been frequently provided alone in
previous research.
Actual outputs of all regression models are attached at the end of the
dissertation as an appendix. Because of the log transformation of dependent
variables, regression coefficients (B value, not Beta value) in actual outputs can be
interpreted approximately as a percentage term. For example, the B coefficient of
the reform index obtained from the regression model of per capita total county
expenditures applied to all counties is about .073. This means that a one-unit
increase on the reform scale causes a 7.3 percent increase of per capital total
expenditures when other independent variables are held constant. Finally, all
discussions in this chapter referring to three policy outcome variables are on the per
capita basis.
5.2 Results and Discussions for All Counties
Table 5.1 shows standardized beta coefficients of the reform index, three reform
items, and other determinant variables. Based on the F values, all models are
statistically significant indicating that linearity exists between each dependent
variable and the independent variables. The adjusted R-squares of these models are
1 0 0
.615, .537, and .658. This means, in the case of the tax revenue model, a 53.7
percent variation of per capita county total tax revenues is explained by the
independent variables in the model.
Table 5.1: Regression Results for All Counties
Variables Expenditures Tax Revenues Urban Exp.
Reformed Structures
Reform Index .180* .388* .102*
Executive Form .144* .261* .066*
Administrator Form .072* .210* .068*
At-large Election - - -
Demographic Factors
Population - - -
Density - - .051*
Population Change -.132* -.212* -
Socioeconomic Factors
Income .106* -.075* .060*
Poverty - - -
Functional Scope .624* .364* .628*
State Mandates
Functional Mandates .073* .178* .045*
Property Tax Limit - -.042* -
Debt Limit -.103* -.126* .106*
Region
New England -.178* -.242* -
Mid-Atlantic -.054* -.133* -
North Central -.086* -.072* -.078*
South -.226* -.354* -
Age -.058*
_
.078*
Home Rule Status -.053* -.100* -
Unincorporated Ratio .097* .251* .093*
Row Officers Index -.092* -.191* -.112*
Special Districts -.084* - -
101
Variables Expenditures Tax Revenues Urban Exp.
F-ratio 167* 121* 188*
Adjusted R2 .615 .537 .658
Cases (Total=2,246) 2,185 2,187 2,044
* Significant at .01 level.
Countv Reformism Variables
The county structural reform index actually increases all three policy outcome
variables (beta=.180, .388, and .102, pC.Ol), as do the county executive and
administrative forms. The county executive form is more important than the county
administrator form for total expenditures (beta=.144 vs. .072) and tax revenues
(beta=.261 vs. .210). But it is slightly less important than the county administrator
form in determining the level of urban expenditures (beta=.066 vs. .068).
Individual reform items show lower beta values than the reformism index for all
three policy outcome variables. This indicates that rather than the reform items, the
composed index may better explain overall features of reformed county structures.
In contrast, the at-large election system is found to have no relationship with the
policy outcome variables.
These results are contrary to the research hypotheses and the previous
research on city governments (Lyon, 1978; Lineberry & Fowler, 1967), but
consistent with recent county studies (DeSantis & Renner, 1994; Parks, 1994;
Schneider & Park, 1989). As explained in chapters two and three, county reform is
102
urged and adopted to effectively manage already expanded county functions. In city
research such as that of Liebert (1974, 1976), reformed cities are more likely than
unreformed cities to provide a smaller number of functions.
Why reformed county structures lead to higher policy outcomes becomes
clear if one sees the beta coefficients of the county functional scope. The county
functional scope is not only significant with positive directions, but also as the
strongest independent variable in predicting county total and urban expenditures.
The betas for these two dependent variables record .624 and .628, respectively. The
correlation coefficient between county reform index and functional scope in MSA
counties is also larger than in non-MSA counties (r=.489 vs. r=.285, p < .01).
For county tax revenues, the betas of functional scope are far lower than
those of models for the other two dependent variables (beta=.364, p < .01). This
indicates that the number of functions that county governments provide is a better
predictor of their expenditures than their tax revenues. As explained in chapter
four, the best indicator for county tax revenues may be state aid to county
governments if it is included in the analysis. Their correlation coefficient is almost
perfect, or .9867 (pc.O l).
Demographic Variables
Population has no relation to the three output variables, meaning that county size has
no scale economies. This result contradicts a previous study by Duncombe et al.
103
(1991), but is consistent with the studies by DeSantis and Renner (1994) and Park
(1994). Density, as expected, shows positive effects only on the level of county
urban expenditures (beta=.051, p < .01).
Population change shows negative impact on the level of total expenditures
(beta=-.132, p c .O l) and tax revenues (beta=-.212, p c .O l). As stated in chapter
two, population growth during the last decade took place primarily in suburban
areas, while populated and rural counties experienced a decline in population. The
negative impact of population change on county expenditures indicates that the
increased population has demanded public services primarily from suburban cities
rather than county governments. Park’s study (1994) found that population change
has a negative impact on the level of total expenditures of central cities and counties,
but a positive impact on that of suburban cities. Demand for public services may be
increased in some counties by residents newly settled in unincorporated areas. But
at the all-county level, this increase of services may be outweighed by the decreased
demand due to the loss of population in other county areas. In addition, county
governments with population decreases easily reduce the level of public services, but
county governments with a population increase can provide the newly demanded
services only if they raise taxes, impose user charges, or obtain outer resources like
intergovernmental aid. This indicates a relative ease in cutting public services rather
than increasing them.
1 0 4
This upward rigidity is very prominent in taxation. Population decrease,
coupled with its damaging consequences, immediately reduces county tax revenues,
but its increase does not necessarily lead to immediate county tax revenue increases.
This situation renders a negative impact of population change on county tax
revenues. The flypaper effect identified by various research on intergovernmental
aid is a rigidity example for taxation (Schneider, 1989; Fisher, 1987).
Socioeconomic Variables
Personal median income which is commonly regarded as an indicator of public
service demand on a community has a positive impact, as expected, on the level of
county total expenditures (beta=.106, p c .O l) and expenditures for urban functions
(beta=.060, p c .O l). Its negative impact on county tax revenues is unexpected
(beta=.075, p c .O l), but may be because counties with higher incomes tend to have
property tax limits imposed by propositions or state laws more than in counties with
lower incomes. County median income level and the presence of property tax limit
are positively correlated (r=.1295 after controlling for 1990 population, p c.O l).
The percentage of poverty is found to have no relationship to the three dependent
variables.
1 0 5
Functional Scope
As stated above, the number of functions that county governments provide is the
most important predictor for how much they spend for overall and urban functions.
This result is congruent with previous research on both city and county governments
which shows the significant importance of functional scope (Park, 1994; Schneider
& Park, 1989; Schneider, 1989; Benton & Rigos, 1985; Dye & Garcia, 1978).
State Mandates
All three state mandate variables demonstrate expected directions for county policy
outcomes. State functional mandates, though the data are outdated, have affected
positively county total expenditures (beta=.073, p c .O l), tax revenues (beta=.178,
p < .01), and urban expenditures (beta=.045, p < .01). The property tax limit has a
negative impact on the level of county tax revenues (beta=-.042, p c .O l), but has
no relationship with the two county expenditures. This again demonstrates the
loosened linkage between county tax revenues and expenditures. In other words,
county expenditures are not affected substantially by state limits on taxes.
The debt limit decreases county total expenditures (beta=-.103, p c .O l) and
tax revenues (beta=-. 126, p < .01). The negative impact of debt limit on county tax
revenues is contrary to expectations. Facing the state’s debt limit, county
governments are supposed to raise taxes. This unexpected direction may be because
both a property tax limit and debt limit are given to county governments as a single
1 0 6
package. Thus, the effect of the property tax limit overwhelms the desire to
increase taxes. Thirty-one states, or 65 percent out of forty-eight states, mandate
both limits on their local governments (Rhode Island and Connecticut do not have
county governments).
The state debt limit has a positive impact on county urban expenditures
(beta=.106, p c.O l). As shown in tables 5.5 and 5.6, this positive impact of the
debt limit at the all-county level is produced mainly because this relationship exists
more strongly for non-MSA counties (beta=.157, p c .O l), but disappears for MSA
counties. This phenomenon is contrary to usual thoughts about rural counties, but
demonstrates that non-MSA counties in fact act as a main urban service-provider in
rural areas. Thus, as a main provider, non-MSA counties have to provide urban
functions for their residents, although the debt limit exists.
Region
Expectedly, the result shows that counties in states other than the Western states
spend less in total expenditures and levy less in taxes. As in previous research, this
result confirms the importance of a regional variable for explaining the level of
policy outcome variables. Counties in Mid-Atlantic and Southern states are not
different from Western counties in spending for urban functions. The information
presented by table 4.3 in chapter four is suggestive of these results. It shows that
per capita expenditures for urban functions spent by Mid-Atlantic counties are higher
107
than those of Western counties, and Southern counties spend as much in
expenditures as do Western counties.
Age
Age of county governments reflects mixed results. Age is positively related to the
level of urban expenditures (beta=.078, p c .O l), but is found to have a negative
impact on the level of total expenditures (beta=-.058, p c .O l). Age has no
relationship with county tax revenues.
Countv Home Rule Status
The results show that counties with a home rule charter spend less for county total
expenditures (beta=-.053, p c .O l) and collect less taxes (beta=-.100, p <.01).
This result is very suggestive of the overall influence of home rule adoption for
county fiscal performance. It is also a meaningful result considering the mean
values of the three policy outcome variables between home rule counties and non
home rule counties. Table 5.2 demonstrates higher averages for home rule counties
in all three policy outcomes. In contrast to the simple analysis, multiple regression
analysis renders an opposite direction indicating that when other factors are taken
into account-or with other factors held constant, a home rule charter has the effect
of lowering the level of county spending and taxes.
1 0 8
Table 5.2: Mean Test for Three Dependent Variables by County Home Rule
Status
Home Rule
Counties
Non-Home Rule
Counties
t-values
Total Expenditures 976.6 634.3 -5.7*
Tax Revenues 244.5 200.7 -1.8
Urban Expenditures 150.5 38.3 -12.2*
* Significant at .01 level.
a. Sources: Bureau of the Census (1992), NACo (1989).
b. Nine Alaska boroughs are excluded for the same reason explained in table 4.3.
Countv Unincorporated Ratio
Expectedly, counties with higher unincorporated ratios tend to spend more for total
expenditures (beta=.097, p c .O l) and urban functions (beta=.251, p <.01), and
levy more taxes (beta=.093, p< .01) than counties with lower unincorporated
ratios. The reason for this result is self-evident because county governments are
primary units of local government in providing public services for residents in
unincorporated areas.
Countv Row Officers Index
The number of county elected row officer positions has a negative impact on all
policy outcome variables: betas are -.092 for total expenditures, -.191 for tax
revenues, and -.112 for urban expenditures. All are statistically significant. This
is contrary to the usual thought that the lower the number, the more reformed the
county government. This result may occur because state laws mandate which row
officers should be elected (Salant & Martin, 1993; Jeffery et al., 1989). Thus,
county governments have to abide by the laws and elect them all although they
might not prefer to elect such a large number of row officers.
Table 5.3: Correlation Coefficients Between Row Officers Index and Three
Dependent Variables
Expenditures Tax Revenues Urban Exp.
Row Officers Index -.1721* -.3364* -.1457*
* Significant at .01 level.
Table 5.3 shows all negative correlation coefficients between the county row
officers index and the three dependent variables. This table states that counties with
higher indexes tend to spend less for total expenditures, urban expenditures, and
collect less taxes than counties with lower indexes.
Table 5.4: Correlation Coefficients Between County Row Officers Index and
Selected County Characteristic Variables
Population Density Income Scope
Row Officers Index -.0511* .-1159* -.1733* -.2706*
* Significant at .01 level.
Table 5.4 presents all negative correlation coefficients between the row
officers index and selected county characteristics. According to this table, counties
110
which are smaller in size, less dense, have lower incomes, and provide fewer
functions tend to have a larger number of elected row officer positions. Based on
this information, one can conclude that counties with a larger number of row officer
positions are rural, smaller in size, lower in functional scope, and poor. Thus, it is
not difficult to expect that these counties would reduce their row officer positions
but for the state mandate.
Special Districts
As expected, the number of special districts reduces county total expenditures
(beta=-.084, pC .O l). However, different from the arguments that districts are
created to avoid state tax and debt limits (Herson & Bolland, 1990; Chicoine &
Walzer, 1985; ACIR, 1982), the number of special districts has no relation to
county tax revenues. This number also is not related to the level of county urban
expenditures. It is rather obscure why the number of special districts in a county
area does not affect the level of county urban expenditures, considering that special
districts are usually created to take care of large-scale urban functions. For bivariate
analysis, the number of special districts is significantly correlated with urban
phenomena like population (r=.6085), density (r=.3350), income (r=.3304), and
county age (r=.5018). Future research is warranted in this area.
Ill
5.3 Results for MSA Counties
Table 5.5 below shows that all three regression models applied to MSA counties are
statistically significant (see F-ratios). Their adjusted R-squares record .700, .577,
and .722. The models are improved compared to those models applied to all
counties. This means that listed independent variables better explain three county
policy outcome variables at the MSA county level than at the all county level.
These models’ improvement has been produced at the cost of a smaller number of
significant independent variables.
In this section, the county reform index and individual reform items are
discussed at some length to examine whether or not the research hypotheses are
verified. Regarding other determinant variables, however, discussions are brief
because their directions are not changed at all compared to the previous tables for all
counties.
Table 5.5: Regression Results for MSA Counties
Variables Expenditures Tax Revenues Urban Exp.
Reformed Structures
Reform Index .246* .316* .128*
Executive Form .150* .212* .067*
Administrator Form .090* .141* .079*
At-large Election - - -
Demographic Factors
Population - - -
Density - - .092*
Population Change -.104* -.242* -
1 1 2
Variables Expenditures Tax Revenues Urban Exp.
Socioeconomic Factors
Income - -.119* -
Poverty - - -
Functional Scope .695* .417* .637*
State M andates
Functional Mandates -
.213*
-
Property Tax Limit - - -
Debt Limit -.115* -.126* -
Region
New England -.149* -.268* -
Mid-Atlantic - -.193* -
North Central - - -
South - -.389* -
Age
. _
.136*
Home Rule Status - -.121* -
Unincorporated Rate - .216* .146*
Row Officers Index -
-.186* -.121*
Special Districts - - -
F-ratio 75* 44* 79*
Adjusted R“ .700 .577 .722
Cases (Total =676) 668 669 632
* Significant at .01 level.
Contrary to the research hypotheses, the county structural reform index is
found to have a positive impact on the level of all policy outcome variables
(beta=.246, .316, and .128, p< .01). Compared to the results for all counties, its
relative importance is increased for total expenditures from .180 to .246 and urban
expenditures from .102 to .128, while it is decreased for county tax revenues from
1 1 3
.388 to .316. The same change in relative importance also happens to the individual
reform items, indicating that at the MSA county level, reformed structures play a
more important role in the county’s spending decisions than in taxation decisions.
The county executive and administrator forms are also found to have a
positive impact on the level of all policy outcomes. The county executive form
remains more important than the county administrator form for total expenditures
(beta=.150 vs. .090) and tax revenues (beta=.212 vs. .141). Like the results at the
all-county level, it is less important than the county administrator form for county
urban expenditures (beta=.067 and .079, p< .01). The at-large election method is
also found to have no impact on any of the three policy outcomes.
Judging from the results of the other independent variables, regional
division, income level, age, the number of special districts, property tax limit, and
state functional mandates are less discriminating in the variance of county policy
outcomes at the metropolitan level than at the all-county level. Regionally, only
counties in New England states spend less for overall expenditures (beta=-.149,
p< .01) than those in Western states. Interestingly, all counties are not different in
county spending decisions for urban functions, indicating that all MSA counties are
equivalently involved in providing urban functions. County income level also turns
out to be indifferent to the level of county expenditures. County age is positively
related only to county urban expenditures. The number of special districts has no
relationship to any of the three policy outcomes. County home rule status and row
1 1 4
officers index become insignificant for county total expenditures. In contrast, along
with reformed structures, functional scope and population change are still influential
for decisions about county policy outcomes in metropolitan areas.
5.4 Results for Non-MSA Counties
Table 5.6 demonstrates the significance of all three regression models for non-MSA
counties (see F-values). Their adjusted R-squares are decreased for all policy
outcome variables compared to the models applied to all counties. Like the previous
models, it can also be concluded from this table that reformed county structures,
either measured by the reform index or by individual items, are positively related to
the level of the three policy outcome variables. However, having at-large elections
is found to have a positive impact on county tax revenues (beta=.057, p < .01).
Other independent variables produce similar results compared to those of all
counties in their significance and relative importance. In demographic factors,
density becomes insignificant while population is changed into being negatively
related to county total expenditures (beta=-. 137, pC.Ol). This indicates that scale
economies of county governments exist in rural areas. Interestingly, county home
rule status is found to have no relationship with county total expenditures, but
unincorporated ratio and row officers index still remain as important variables to the
three policy outcome variables.
1 1 5
Table 5.6: Regression Results for Non-MSA Counties
Variables Expenditures Tax Revenues Urban Exp.
Reformed Structures
Reform Index .216* .399* .124*
Executive Form .181* .279* .078*
Administrator Form .107* .230* .083*
At-large Election - .057* -
Demographic Factors
Population -.137* - -
Density - - -
Population change -.073* -.142* -
Socioeconomic Factors
Income .146* - .104*
Poverty - - -
Functional Scope .590* .315* .602*
State M andates
Functional Mandates .081* .151*
-
Property Tax Limit - -.063* -
Debt Limit -.070* -.111* .157*
Region
New England
*
IT)
£
r
-.301* -
Mid-Atlantic - -.089* -
North Central -.087* -
-.129*
South -.257* -.351* -
Age
_ _
Home Rule Status - -.071* -
Unincorporated Rate .091* .281* .073*
Row Officers Index -.116* -.207* -.113*
Special Districts - - -
F-ratio 104* 79* 109*
Adjusted R2 .587 .518 .617
Cases (Total = 1,570) 1,519 1,526 1,417
* Significant at .01 level.
1 1 6
5.5 Long-term Impact of Reformed County Structures
Table 5.7 presents beta coefficients for two county reformed items and other
determinant variables rendered by regression analysis applied to the counties which
have not changed their governing forms during 1978-1989. The three regression
models are statistically significant (F=160, 121, and 169, p< .0 1 ), and adjusted R-
squares are higher than those of models for all and non-MSA counties, but lower
than those for MSA counties.
Like other analyses, the two county reformed forms are found to have a
positive impact on county tax revenues (beta=.056 and .263, p < .01) and urban
expenditures (beta=.068 and .050, p< .01). But there is an important change
showing that the beta value of the county administrator form becomes much higher
than that of the county executive form in county tax revenues. Another change is
that the county executive form is insignificant for decisions about county total
expenditures. This may be because all 119 Kentucky and 87 Tennessee executive
counties were excluded in the analysis. These counties comprise 93 percent of all
221 counties which have changed their governing forms from commission to the
executive form during this period. As stated in chapter two, Kentucky and
Tennessee state governments mandated their counties to adopt only the county
executive form. This result, thus, raises a question about the long-term impact of
the county executive form on county policy outcomes. Does the county executive
form -a reformed structure-not increase the level of county expenditures when a
1 1 7
county has had this form, for example, for more than ten years? Future research is
warranted by this question.
Table 5.7: Regression Results for the Counties With No Formal Change During
the Period of 1978-1989
Variable Expenditures Taxes Urban Exp.
Reformed Structures
Executive Form - .056* .068*
Administrator Form .076* .263* .050*
Demographic Factors
Population - - -
Density - - .046*
Population -.113* -.195* .035*
Socioeconomic Factors
Income .142* -
.061*
Poverty - -.045* -
Functional Scope .572* .257* .608*
State Mandates
Functional Mandates .150* .245* .077*
Property Tax Limit .087* .049* .033*
Debt Limit - .067* .143*
Region
New England -.163* -.235* -
Aid-Atlantic -.074* -.131* -
North Central -.051*
-
-.062*
South -.103* -.185* -
Age -.060*
_
.071*
Home Rule Status -
-.038*
-
Unincorporated Ratio .097* .255* .080*
Row Officers Index -.091* -.211* -.123*
Special Districts -.111* - -.070*
1 1 8
Variable Expenditures Taxes Urban Exp.
F-ratio 160* 121* 169*
Adjusted R2 .653 .586 .682
Cases (Total = 1,745) 1,694 1,701 1,567
* Significant at .01 level.
5.6 Summary of Mtyor Research Findings
Reformed Countv Structures
(1) Contrary to city reform studies and research hypotheses in this dissertation,
reformed county structures, except for at-large election, lead county
governments to spend more in expenditures for overall and urban functions and
levy more taxes. This impact of reformed structures is very clear regardless of
county metro or non-metro status. This is also true when county reform is
measured by a composed reform index or by individual reform items.
(2) The county structural reform index is more important than the individual reform
items, judging from the standardized beta values. This suggests that the
composed index developed in this study better represents the reform features of
county governments than the individual reform items.
(3) The county executive form is more important than the county administrator form
in determining the level of county total expenditures and tax revenues. But in
urban expenditures, the county administrator form is more important than the
county executive form.
119
(4) The county at-large election is significant only to the level of county tax
revenues in non-metro areas.
(5) For the long-term impact of reformed county structures, both the executive and
administrator forms positively affect the level of county tax revenues and urban
expenditures. In the area of county total expenditures, the executive form turns
out to be insignificant while the administrator form remains significant.
Other Determinant Variables
(1) County functional scope is the most stable and influential predictor for the level
of the three county policy outcomes-total expenditures, tax revenues, and urban
expenditures.
(2) Population change in 1980-1990 is consistently significant with county total
expenditures and tax revenues (negative relationship) for all analyses. Density is
a predictor only for county urban expenditures in all and metro counties,
whereas population is negatively related only to county total expenditures in non-
metro areas, showing the existence of scale economies for county size.
(3) Personal median income is a consistent predictor for the level of all policy
outcome variables. But it becomes significant to county tax revenues only when
applied to metro counties, and only to county total and urban expenditures in
non-metro areas. Percentage of persons below the poverty line has no effect on
the level of policy outcomes for all analyses.
120
(4) Debt limit is a significant and consistent predictor for all county policy outcomes
in non-metro areas. But it becomes insignificant to the level of county urban
expenditures in metro areas. Property tax limits have no differential power for
all policy outcome variables in metro areas, but they do affect tax revenues in
non-metro areas. State functional mandates are significant only for the level of
county tax revenues in metro areas, and for county total expenditures and tax
revenues in non-metro areas.
(5) Regional division has a significant predicting power for the level of total
expenditures and tax revenues in non-metro areas, but this power disappears in
metro counties except in New England. For county urban expenditures, region
does not make any difference in metro areas, suggesting that all metro counties
regardless of their locations are actively involved in urban functions.
(6) Age has a positive impact only on the level of county urban expenditures in
metro areas. But at the all-county level, it shows a negative relationship with
county total expenditures.
(7) County home rule status has a negative impact only on the level of county tax
revenues both in metro and non-metro areas. But for all counties, it is
negatively related to county total expenditures.
(8) County unincorporated ratio and the row officers index are significant predictors
for all county policy outcomes in all counties and non-metro counties. But they
are not significant for county total expenditures in metro areas.
121
(9) The number of special districts, unexpectedly, has no impact on the level of all
county policy outcomes. It, however, does have a negative impact on county
total expenditures for the analysis of all counties.
This chapter presented and discussed research findings according to the order
of hypotheses stated in chapter four. The next and final chapter summarizes the
main contents of this dissertation, draws out the practical and theoretical
implications, and provides recommendations for future county studies.
1 2 2
CHAPTER VI
CONCLUSIONS
6.1 Introduction
This chapter briefly summarizes the previous chapters, then it discusses the
implications of the research findings for the county reform movement. Finally, it
presents the theoretical implications and offers recommendations for future studies
regarding county reforms.
6.2 Summary of the Previous Chapters
This dissertation has tried to answer one main question, "do reformed county
political structures make a difference in achieving county policy outcomes?"
Chapter one introduces the reason why this question is worthy of pursuit by
reviewing the current state of county research. Compared to the richness of city
research concerning reformed structures and policy outcomes, county research has
been scarce. It is only recently that academic interest in county governments has
gradually increased. This heightened academic interest firmly demonstrates that,
after a long period of neglect, county governments have emerged not only as
significant subjects worthy of research but as important units of local government.
The latter image of county governments has begun to be revealed, after being
overshadowed by the county's other image-as an administrative arm of the state.
1 2 3
Chapter two focuses on suburbanization as a motivating force behind county
changes and explains the subsequent expansion of counties' functional
responsibilities and their adoption of reformed structures. The rapid suburbanization
process which has taken place since World War II has led county governments to
emerge as important area-wide service providers and regional problem solvers.
During the 1967-1987 period, county urban functions such as fire protection, police,
housing renewal, libraries, and higher education have increased in terms of their
relative size of public expenditures. Similarly, counties still remain strong, or have
grown in traditional functional areas like corrections and public welfare.
Nonetheless, reformed county political structures have not been adopted
concurrently with the growing pace of their functional responsibilities. About sixty-
three percent of counties still maintain the commission form, while only twelve
percent of counties have adopted the executive form as their governing structure.
Regarding the representative method for governing boards of supervisors, only
thirty-five percent of counties use an at-large constituency. Furthermore, the actual
adoptions of home rule charters is not impressive; less than ten percent of counties
have such charters. However, once home rule is implemented, counties with home
rule charters are more likely than those without home rule charters to have reformed
governing forms and a smaller number of elected county row officer positions.
Chapter three introduces ethos theory as the basis of arguments about city
reformism, and it reviews related research on reformed city structures. Empirical
1 2 4
results of recent county studies are also discussed at the end of the chapter. The
variables identified in the literature review are: (1) reformed political structures; (2)
regional division; (3) functional scope; (4) age; (5) demographic and socioeconomic
factors; and (6) seven variables unique to county studies. These variables are
separately discussed as predictors of county policy outcomes.
Chapter four, the methodology part of the dissertation, selects three per
capita measurements as county policy outcome variables: total expenditures, total tax
revenues, and urban expenditures. This chapter explains the cross-sectional and
cause-effect analyses as used in the research design of this dissertation. An ordinary
linear regression technique is utilized in order to isolate the impacts of reformed
county political structures and other determinant variables. Using factor analysis, a
county structural reform index is formulated using county home rule status, county
reformed governing forms, and a dichotomous row officers index. This county
reform index is applied separately to regression models along with individual reform
items such as governing forms and representative methods. Operationalization of
other determinant variables is also made in this chapter. Finally, three research
hypotheses are stated regarding the impact of reformed county political structures.
The main content of these research hypotheses is that the more county governments
are reformed in their political structures, the lower are levels of county policy
outcomes. This direction is consistent with normative and empirical arguments
about city reform, but they differ from the results of recent county research.
1 2 5
Chapter five presents and discusses the research findings rendered by the
various regression models. These research findings invariably demonstrate a
positive impact from reformed county structures on policy outcomes. These findings
reject the original hypotheses of this dissertation. These positive impacts are very
clear regardless of whether county reform is measured by a composed reform index
or by individual reform items, and of whether these analyses are applied separately
to metro and non-metro areas. The long-term effects of reformed county structures
also show the same outcomes. As for the other determinant variables, county
functional scope, population change, debt limit, regional division, county home rule
status, county unincorporated ratio, and the number of elected row officer positions
are identified as significant and consistent predictors for county policy outcomes.
But the degree of significance of these variables varies by metro and non-metro
areas.
6.3 Implications of This Dissertation
The main result of this dissertation-that there are positive impacts from reformed
county political structures-makes it hard to draw a policy recommendation for
county reform practitioners and concerned citizens. As stated in chapter five, the
adoption of reformed county governing forms, particularly the county executive
form, are urged to effectively serve already expanded county functions. Thus,
simply saving costs in implementing mandated or optional programs is not the
1 2 6
primary purpose in changing to a reformed form. Nonetheless, at this time it is not
right for academic scholars and county reformists simply to urge county
governments to adopt the executive form. In other words, this simple argument
regarding the county executive form is not valid in terms of efficiency gains unless
other county research verifies the "effectiveness" of the county executive form.
Yet, county home rule is warranted by this study. As identified in chapter
five, county home rule status has a significant effect in lowering total county
expenditures and tax revenues when other related factors are held constant. As
reported in chapter two, counties with home rule tend to have reformed forms and
smaller numbers of elected row officer positions more often than those without home
rule. This is very suggestive of what should be an ultimate objective of the county
reform movement. Helping counties adopt home rule charters produces more than
merely changing governing forms; it could make the administration of county
governments more effective in serving their growing functional responsibilities in
local affairs. The merits of adoption of home rule, however, depend upon the
degree of latitude authorized by state governments. Compared to the forty states
which authorize the structural home rule option for their city governments, only
twenty-four states grant it to their county governments (ACIR, 1993b). Enhancing
county reform will require a collaborative effort by state governments and county
reformists. This collaboration must be aimed at achieving legal authorization for a
comprehensive home rule option, not just at limited structural change. The home
127
rule option coupled with subsequently enlarged autonomy benefits the administration
of county governments far more than the latter option alone. With the home rule
option, counties can decide whether to adopt home rule charters or to remain in
general-law status depending on their local circumstances. Besides legal
authorization of the home rule option, many formal and informal constraints exist
which increase the cost of adopting, maintaining, and altering county home rule
charters (Salant, 1988 and see chapter two). Ways to alleviate these constraints need
to be devised for the progress of overall county reform.
6.4 Recommendations for Future Study
Future studies of the impact of reformed county structures might not produce
meaningful outcomes if they utilize the same research design as employed by this
study. Previous county research (DeSantis & Renner, 1994; Schneider & Park,
1989) found a consistent and positive relationship between reformed forms and
policy outcomes. This dissertation also confirms the same relationship. It is
desirable to diversify study areas about county reformed structures. One area which
is most uncertain is why county governing forms, though reformed, render higher
level of per capita total expenditures and taxes. It is not sufficient to explain positive
impact of reformed county structures as being a function of expanded county
functional responsibilities. This dissertation explains expansion of county functions
and change of county governing forms. It is also implied that there is a probable
1 2 8
association between the two. But whether expanding county functions actually have
caused the adoption of reformed structures should be further explored. Other studies
can be expected to answer this question by looking at the internal managerial aspects
of reformed counties rather than at their structural differences. As noted in chapter
two, the county modernizing movement has been summarized as an effort to
improve the overall problem-solving capacity of county governments (Duncombe,
1977).
County structural reform is not a one-dimensional phenomenon. Other
aspects of reformed county structures surely warrant future studies. This
dissertation introduces three reform elements: (1) county home rule status; (2) a row
officers index; and (3) a composed reform index. The number of county supervisors
is also a useful element for the study of county reform, though not utilized in this
study. These variables are not only unique to county studies but have been found to
have individual and distinctive effects on the level of county outcomes.
The county structural reform index is far more important than separate
reform items in determining the level of county policy outcomes (see chapter five).
Future studies using this index, however, have to improve the internal consistency
among the included elements-home rule status, reformed governing forms, and the
row officers index. The best way to improve the reliability of this index is to
explicate the relative importance, or weight, of each item. If their weights are
1 2 9
appropriately measured, this index could become a reliable representative variable
for reformed county structures.
The county row officers index is found to be negatively related to home rule
status and the executive/administrator form. It also well reflects the attributes of
counties which are rural, smaller in size, lower in functional scope, and poor.
Thus, the utility of this index can be increased by future studies which specially
focus on rural counties (or non-metro counties). Recent academic interest in county
governments largely ignores rural counties compared to urban counties. This
dissertation demonstrates by analyzing non-metro counties that their status is quite
different from that of metro counties. In rural areas, the county government is the
most important service-provider and faces no competition from municipalities.
Finally, the county unincorporated ratio shows significant research potential
for research on the relationships between county urban functions and population
change. This ratio is also related to municipal incorporation and annexation efforts.
Thus, the utility of this ratio may benefit from research on city governments.
Even though studies on county governments have been burgeoning since the
late 1980s, potential areas for county reform still remain unanswered and await
future research. It is expected by the author that the various reform elements
identified and measured by this dissertation will make a contribution to future
studies which will result in better understandings about the American county system.
1 3 0
BIBLIOGRAPHY
Advisory Commission on Intergovernmental Relations, 1993a. Metropolitan
Organization: Comparison o f the Allegheny and St. Louis Case Studies.
Washington, DC: ACIR, SR-15.
1993b. State Laws Governing Local Government Structure and
Administration. Washington, DC: ACIR, M-186.
1993c. Local Government Autonomy: Needs for State Constitutional
Statutory, and Judicial Clarification. Washington, DC: ACIR, A-127.
1992. Metropolitan Organization: The Allegheny County Case. Washington,
DC: ACIR, M-181.
1988. Metropolitan Organization: The St. Louis Case. Washington, DC:
ACIR, M-158.
1987. The Organization o f Local Public Economies. Washington, DC:
ACIR, A-109.
1982. State and Local Roles in the Federal System. Washington, DC: ACIR,
A-88.
1978. State Mandating o f Local Expenditure. Washington, DC: ACIR, A-67.
1972. Profile o f County Government. Washington, DC: ACIR, M-72.
Alford, Robert and Harry Scoble, 1965. “Political and Socioeconomic
Characteristics of American Cities.” The Municipal Year Book 1965. Chicago:
International City Management Association.
Bahl, Roy, Marvin Johnson, and Michael Wasylenko, 1980. “State and Local
Government Expenditure Determinants: The Traditional View and a New
Approach.” in Roy Bahl, Jesse Burkhead, and Bernard Jump eds., Public
Employment and State and Local Government Finance, Cambridge: Ballinger
Publishing Company, pp. 65-119.
Baldassare, Mark, 1994. “Regional Variations in Support For Regional
Governance.” Urban Affairs Quarterly, Vol. 30, No. 2, pp. 275-284.
131
1992. “Suburban Communities.” Annual Review o f Sociology, Vol. 18, pp.
475-494.
Banfield, Edward C., 1974. The Unheavenly City Revisited. Prospect Heights, IL:
Waveland Press, Inc.
Banfield, Edward C. and James Q. Wilson, 1963. City Politics. New York: Vintage
Books.
Benton, J. Edwin and Donald C. Menzel, 1993. “County Services: The Emerging of
Full-Service Government.” In David R. Berman ed., County Governments in
An Era o f Change. Westport, CT: Greenwood Press, pp. 53-69.
1991. “County Service Trends and Practices: The Case of Florida.” State
and Local Government Review, Vol. 28 (Spring), pp. 68-75.
Benton, J. Edwin and Platon N. Rigos, 1985. “Patterns of Metropolitan Service
Dominance: Central City and Central County Service Roles Compared.”
Urban Affairs Quarterly, Vol. 20, No. 3, pp. 285-302.
Benton, J. Edwin and Darwin Gamble, 1984. “City/County Consolidation and
Economies of Scale: Evidence From a Time-Series Analysis in Jacksonville,
Florida.” Social Science Quarterly, Vol. 65 (March), pp. 190-198.
Berman, David R., 1993a. “Counties, Other Governments, and the Future.” In his
edited book, County Governments in An Era o f Change. Westport, CT:
Greenwood Press.
1993b. “State-Local Relations: Patterns and Problems.” in The Municipal
Year Book 1993. Washington, DC: International City Management
Association, pp. 87-93.
Berman, David R. and Lawrence L. Martin, 1988. “State-Local Relations: An
Examination of Local Discretion.” Public Administration Review, Vol. 48
(March/April), pp. 637-641.
Berman, David R., Lawrence L. Martin, and Laura A. Kajfez, 1985. “County
Home Rule: Does Where You Stand Depend on Where You Sit? ” State and
Local Government Review, Vol. 17, pp. 232-234.
Bernstein, Richard, 1978. The Restructuring o f Social and Political Theory.
Philadelphia: University of Pennsylvania Press.
1 3 2
Bish, Robert L. and Vincent Ostrom, 1975. Understanding Urban Government:
Metropolitan Reform Reconsidered. Washington, DC: American Enterprise
Institute for Public Policy Research.
Braaten, Kaye, 1991. “Rural Counties: The Challenges Ahead.” Intergovernmental
Perspective, Vol. 17, No. 1, pp. 38-40.
Bryant, Stephen, 1976. “The Dimensions of Reformism in Urban Policy Analysis.”
Urban Affairs Quarterly, Vol. 12, No. 1, pp. 117-124.
Burrell, Gibson and Garath Mogan, 1979. Sociological Paradigms and
Organizational Analysis: Elements o f the Sociology o f Corporate Life.
Portsmouth, New Hampshire: Heinemann.
Cassella, William N., 1971. “Improving Counties: Form Follows Function.” Public
Management, Vol. 53, No. 4, pp. 6-8.
Chicoine, David L. and Norman Walzer, 1985. Governmental Structure and Local
Public Finance. Boston: Oelgeschlager, Gunn & Hain, Publishers, Inc.
Cigler, Beverly A., 1995. “County Governance in the 1990s.” State and Local
Government Review, Vol. 27, No. 1, pp. 55-77.
1994. “The County-State Connection: A National Study of Associations of
Counties.” Public Administration Review, Vol. 54 (January/February), pp. 3-11.
1993. “The Special Problems of Rural County Governments.” In David R.
Berman ed., County Governments in An Era o f Change. Westport, CT:
Greenwood Press, pp. 89-105.
Clark, Terry Nichols, 1994. “Race and Class Versus the New Political Culture.” In
his edited book, Urban Innovation: Creative Strategies fo r Turbulent Times,
Thousand Oaks, CA: SAGE Publications, pp. 21-78.
1968. “Community Structure, Decision-Making, Budget Expenditures, and
Urban Renewal in 51 American Communities.” American Sociological
Review, Vol. 3 (August), pp. 576-593.
Clark, Terry Nichols, Lorna C. Ferguson, and Robert Y. Shapiro, 1982.
“Functional Performance Analysis: A New Approach to the Study of
1 3 3
Municipal Expenditures and Debt." Political Methodology, Vol. 8, No. 2,
pp. 87-123.
Cole, Richard L., 1971. "The Urban Policy Process: A Note on Structural and
Regional Influences." Social Science Quarterly, Vol. 52, pp. 645-655.
Committee for Economic Development, 1970. Reshaping Government in
Metropolitan Areas. New York: CED.
— 1966. Modernizing Local Government to Secure a Balanced Federalism. New
York: CED.
de Soto, Hernando, 1989. The Other Path: The Invisible Revolution In The Third
World. New York: Harper & Row Publisher.
DeSantis, Victor S., 1989. "County Government: A Century of Change." In The
Municipal Year Book 1989, Washington, DC: International City Management
Association, pp. 55-84.
DeSantis, Victor S. and Tari Renner, 1994. "The Impact of Political Structures on
Public Policies in American Counties." Public Administration Review, Vol.
54, No. 3, pp. 291-295.
-— 1993. "Governing the County: Authority, Structure, and Elections. In David R.
Berman ed., County Governments in An Era of Change. Westport, CT:
Greenwood Press, pp. 15-28.
Dilulio, John J., Gerald Garvey, and Donald F. Kettl, 1993. Improving Government
Performance: An Owner's Manual. The Brookings Institution. Washington,
DC: The Brookings Institution.
Cleveland Foundation, 1980. Reorganizing Our Counties: A Catalog o f County
Government Reorganization Experience in America. Cleveland, OH:
Governmental Research Institute.
Dodge, William R., 1990. "Regional Problem Solving in the 1990s:
Experimentation with Local Governance for the 21st Century." National
Civic Review, Vol. 79 (Spring), pp. 354-366.
Dolan, Drew A., 1990. "Local Government Fragmentation: Does It Drive Up the
Cost of Government?" Urban Affairs Quarterly, Vol. 26, No. 1, pp. 28-45.
134
Downing, R. G ., 1991. “Urban County Fiscal Stress: A Survey of Public Official’s
Perceptions and Government Experience.” Urban Affairs Quarterly, Vol. 27,
No. 2, pp. 314-325.
Downs, Anthony, 1994. New Visions fo r Metropolitan America. Washington, DC:
The Brookings Institution.
Duncombe, Sydney, William Duncombe, and Richard Kinney, 1992. “Factors
Influencing the Politics and Process of County Government Budgeting.”
State and Local Government Review, Vol. 24, No. 1, pp. 19-27.
Duncombe, Herbert Sydney, 1977. Modem County Government. Washington, DC:
National Association of Counties.
Dye, Thomas R., 1965. “City-Suburban Social Distance and Public Policy.” Social
Force, Vol. 4, pp. 100-106.
Dye, Thomas R. and John A. Garcia, 1978. “Structure, Function, and Policy in
American Cities.” Urban Affairs Quarterly, Vol. 14, No. 1, pp. 103-122.
Dye, Thomas R. and Susan A. MacManus, 1990. “State vs. Local Public Sector
Growth: A Comparison of Determinant Models.” Policy Studies Journal,
Vol. 18, No. 3, pp. 645-657.
1976. “Predicting City Government Structure.” American Journal o f Political
Science, Vol. 20, pp. 257-271.
Ebel, Robert D., 1991. “A Profile of County Finances.” Intergovernmental
Perspective, Vol. 17, No. 1, pp. 14-17.
Edwards, Patricia K. and James R. Bolland, 1991. “Reform and Economic
Development: Attitudinal Dimensions of Metropolitan Consolidation.”
Journal o f Urban Affairs, Vol. 13, No. 4, pp. 461-478.
Elster, Jon, 1989. Nuts and Bolts fo r the Social Sciences. New York: Cambridge
University Press.
Fabricant, Solomon, 1952. The Trend o f Government Activity in the United States
since 1900. New York: National Bureau of Economic Research, Inc.
Fairlie, John A., 1906. Local Government in Counties, Towns and Villages. New
York: The Century Co.
1 3 5
Farnham, Paul G. and Stephen N. Bryant, 1985. “Form of Local Government:
Structural Policies of Citizen Choice.” Social Science Quarterly, Vol. 66
(January), pp. 386-400.
Fava, Sylvia F., 1989. “Public Policy and Suburban Development.” In Barbara M.
Kelly ed., Suburbia Re-examined. New York: Greenwood Press, pp. 193-
197.
Ferris, James M. and Graddy, Elizabeth, 1991. “Production Costs, Transaction
Costs, and Local Government Contractor Choice,” Economic Inquiry, Vol.
24 (July), pp. 541-554.
Ferris, James M. and Shui-Yan Tang, 1993. “The New Institutionalism and Public
Administration: An Overview.” Journal o f Public Administration Research
and Theory, Vol. 3, No. 1, pp. 4-10.
Fisher, Ronald C., 1988. State and Local Public Finance. Glenview, IL: Scott,
Foresman and Company.
Fishman, Robert, 1990. “America’s New City: Megalopolis Unbound.” Wilson
Quarterly, (Winter), pp. 24-25.
Fosler, R. Scott, 1991. “The Suburban County: Governing Mainstream Diversity.”
Intergovernmental Perspective, Vol. 17, No. 1, pp. 33-37.
Fox, William F., Jerome M. Stam, W. Maureen Godsey, and Susan D. Brown,
1979. Economics o f Size in Local Government: An Annotated Bibliography.
Washington, DC: U.S. Department of Agriculture.
Frisbie, W. Parker and John D. Kasarda, 1988. “Spatial Processes.” In Neil J.
Smelser ed., Handbook o f Sociology, SAGE Publications: Newburry Park,
pp. 629-666.
Garreau, Joel, 1988. Edge City: Life on The New Frontier. New York: Anchor
Books Doubleday.
Gilbertson, H. S., 1917. The County: The “ Dark Continent” o f American Politics.
New York: The National Short Ballot Organization.
Giuliano, Genevieve and Kenneth A. Small, 1991. “Subcenters in the Los Angeles
Region.” Regional Science and Urban Economics, Vol. 21, pp. 163-182.
1 3 6
Glaab, Charles N. and A. Theodore Brown, 1967. A History o f Urban America.
London: The Macmillan Company.
Gore, Al, 1994. “The New Job of the Federal Executive,” Public Administration
Review, Vol. 54 (July/August), pp. 317-321.
Greene, Barbara P., 1987. “Counties and the Fiscal Challenges of the 1980s.”
Intergovernmental Perspective, Vol. 13 (Winter), pp. 14-19.
Gujarati, Damodar N., 1988. Basic Econometrics. New York: McGraw-Hill Book
Company.
Gustely, Richard D., 1977. “The Allocational and Distributional Impacts of
Governmental Consolidation: The Dade County Experience.” Urban Affairs
Quarterly, Vol. 12, No. 3, pp. 349-364.
Harthshorn, Truman A. and Peter O. Muller, 1992. “The Suburban Downtown and
Urban Economic Development Today.” In Edwin S. Mills, John F.
McDonald, and Mary L. Mclean eds., Sources o f Metropolitan Growth, New
Brunswick: Center for Urban Policy Research, pp. 147-158.
Hawkins, Brett W. and Stephen L. Percy, 1991. “On Anti-Suburban Orthodoxy.”
Social Science Quarterly, Vol. 72, No. 3, pp. 478-490.
Hawkins, Brett W. and Stephen L. Percy, 1991. “Rejoinder to Neiman and Logan.”
Social Science Quarterly, Vol. 72, No. 3, pp. 501-503.
Herson, Lawrence J. R. and John M. Bolland, 1990. The Urban Web: Politics,
Policy, and Theory. Nelson-Hall Publishers: Chicago.
Hill, Melvin B., 1978. State Laws Governing Local Government Structure and
Administration. Athens: Institute of Government, University of Georgia.
Hirsch, Werner Z., 1995. “Contracting Out by Urban Governments: A Review.”
Urban Affairs Quarterly, Vol. 30, No. 3, pp. 458-472.
Hofstadter, Richard, 1955. The Age of Reform: From Bryan to F.D.R. New York:
Vintage Books.
Hughes, Holly L., 1993. “Metropolitan Structure and the Suburban Hierarchy.”
American Sociological Review, Vol. 58 (June), pp. 417-433.
1 3 7
ICMA, 1992. Service Contracting: A Local Government Guide. Washington, DC:
International City/County Management Association.
1 9 8 9. Service Delivery in the 90s: Alternative Approaches For Local
Governments. Special Report. Washington, DC: International City/County
Management Association.
Inman Robert P., 1979. “The Fiscal Performance of Local Governments: An
Interpretative Review.” in Peter Mieszkowski and Mahlon Straszheim eds.,
Current Issues in Urban Economics. Baltimore: Johns Hopkins University
Press, pp. 270-321.
Jeffery, Blake R., Tanis J. Salant, and Alan L. Boroshok, 1989. County Government
Structure: A State by State Report. Washington, DC: National Association of
Counties.
Kantor, Paul, 1988. The Dependent City: The Changing Political Economy of Urban
America. Glenview: Scott, Foresman and Company.
Karnig, Albert K., 1975. “Private-Regarding Policy, Civil Rights Groups, and the
Mediating Impact of Municipal Reforms.” American Journal of Political
Science, Vol. 19, No. I, pp. 91-106.
Keating, Ann Durkin, 1989. “Real Estate Developers: Creators of Improved
Subdivisions, Mentors of Suburban Government.” In Barbara M. Kelly ed.,
Suburbia Re-examined. New York: Greenwood Press, pp. 199-206.
Kenyon, Daphne A. and John Kincaid eds., 1991. Competition among States and
Local Governments: Efficiency and Equity in American Federalism.
Washington, DC: The Urban Institute Press.
Klinger, Ann, 1991. “County Leadership and Models for Change.”
Intergovernmental Perspective, Vol. 17, No. 1, pp. 45-48.
Kneier, Charles M., 1930a. “Development of Newer County Functions.” American
Political Science Review, Vol. 24 (February), pp. 134-140.
1930b. “The Legal Nature and Status of the American County.” Minnesota
Law Review, Vol. 14, pp. 141-156.
1 3 8
Knoke, David, 1982. “The Spread of Municipal Reform: Temporal, Spatial, and
Social Dynamics.” American Journal of Sociology, Vol. 87, No. 6, pp.
1314-1339.
Koehler, Cortus T., 1983. Managing California Counties: Serving People Solving
Problems. Sacramento: County Supervisors Association of California.
Ladd, Helen F. and William Wheaton, 1991. “Causes and Consequences of the
Changing Urban Form.” Regional Science and Urban Economics, Vol. 21,
pp. 157-162.
Lawrence, Carolyn B. and John M. DeGrove, 1976. “County Government
Services” in The County Year Book 1976, Washington, DC: National
Association of Counties, pp. 91-99.
Lewis, Edward B. and George A. Taylor, 1994. “Policy Making/Implementing
Activities of Elected County Executives and Appointed County
Administrators: Does Form of Government Make a Difference? ”
International Journal o f Public Administration, Vol. 17, No. 5, pp. 935-954.
Liebert, Ronald J., 1976. Disintegration and Political Action: The Changing
Functions o f City Governments in America. New York: Academy Press.
1974. “Municipal Functions, Structure, and Expenditures: A Reanalysis of
Recent Research.” Social Science Quarterly, Vol. 54 (March), pp. 765-783.
Lineberry, Robert L. and Edmund P. Fowler, 1967. “Reformism and Public
Policies in American Cities.” American Political Science Review, Vol. 61,
pp. 701-717.
Logan, John R., 1991. “Blaming the Suburbs? ”, Social Science Quarterly, Vol. 72,
No. 3, pp. 496-500.
Logan, John R. and Harvey L. Molotch, 1987. Urban Fortunes: The Political
Economy o f Place. Berkeley: University of California Press.
Lyons, William, 1978. “Reform and Response in American Cities: Structure and
Policy Reconsidered.” Social Science Quarterly, Vol. 59, No. 1, pp. 118-
132.
1977. “Urban Structures and Policy: Reassessing Additive Assumptions of
Reform.” Political Methodology, Vol. 4, No. 2, pp. 213-226.
1 3 9
Lyons, William E., David Lowery, and Ruth Hoogland DeHoog, 1992. The Politics
o f Dissatisfaction: Citizens, Services, and Urban Institutions. Armonk, NY:
M. E. Sharpe, Inc.
MacManus, Susan A. and William J. Pammer, 1990. “Cutbacks in The Country:
Retrenchment in Rural Village, Townships, and Counties.” Public
Administration Quarterly, Vol. 14, pp. 302-323.
Marando, Vincent L., 1979. “City-County Consolidation: Reform, Regionalism,
Referenda and Requiem.” Western Political Quarterly, Vol. 32 (December),
pp. 409-421.
Marando, Vincent L. and C. Douglas Baker, 1993. “Metropolitan Counties and
Urbanization.” In David R. Berman ed., County Governments in An Era of
Change. Westport, CT: Greenwood Press, pp. 71-87.
Marando, Vincent L. and Mavis Mann Reeves, 1991a. “Counties as local
Governments: Research Issues and Questions.” Journal o f Urban Affairs,
Vol. 13, No. 1, pp. 45-53.
1991b. “Counties: Evolving Local Governments, Reform and
Responsiveness.” National Civic Review, Vol. 80 (Spring), pp. 222-226.
Marando, Vincent L. and Robert D. Thomas, 1977. The Forgotten Governments:
County Commissions as Policy Makers. Gainesville: A Florida Atlantic
University Book.
March, James G., 1992. “The War Is Over, The Victors Have Lost.” Journal o f
Public Administration Research and Theory, Vol. 2, No. 3, pp. 225-231.
Martin, David L., 1990. Running City Hall: Municipal Administration in America.
Tuscaloosa: The University of Alabama Press.
Martin, Lawrence L., 1993. “American County Government: An Historical
Perspective.” In David R. Berman ed., County Governments in An Era o f
Change. Westport, CT: Greenwood Press, pp. 1-13.
Martin, Lawrence L. and Ronald C. Nyhan, 1994. “Determinants of County Home
Rule.” International Journal of Public Administration, Vol. 17, No. 5, pp.
955-970.
1 4 0
McArthur, Alastair, 1971. “3,049 Labs for Local Government Testing.” Public
Management, Vol. 53, No. 4, pp. 3-5.
Menzel, Donald C., Vincent L. Marando, Roger B. Parks, William L. Waugh,
Beverly A. Cigler, James H. Svara, Mavis Mann Reeves, J. Edwin Benton,
Robert D. Thomas, Gregory Streib, and Mark Schneider, 1992. “Setting a
Research Agenda for the Study of the American County.” Public
Administration Review, Vol. 52, No. 2, pp. 173-182.
Miller, Delbert C., 1991. Handbook o f Research Design and Social Measurement.
Newburry Park, CA: SAGE Publications.
Miller, Gary J., 1981. Cities by Contract: The Politics o f Municipal Incorporation.
Cambridge: The MIT Press.
Miranda, Rowan A. and Norman Walzer, 1994. “Growth and Decline of City
Government.” In Terry Nichols Clark ed., Urban Innovation: Creative
Strategies fo r Turbulent Times. Thousand Oaks, CA: SAGE Publications, pp.
146-166.
Mohl, Raymond A., 1985. The New City: Urban America in the Industrial Age,
1860-1920. Arlington Heights, IL: Harlan Davison, Inc.
Morgan, David R. and John P. Pelissero, 1980. “Urban Policy: Does Political
Structure Matter? ” American Political Science Review, Vol. 74, No. 4, pp.
999-1006.
Mueller, Dennis C., 1989. Public Choice II. New York: Cambridge University
Press.
Muller, Peter O., 1989. “The Transformation of Bedroom Suburbia into the Outer
City: An Overview of Metropolitan Structural Change Since 1947.” In
Barbara M. Kelly ed., Suburbia Re-examined. New York: Greenwood Press,
pp. 37-44.
Murray, Charles, 1993. “The Coming White Underclass.” Wall Street Journal,
October 29, Editorial Page.
National Association of Counties, 1978. The County Year Book, 1978. Washington,
DC: NACo.
141
Neiman, Max, 1991. “What, Me hate the Suburbs? The Irrelevance of Anti-
Suburban Orthodoxy.” Social Science Quarterly, Vol. 72, No. 3, pp. 491-
495.
Norusis, Marija J., 1994. SPSS fo r Windows: Base System User’ s Guide Release
6.1. Chicago: SPSS Inc.
North, Douglass C., 1994. “Economic Performance Through Time,” American
Economic Review, Vol. 84, No. 3, pp. 359-368.
1990. Institutions, Institutional Change and Economic Performance. New
York: Cambridge University Press.
O’Connell, Brian J., 1989. “The Federal Role in the Suburban Boom.” In Barbara
M. Kelly ed., Suburbia Re-examined. New York: Greenwood Press, pp.
183-192
Ostrom, Elinor, 1991. “Rational Choice Theory and Institutional Analysis: Toward
Complementary.” American Political Science Review, Vol. 85, No. 1, pp. 237-
243.
1986. “Agenda For the Study of Institutions,” Public Choice, Vol. 48, pp.
3-25.
1972. “Metropolitan Reform: Propositions Derived From Two Traditions.”
Social Science Quarterly, Vol. 53, No. 4, pp. 474-493.
Ostrom, Vincent, Robert Bish, and Elinor Ostrom, 1988. Local Government in the
United States. San Francisco: ICS Press.
Ostrom, Vincent, Charles M. Tieboutand Robert Warren, 1961. “The Organization
of Government in Metropolitan Areas: A Theoretical Inquiry.” American
Political Science Review, Vol. 55, pp. 831-842.
Park, Kee Ok, 1994. “Expenditure Patterns and Interactions among Local
Governments in Metropolitan Areas.” Urban Affairs Quarterly, Vol. 29, No.
4, pp. 535-564.
Parks, Roger B., 1991. “Counties in the Federal System: The Interlocal
Connection.” Intergovernmental Perspective, Vol. 17, No. 1, pp. 29-32.
142
Peterson, Don, 1977. “The Functional Role of the County. In Linda K.
Ganschinietz ed., Decade fo r Decisions: 1976-1986. Washington, DC:
National Association of Counties, pp. 47-53.
Peterson, Paul E., 1987. “Analyzing Developmental Politics: A Response to
Sanders and Stone.” Urban Affairs Quarterly, Vol. 22, No. 4, pp. 540-547.
1981. City Limits. Chicago: The University of Chicago Press.
Pfeffer, Jeffrey, 1982. Organizations and Organization Theory. Marshfield,
Massachusetts: Pitman Publishing Inc.
Phares, Donald, 1989. “Bigger Is Better, or Is It Smaller? : Restructuring Local
Government in the St. Louis Area.” Urban Affairs Quarterly, Vol. 25, No.
1, pp. 5-17.
Porter, Kirk H., 1922. County and Township Government in The United States.
New York: The Macmillan Company.
Prager, Jonas, 1994. “Contracting Out Government Services: Lessons From the
Private Sector,” Public Administration Review, Vol. 54 (March/April), pp. 176-
184.
Ramos, Alberto G., 1981. The New Science of Organizations: A Reconceptualization
o f the Wealth o f Nations. Toronto: University of Toronto Press.
Reason Foundation, 1994. Competitive Government fo r a Competitive Los Angeles.
Edited by William D. Eggers. Los Angeles: Reason Foundation.
Richardson, Harry W., 1988. “Monocentric vs. Polycentric Models: The Future of
Urban Economics in Regional Science.” The Annals o f Regional Science,
Vol. 22, pp. 1-12.
Ross, Bernard H., Myron A. Levine, Murray S. Stedman, 1991. Urban Politics:
Power in Metropolitan America. Itasca, IL: F. E. Peacock Publishers, Inc.
Rusk, David, 1993. Cities Without Suburbs. Washington, DC: The Woodrow
Wilson Center Press.
Salant, Tanis J., 1993. “Shifting Roles in County-State Relations.” In David R.
Berman ed., County Governments in An Era of Change. Westport, CT:
Greenwood Press, pp. 107-121.
143
1991. “County Government: An Overview.” Intergovernmental Perspective,
Vol. 17, No. 1, pp. 5-9.
1988. County Home Rule: Challenging the Tenets of Reform With Emphasis
on Counties in Arizona. DPA Dissertation. Sacramento: Sacramento Center,
School of Public Administration, University of Southern California
Salant, Tanis J. and Lawrence L. Martin, 1993. “County Constitutional Officers: A
Preliminary Investigation.” State and Local Government Review, Vol. 25, No.
3, pp. 164-172.
Sanders, Heywood T. and Clarence N. Stone, 1987. “Developmental Politics
Reconsidered.” Urban Affairs Quarterly, Vol. 22, No. 4, pp. 521-539 and
548-551.
Schiesl, Martin J., 1977. The Politics o f Efficiency: Municipal Administration and
Reform in America: 1880-1920. Berkeley: University of California Press.
Schneider, Mark, 1989. The Competitive City: The Political Economy of Suburbia.
Pittsburgh: University of Pittsburgh Press.
Schneider, Mark and Kee Ok Park, 1989. “Metropolitan Counties as Service
Delivery Agents: The Still Forgotten Governments.” Public Administration
Review, Vol. 49 (July/August), pp. 345-352.
Schneider, Mark and Paul Teske, 1993. “The Progrowth Entrepreneur in Local
Government.” Urban Affairs Quarterly, Vol. 29, No. 2, pp. 316-327.
Scott, W. Richard, 1992. Organizations: Rational, Natural, and Open Systems.
Englewood Cliffs, NJ: Prentice-Hall
Shannon, John, 1991. “ Federalism’s ‘Invisible Regulator’-Interjurisdictional
Competition.” in Daphne A. Kenyon and John Kincaid eds., Competition
among States and Local Governments: Efficiency and Equity in American
Federalism. Washington, DC: The Urban Institute Press, pp. 117-125.
Sharp, Elaine B., 1986. “The Politics and Economics of the New City Debt.”
American Political Science Review, Vol. 80, No. 4, pp. 1271-1288.
144
Simon, Herbert A., 1985. “Human Nature in Politics: The Dialogue of Psychology
With Political Science.” American Political Science Review, Vol. 79, pp. 293-
304.
Snider, Clyde F ., 1952. “ American County Government: A Mid-Century Review.”
American Political Science Review, Vol. 46, No. 1, pp. 66-80.
Stanback, Thomas M., 1991. The New Suburbanization: Challenge to the Central
City. Boulder: Westview Press.
Stein, Robert M., 1989. “Market Maximization of Individual Preferences and
Metropolitan Municipal Service Responsibility.” Urban Affairs Quarterly,
Vol. 25, No. 1, pp. 86-116.
1982. “The Political Economy of Municipal Functional Responsibility.”
Social Science Quarterly, Vol. 63, No. 3, pp. 530-548.
Streib, Gregory and William L. Waugh, 1991a. “The Changing Responsibilities of
County Government: Data from a National Survey of County Leaders.”
American Review o f Public Administration, Vol. 21, No. 2, pp. 139-156.
Streib, Gregory and William L. Waugh, 1991b. “Probing the Limits of County
Reform in an Ear o f Scarcity: A National Survey of County Administrators
and Executives.” Public Administrative Quarterly, Fall, pp. 378-395.
Teaford, Jon C ., 1993. The Twentieth-Century American City. Baltimore: The Johns
Hopkins University Press.
Teske, Paul and Mark Schneider, 1994. “The Bureaucratic Entrepreneur: The Case
of City Managers.” Public Administration Review, Vol. 54, No. 4, pp. 331-
340.
Thomas, John P., 1991a. “Financing County Government: An Overview.”
Intergovernmental Perspective, Vol. 17, No. 1, pp. 10-13.
Thomas, Robert D., 1991b. “Counties in Transition: Issues and Challenges.”
Intergovernmental Perspective, Vol. 17, No. 1, pp. 41-44.
Todd, Barbara, 1991. “Counties in the Federal System: The State Connection.”
Intergovernmental Perspective, Vol. 17, No. 1, pp. 21-25.
1 4 5
U.S. Bureau of the Census, 1994a. Census of Governments 1992: Government
Organization, Volume 1 Number 1. Washington, DC: U.S. Department of
Commerce.
1994b. USA Counties 1994: A Statistical Abstract Supplement [CD-Rom
Version]. Washington, DC: U.S. Department of Commerce.
1994c. Population o f Counties by Decennial Census: 1900 to 1990
Washington, DC: U.S. Department of Commerce.
1993. Current Population Reports, Series P23-185, Population Profile o f the
United States: 1993. Washington, DC: U.S. Department of Commerce.
1992. Annual Survey o f Governments, 1990: Finance Statistics [Computer
File]. Washington, DC: U.S. Department of Commerce.
1990a. Census of Governments 1987: Government Organization File
[Computer File]. Washington, DC: U.S. Department of Commerce.
1990b. Census of Government 1987: Local Government Finances File
[Computer File A]. Washington, DC: U.S. Department of Commerce.
1990c. Census o f Governments 1987: Compendium of Government Finances.
Volume 4, Number 5. Washington, DC: U.S. Department of Commerce.
1975. Historical Statistics o f the United States, Colonial Times to 1972. Part
I. Washington, DC: U.S. Department of Commerce.
Various Years. Statistical Abstract o f the United States. Washington, DC:
U.S. Department of Commerce.
Waugh, William L. and Gregory Streib, 1993. “County Capacity and
Intergovernmental Relations.” In David R. Berman ed., County Governments
in An Era o f Change. Westport, CT: Greenwood Press, pp. 43-52.
Welch, Susan and Timothy Bledsoe, 1988. Urban Reform and Its Consequences: A
Study in Representation. Chicago: The University of Chicago Press.
Wilson, James Q. and Edward C. Banfield, 1971. “Political Ethos Revisited.”
American Political Science Review, Vol. 65 (December), pp. 1048-1062.
146
1964. “Public-Regardingness as a Value Premise in Voting Behavior.”
American Political Science Review, Vol. 58, pp. 876-887.
Wolfinger, Raynold E. and John Osgood Field, 1966. “Political Ethos and the
Structure of City Government.” American Political Science Review, Vol. 60,
pp. 306-326.
Wong, Kenneth K., 1988. “Economic Constraint and Political Choice in Urban
Policy-Making.” American Journal of Political Science, Vol. 32, No. 1, pp. 1-
18.
Zax, Jeffrey, 1989. “Is There a Leviathan in Your Neighborhood? ” The American
Economic Review, Vol. 79, No. 3, pp. 560-567.
Zeller, Florence, 1975. “Forms of County Government.” in The County Year Book
1975, Washington, DC: National Association of Counties, pp. 27-36.
APPENDIX
Variable Descriptions
Dependent Variables
1. LOGEX log (base 10) of total expenditures per capita
2. LOGTAX log (base 10) of total tax revenues per capita
3. LOGUEX log (base 10) of total urban expenditures per capita
Independent Variables
4. REFORM county structural reform index
5. EXECU county executive form
6. ADMI county administrator form
7. ATLARG at-large election
8. POP90 population of 1990
9. DENSITY population per square mile, 1990
10. POPCH98 population change of 1980-1990
11. INCOME per capita personal iticome, 1990
12. POVERTY percent persons below poverty level, 1989
13. SCOPE county functional scope
14. MDFUNC state functional mandates
15. PRTAX property tax limit
16. DEBT debt limit
17. NEWENG New England counties
18. NOEAST Mid-Atlantic counties
19. NOCENT North Central counties
20. SOUTH South counties
21. AGE age of county by decades
22. HRULE home rule charter county surveyed in 1989
23. UNIN unincorporation rate of 1987
24. ROWOFF number of elected row officer positions, 1987
25. DISTRICT number of special districts, 1987
1 4 8
ALL-COUNTY LEVEL
TOTAL EXPENDITURES WITH REFORM INDEX
Multiple R .78741
R Square .62002
Adjusted R Square .61668
Standard Error .20442
Analysis of Variance
DF Sum of Squares
Regression 19 147.28127
Residual 2160 90.26246
F = 185.49854 Signif F = .0000
****MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGEX log (base 10) of expenditure
Variables in the Equation
Variable B SEB Beta T SigT
REFORM .072964 .009157 .180818 7.968 .0000
POP90 2.09562E-08 2.2952E-08 .018953 .913 .3613
DENSITY 1.15235E-05 1.1780E-05 .017002 .978 .3281
POPCH98 -.002810 3.0217E-04 -.148396 -9.299 .0000
INCOME 9.85862E-06 1.9583E-06 .102146 5.034 .0000
POVERTY 7.19333E-04 8.8840E-04 .016111 .810 .4182
SCOPE .071306 .002050 .607770 34.778 .0000
MDFUNC .001886 5.4137E-04 .062520 3.483 .0005
PRTAX -.015755 .011193 -.020552 -1.408 .1594
DEBT -.090853 .012939 -.109806 -7.022 .0000
NEWENG -.523697 .045414 -.199401 -11.531 .0000
NOEAST -.072457 .027341 -.048046 -2.650 .0081
NOCENT -.063483 .015713 -.091273 -4.040 .0001
SOUTH -.151853 .018233 -.228637 -8.328 .0000
AGE -.006241 .001774 -.065300 -3.517 .0004
HRULE -.129780 .027297 -.081066 -4.754 .0000
UNIN .001376 2.5479E-04 .090550 5.402 .0000
ROWOFF -.002104 .002790 -.013481 -.754 .4509
DISTRICT -.001138 3.0415E-04 -.083211 -3.742 .0002
(Constant) 1.838030 .060841 30.211 .0000
T o ta l C a s e s = 2 1 8 0
Mean Square
7.75165
.04179
1 4 9
T O T A L E X P E N D I T U R E S W I T H T H R E E R E F O R M I T E M S
Multiple R .78691
R Square .61923
Adjusted R Square .61553
Standard Error .20529
Analysis of Variance
Regression
Residual
F = 167.50454
DF Sum of Squares Mean Square
21 148.24300 7.05919
2163 91.15591 .04214
Signif F =.0000
***MULTIPLE REGRESSION * * * *
Equation Number 1 Dependent Variable.. LOGEX log (base 10) of expenditure
Variables in the Equation
Variable B SE B Beta T SigT
EXECU .144224 .016395 .144071 8.797 .0000
ADMI .051687 .012035 .072640 4.295 .0000
ATLARG -.010727 .009824 -.015656 -1.092 .2750
POP90 2.13320E-08 2.3050E-08 .019219 .925 .3548
DENSITY 9.74861 E-06 1.1850E-05 .014329 .823 .4108
POPCH98 -.002512 3.0740E-04 -.132198 -8.172 .0000
INCOME 1.02879E-05 1.9651 E-06 .106377 5.235 .0000
POVERTY 6.20355E-04 8.9543E-04 .013857 .693 .4885
SCOPE .073531 .002074 .624793 35.453 .0000
MDFUNC .002226 5.4349E-04 .073586 4.096 .0000
PRTAX .001044 .011357 .001357 .092 .9268
DEBT -.085541 .013098 -.103015 -6.531 .0000
NEWENG -.471381 .045524 -.178788 -10.355 .0000
NOEAST -.082122 .027613 -.054248 -2.974 .0030
NOCENT -.060178 .016071 -.086265 -3.745 .0002
SOUTH -.150604 .018230 -.226152 -8.261 .0000
AGE -.005586 .001788 -.058242 -3.123 .0018
HRULE -.086051 .025638 -.053808 -3.356 .0008
UNIN .001487 2.5282E-04 .097662 5.882 .0000
ROWOFF -.014501 .002649 -.092755 -5.473 .0000
DISTRICT -.001159 3.0559E-04 -.084456 -3.793 .0002
(Constant)
Total Cases
1.865175
= 2185
.061857 30.153 .0000
1 5 0
T O T A L T A X E S W I T H R E F O R M I N D E X
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 134.03744
.73519
.54051
.53647
.30717
DF
19
2165
SignifF =
Sum of Squares
240.29367
204.27749
.0000
Mean Square
12.64704
.09435
♦ ♦ ♦♦ m u l t i p l e REGRESSION ****
Equation Number 1 Dependent Variable.. LOGTAX
------------------Variables in the Equation —
log (base 10) of taxes
Variable B SEB Beta T SigT
REFORM .214531 .013770 .388657 15.580 .0000
POP90 1.04210E-08 3.4613E-08 .006888 .301 .7634
DENSITY 1.36495E-05 1.7705E-05 .014718 .771 .4408
POPCH98 -.006163 4.5532E-04 -.237421 -13.536 .0000
INCOME -1.10659E-05 2.9722E-06 -.083756 -3.723 .0002
POVERTY -.002075 .001335 -.034128 -1.554 .1203
SCOPE .054708 .003073 .339395 17.805 .0000
MDFUNC .006319 8.1133E-04 .153730 7.788 .0000
PRTAX -.079446 .016840 -.075397 -4.718 .0000
DEBT -.132828 .019509 -.116350 -6.809 .0000
NEWENG -1.087734 .072162 -.275962 -15.074 .0000
NOEAST -.261695 .041136 -.126854 -6.362 .0000
NOCENT -.092458 .023508 -.097382 -3.933 .0001
SOUTH -.340463 .027330 -.375130 -12.457 .0000
AGE -.001719 .002677 -.013107 -.642 .5208
HRULE -.383351 .040783 -.175904 -9.400 .0000
UNIN .004790 3.8436E-04 .230123 12.463 .0000
ROWOFF -.013789 .004193 -.064478 -3.289 .0010
DISTRICT 3.38315E-04 4.6098E-04 .018024 .734 .4631
(Constant)
Total Cases
1.610533
= 2185
.091445 17.612 .0000
151
T O T A L T A X E S W I T H T H R E E R E F O R M I T E M S
Multiple R .73591
R Square .54157
Adjusted R Square .53712
Standard Error .30714
Analysis of Variance
Regression
Residual
F = 121.79211
DF
21
2165
Sum of Squares
241.27351
204.23449
SignifF = .0000
Mean Square
11.48921
.09433
****MULTIPLE REGRESSION****
Equation Number 1 Dependent Variable.. LOGTAX log (base 10) of taxes
----------------- Variables in the Equation-----------------
Variable B SEB
EXECU .355871 .024521
ADMI .204316 .017999
ATLARG .035054 .014707
POP90 1.28438E-08 3.4612E-08
DENSITY 1.52539E-05 1.7739E-05
POPCH98 -.005492 4.5867E-04
INCOME -9.95992E-06 2.9697E-06
POVERTY -.002163 .001338
SCOPE .058786 .003103
MDFUNC .007356 8.1209E-04
PRTAX -.045027 .017010
DEBT -.144125 .019670
NEWENG -.956600 .072124
NOEAST -.275274 .041411
NOCENT -.069281 .023964
SOUTH -.321973 .027220
AGE -2.95288E-04 .002692
HRULE -.218278 .038387
UNIN .005243 3.8017E-04
ROWOFF -.040919 .003973
DISTRICT 3.42770E-04 4.6139E-04
(Constant) 1.633089 .092505
Total Cases = 2187
Beta T SigT
.261427 14.513 .0000
.210453 11.351 .0000
.037487 2.384 .0172
.008478 .371 .7106
.016420 .860 .3899
-.212608 -11.974 .0000
-.075338 -3.354 .0008
-.035539 -1.616 .1062
.364058 18.943 .0000
.178357 9.058 .0000
-.042748 -2.647 .0082
-.126126 -7.327 .0000
-.242439 -13.263 .0000
-.133299 -6.647 .0000
-.072935 -2.891 .0039
-.354477 -11.828 .0000
-.002247 -.110 .9127
-.100055 -5.686 .0000
.251851 13.792 .0000
-.191546 -10.299 .0000
.018251 .743 .4576
17.654 .0000
1 5 2
U R B A N E X P E N D I T U R E S W I T H R E F O R M I N D E X
Multiple R .81333
R Square .66151
Adjusted R Square .65833
Standard Error .37661
Analysis of Variance
Regression
Residual
F = 208.07858
DF Sum of Squares
19 560.74260
2023 286.93166
Mean Square
29.51277
.14183
Signif F = .0000
****MULTIPLE REGRESSION * * * *
Equation Number 1 Dependent Variable.. LOGUEX log (base 10) of uex
------------------Variables in the Equation------------------
Variable B SEB
REFORM .079956 .017272
POP90 -5.37357E-08 4.2795E-08
DENSITY 6.50122E-05 2.1792E-05
POPCH98 9.09700E-04 5.5490E-04
INCOME 1.13876E-05 3.7743E-06
POVERTY -.002051 .001692
SCOPE .150569 .003990
MDFUNC .002306 .001018
PRTAX .010398 .021135
DEBT .178518 .025259
NEWENG -.310399 .1076%
NOEAST .029289 .050861
NOCENT -.118934 .029435
SOUTH .003554 .034124
AGE .015374 .003390
HRULE -.065171 .050386
UNIN .002680 4.8298E-04
ROWOFF -.024353 .005297
DISTRICT -.001160 5.7652E-04
(Constant) -.783494 .115545
Total Cases = 2043
Beta T SigT
.102569 4.629 .0000
-.025506 -1.256 .2094
.050447 2.983 .0029
.025282 1.639 .1013
.060195 3.017 .0026
-.023508 -1.212 .2256
.622006 37.737 .0000
.039735 2.265 .0236
.006921 .492 .6228
.108598 7.067 .0000
-.043774 -2.882 .0040
.010263 .576 .5648
-.086380 -4.041 .0001
.002749 .104 .9171
.082141 4.536 .0000
-.021621 -1.293 .1960
.090923 5.550 .0000
-.080364 -4.598 .0000
-.044083 -2.012 .0444
-6.781 .0000
1 5 3
U R B A N E X P E N D I T U R E S W I T H T H R E E R E F O R M I T E M S
Multiple R .81383
R Square .66232
Adjusted R Square .65881
Standard Error .37714
Analysis of Variance
Regression
Residual
F = 188.85279
DF
21
2022
Sum of Squares
564.08754
287.59740
SignifF = .0000
Mean Square
26.86131
.14223
****MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGUEX log (base 10) of uex
Variables in the Equation
Variable B SEB Beta T SigT
EXECU .128216 .030967 .066936 4.140 .0000
ADMI .093978 .022455 .068673 4.185 .0000
ATLARG .005985 .018704 .004490 .320 .7490
POP90 -5.17725E-08 4.2862E-08 -.024517 -1.208 .2272
DENSITY 6.67890E-05 2.1856E-05 .051704 3.056 .0023
POPCH98 9.16475E-04 5.6270E-04 .025436 1.629 .1035
INCOME 1.15195E-05 3.7849E-06 .060749 3.044 .0024
POVERTY -.002109 .001704 -.024121 -1.238 .2159
SCOPE .152421 .004063 .628670 37.515 .0000
MDFUNC .002664 .001019 .045781 2.613 .0090
PRTAX .017333 .021445 .011511 .808 .4190
DEBT .175496 .025470 .106619 6.890 .0000
NEWENG -.248602 .108237 -.034976 -2.297 .0217
NOEAST .035502 .051231 .012411 .693 .4884
NOCENT -.108240 .030104 -.078437 -3.595 .0003
SOUTH .011187 .034066 .008635 .328 .7426
AGE .014754 .003407 .078742 4.331 .0000
HRULE .003974 .047570 .001315 .084 .9334
UNIN .002772 4.7936E-04 .093833 5.783 .0000
ROWOFF -.034111 .004958 -.112357
-6.879 .0000
DISTRICT -.001147 5.7768E-04 -.043480 -1.985 .0473
(Constant) -.769827 .117294 -6.563 .0000
T o ta l C a s e s = 2 0 4 4
1 5 4
Mean Square
2.83868
.03544
REGRESSION ****
METROPOLITAN COUNTIES
TOTAL EXPENDITURES WITH REFORM INDEX
.83729
.70105
.69229
.18825
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 80.10040 SignifF = .0000
****MULTIPLE
Equation Number 1 Dependent Variable.. LOGEX
DF Sum of Squar
19 53.9341
649 22.999<
Variable B
REFORM .089517
POP90 5.59946E-09
DENSITY 2.35325E-05
POPCH98 -.002076
INCOME 1.36437E-06
POVERTY -.002493
SCOPE .072448
MDFUNC 5.68743E-04
PRTAX -.030809
DEBT -.107102
NEWENG -.408841
NOEAST .014261
NOCENT -.019882
SOUTH -.066365
AGE -.004822
HRULE -.112818
UNIN 7.86844E-04
ROWOFF .009131
DISTRICT -4.50285E-04
(Constant) 1.912251
Total Cases = 669
- Variables in the Equation
SE B Beta
.015726 .246368
2.3688E-08 .008320
1.1550E-05 .056585
4.6223E-04 -.119922
3.1346E-06 .015114
.002158 -.039775
.003592 .668517
8.8545E-04 .020333
.019347 -.037801
.024254 -.130452
.073114 -.184203
.038347 .013268
.029559 -.026643
.034033 -.097469
.002737 -.054156
.033972 -.106172
4.2756E-04 .057565
.004843 .063488
3.6094E-04 -.046213
.107468 17.794
log (base 10) of expenditure
T SigT
5.692 .0000
.236 .8132
2.037 .0420
-4.492 .0000
.435 .6635
-1.155 .2484
20.169 .0000
.642 .5209
-1.592 .1118
-4.416 .0000
-5.592 .0000
.372 .7101
-.673 .5014
-1.950 .0516
-1.762 .0786
-3.321 .0009
1.840 .0662
1.885 .0598
-1.248 .2126
.0000
1 5 5
T O T A L E X P E N D I T U R E S W I T H T H R E E R E F O R M I T E M S
Multiple R .84250
R Square .70980
Adjusted R Square .70037
Standard Error .18575
Analysis of Variance
DF Sum of Squares Mean Square
Regression 21 54.51580 2.59599
Residual 646 22.28825 .03450
F = 75.24189 SignifF = .0000
♦♦♦♦MULTIPLE REGRESSION ****
Equation Number I Dependent Variable.. LOGEX log (base 10) of expenditure
Variables in the Equation
Variable B SEB Beta T SigT
EXECU .140267 .025862 .150565 5.424 .0000
ADMI .062008 .019105 .090445 3.246 .0012
ATLARG -.040080 .015897 -.057826 -2.521 .0119
POP90 1.52234E-08 2.3469E-08 .022638 .649 .5168
DENSITY 2.78683E-05 1.1426E-05 .067055 2.439 .0150
POPCH98 -.001811 4.626 IE-04 -.104574 -3.915 .0001
INCOME 9.84835E-07 3.0973E-06 .010917 .318 .7506
POVERTY -.003481 .002135 -.055578 -1.630 .1035
SCOPE .075344 .003590 .695429 20.990 .0000
MDFUNC .001129 8.6552E-04 .040346 1.304 .1926
PRTAX 6.35388E-04 .019840 7.782E-04 .032 .9745
DEBT -.094995 .024309 -.115780 -3.908 .0001
NEWENG -.331369 .072799 -.149422 -4.552 .0000
NOEAST .015500 .038652 .014431 .401 .6885
NOCENT -.022296 .029344 -.029849 -.760 .4476
SOUTH -.060764 .033638 -.089263 -1.806 .0713
AGE -.004191 .002704 -.047060 -1.550 .1216
HRULE -.055908 .029510 -.052655 -1.895 .0586
UNIN 9.94995E-04 4.1749E-04 .072803 2.383 .0174
ROWOFF -.003281 .004325 -.022823 -.759 .4483
DISTRICT -6.95830E-04 3.5955E-04 -.071161 -1.935 .0534
(Constant)
Total Cases
1.949709
= 668
.106391 18.326 .0000
1 5 6
T O T A L T A X E S W I T H R E F O R M I N D E X
Multiple R .76579
R Square .58643
Adjusted R Square .57432
Standard Error .30735
Analysis of Variance
DF
Regression 19
Residual 649
F = 48.43461 Signif F = .0000
****MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGTAX log (base 10) of taxes
------------------Variables in the Equation-----------------
Variable B SE B Beta T SigT
REFORM .159733 .025706 .316726 6.214 .0000
POP90 1.44344E-08 3.8691E-08 .015450 .373 .7092
DENSITY 2.42199E-05 1.8847E-05 .041950 1.285 .1992
POPCH98 -.006315 7.5630E-04 -.262167 -8.350 .0000
INCOME -1.49598E-05 5.0737E-06 -.119890 -2.949 .0033
POVERTY -.004492 .003528 -.051744 -1.273 .2034
SCOPE .060230 .005874 .399593 10.254 .0000
MDFUNC .007150 .001447 .184121 4.941 .0000
PRTAX -.063158 .031601 -.055825 -1.999 .0461
DEBT -.146784 .039608 -.128480 -3.706 .0002
NEWENG -.916559 .119028 -.297500 -7.700 .0000
NOEAST -.269796 .062341 -.180829 -4.328 .0000
NOCENT -.114063 .047971 -.110119 -2.378 .0177
SOUTH -.370713 .055056 -.392131 -6.733 .0000
AGE -9.65429E-04 .004471 -.007812 -.216 .8291
HRULE -.290717 .055418 -.197100 -5.246 .0000
UNIN .003801 6.9886E-04 .200177 5.439 .0000
ROWOFF -.015250 .007912 -.076376 -1.927 .0544
DISTRICT -4.08764E-05 5.8865E-04 -.003022 -.069 .9447
(Constant)
Total Cases
1.725242
= 669
.176137 9.795 .0000
Sum of Squares Mean Square
86.93015 4.57527
61.30638 .09446
1 5 7
T O T A L T A X E S W I T H T H R E E R E F O R M I T E M S
Multiple R .76891
R Square .59123
Adjusted R Square .577%
Standard Error .30603
Analysis of Variance
DF Sum of Squares Mean Square
Regression 21 87.64136 4.17340
Residual 647 60.59517 .09366
F = 44.56112 Signif F = .0000
** ** m u l t i p l e r e g r e s s i o n ****
Equation Number 1 Dependent Variable.. LOGTAX log (base 10) of taxes
Variables in the Equation
Variable B SEB Beta T SigT
EXECU .274637 .042696 .212228 6.432 .0000
ADMI .134242 .031460 .141207 4.267 .0000
ATLARG -9.40260E-04 .026201 -9.763E-04 -.036 .9714
POP90 2.23562E-08 3.8551E-08 .023929 .580 .5622
DENSITY 2.57643E-05 1.8815E-05 .044625 1.369 .1714
POPCH98 -.005850 7.6382E-04 -.242861 -7.659 .0000
INCOME -1.49579E-05 5.0548E-06 -.119875 -2.959 .0032
POVERTY -.005142 .003521 -.059228 -1.461 .1446
SCOPE .062889 .005914 .417238 10.634 .0000
MDFUNC .008280 .001426 .213215 5.805 .0000
PRTAX -.0267% .032558 -.023685 -.823 .4108
DEBT -.143954 .040056 -.126003 -3.594 .0004
NEWENG -.826225 .119540 -.268179 -6.912 .0000
NOEAST -.289355 .063339 -.193939 -4.568 .0000
NOCENT -.110735 .048041 -.106905 -2.305 .0215
SOUTH -.367997 .054925 -.389257 -6.700 .0000
AGE -5.53200E-05 .004454 -4.476E-04 -.012 .9901
HRULE -.179242 .048616 -.121522 -3.687 .0002
UNIN .004118 6.8795E-04 .216880 5.986 .0000
ROWOFF -.037278 .007090 -.186702 -5.258 .0000
DISTRICT -1.93102E-04 5.8652E-04 -.014275 -.329 .7421
(Constant) 1.781169 .175724 10.136 .0000
T o ta l C a s e s = 6 6 9
1 5 8
U R B A N E X P E N D I T U R E S W I T H R E F O R M IN D E X
Multiple R .85657
R Square .73371
Adjusted R Square .72543
Standard Error .34917
Analysis of Variance
DF Sum of Squares Mean Square
Regression 19 205.25529 10.80291
Residual 611 74.49301 .12192
F = 88.60668 Signif F = .0000
♦♦♦♦MULTIPLE REGRESSION ♦♦♦♦
Equation Number 1 Dependent Variable.. LOGUEX log (base 10) of uex
Variables in the Equation
Variable B SEB Beta T SigT
REFORM .091212 .030273 .128469 3.013 .0027
POP90 -5.77609E-08 4.4276E-08 -.044632 -1.305 .1925
DENSITY 7.38700E-05 2.1488E-05 .092512 3.438 .0006
POPCH98 .001952 8.7647E-04 .057854 2.228 .0263
INCOME 5.2473 IE-07 5.8937E-06 .002993 .089 .9291
POVERTY -.001371 .004080 -.011207 -.336 .7369
SCOPE .145583 .006958 .643653 20.923 .0000
MDFUNC .002297 .001678 .042452 1.369 .1716
PRTAX -.015249 .036728 -.009491 -.415 .6782
DEBT .050093 .047169 .030599 1.062 .2887
NEWENG -.313895 .182000 -.045751 -1.725 .0851
NOEAST .130128 .071942 .063247 1.809 .0710
NOCENT .070140 .055238 .047566 1.270 .2046
SOUTH .047064 .064432 .035244 .730 .4654
AGE .024742 .005282 .140146 4.684 .0000
HRULE -.097577 .064532 -.047698 -1.512 .1310
UNIN .003579 8.1858E-04 .133985 4.372 .0000
ROWOFF -.024999 .009256 -.088511 -2.701 .0071
DISTRICT -4.63396E-04 6.7960E-04 -.024641 -.682 .4956
(Constant)
Total Cases
-.652990
= 631
.204410 -3.195 .0015
1 5 9
U R B A N E X P E N D I T U R E S W I T H T H R E E R E F O R M I T E M S
Multiple R .85546
R Square .73182
Adjusted R Square .72258
Standard Error .35156
Analysis of Variance
DF
Regression 21
Residual 610
F = 79.26455 SignifF = .0000
****MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGUEX log (base 10) of uex
Variables in the Equation
Variable B SEB Beta T SigT
EXECU .122364 .050373 .067764 2.429 .0154
ADMI .106108 .037245 .079030 2.849 .0045
ATLARG .037893 .031015 .027801 1.222 .2223
POP90 -5.43892E-08 4.4613E-08 -.041930 -1.219 .2233
DENSITY 7.43578E-05 2.1678E-05 .092903 3.430 .0006
POPCH98 .002115 8.9359E-04 .062550 2.367 .0182
INCOME 9.31580E-07 5.9404E-06 .005305 .157 .8754
POVERTY -7.96319E-04 .004113 -.006498 -.194 .8465
SCOPE .144521 .007082 .637402 20.408 .0000
MDFUNC .002644 .001670 .048813 1.583 .1139
PRTAX -.007525 .038333 -.004674 -.196 .8444
DEBT .043480 .048314 .026501 .900 .3685
NEWENG -.282539 .184907 -.041081 -1.528 .1270
NOEAST .124620 .073915 .060429 1.686 .0923
NOCENT .082646 .055915 .055928 1.478 .1399
SOUTH .057714 .064971 .043153 .888 .3747
AGE .024206 .005324 .136987 4.546 .0000
HRULE .024220 .056363 .011878 .430 .6676
UNIN .003922 8.1208E-04 .146950 4.830 .0000
ROWOFF -.034277 .008286 -.121078 -4.137 .0000
DISTRICT -5.02973E-04 6.8469E-04 -.026698 -.735 .4629
(Constant)
Total Cases
-.641037
= 632
.206018 -3.112 .0019
Sum of Squares Mean Square
205.72684 9.79652
75.39152 .12359
1 6 0
NONMETROPOLITAN COUNTIES
TOTAL EXPENDITURES WITH REFORM INDEX
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 111.66025
.76538
.58581
.58057
.21292
DF Sum of Squares
19 96.18030
1500 68.00255
Signif F = .0000
Mean Square
5.06212
.04534
** * * m u l t i p l e REGRESSION ****
Equation Number 1 Dependent Variable.. LOGEX
------------------Variables in the Equation
log (base 10) of expenditure
Variable B SEB Beta T SigT
REFORM .097628 .011986 .216465 8.145 .0000
POP90 -1.71283E-06 4.8047E-07 -.133004 -3.565 .0004
DENSITY -2.25693E-05 2.1312E-04 -.003034 -.106 .9157
POPCH98 -.001750 4.5263E-04 -.081415 -3.866 .0001
INCOME 1.59988E-05 2.688 IE-06 .140071 5.952 .0000
POVERTY 7.40458E-04 .001100 .016843 .673 .5010
SCOPE .069288 .002519 .559816 27.505 .0000
MDFUNC .002092 7.0915E-04 .065750 2.950 .0032
PRTAX -.012208 .013992 -.016168 -.873 .3831
DEBT -.058690 .016146 -.069888 -3.635 .0003
NEWENG -.560332 .061101 -.189420 -9.171 .0000
NOEAST -.052650 .044512 -.024028 -1.183 .2371
NOCENT -.068260 .020369 -.099997 -3.351 .0008
SOUTH -.176019 .023893 -.265773 -7.367 .0000
AGE .008424 .004145 .055290 2.032 .0423
HRULE -.125737 .050838 -.044657 -2.473 .0135
UNIN .001338 3.4001E-04 .081269 3.934 .0001
ROWOFF -.003471 .003499 -.021386 -.992 .3213
DISTRICT -.001677 5.6225E-04 -.077507 -2.983 .0029
(Constant)
Total Cases =
1.795445
1520
.079500 22.584 .0000
161
T O T A L E X P E N D I T U R E S W I T H T H R E E R E F O R M I T E M S
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 104.12645
.77046
.59361
.58791
.21107
DF
21
1497
Sum of Squares
97.41505
66.69105
SignifF = .0000
Mean Square
4.63881
.04455
**** m u l t i p l e REGRESSION ****
Equation Number 1 Dependent Variable.. LOGEX
------------Variables in the Equation
log (base 10) of expenditure
Variable B SEB Beta T SigT
EXECU .189235 .021213 .181506 8.921 .0000
ADMI .079720 .016082 .107418 4.957 .0000
ATLARG -.001340 .012391 -.001957 -.108 .9139
POP90 -1.76713E-06 4.8124E-07 -.137269 -3.672 .0002
DENSITY 6.23291 E-05 2.1157E-04 .008381 .295 .7683
POPCH98 -.001582 4.5143E-04 -.073421 -3.504 .0005
INCOME 1.67430E-05 2.6754E-06 .146587 6.258 .0000
POVERTY 9.77037E-04 .001098 .022221 .890 .3737
SCOPE .073081 .002541 .590218 28.757 .0000
MDFUNC .002593 7.065 IE-04 .081456 3.670 .0003
PRTAX 6.22602E-04 .013898 8.247E-04 .045 .9643
DEBT -.059212 .015998 -.070521 -3.701 .0002
NEWENG -.490390 .060762 -.165815 -8.071 .0000
NOEAST -.063827 .044174 -.029135 -1.445 .1487
NOCENT -.059757 .020915 -.087512 -2.857 .0043
SOUTH -.170216 .024000 -.257047 -7.092 .0000
AGE .009389 .004126 .061631 2.276 .0230
HRULE -.071452 .049503 -.025383 -1.443 .1491
UNIN .001513 3.3523E-04 .091864 4.514 .0000
ROWOFF -.018939 .003312 -.116438 -5.719 .0000
DISTRICT -.001417 5.5806E-04 -.065515 -2.540 .0112
(Constant)
Total Cases =
1.797359
1519
.080503 22.327 .0000
1 6 2
T O T A L T A X E S W I T H R E F O R M I N D E X
Multiple R .72805
R Square .53005
Adjusted R Square .52412
Standard Error .31054
Analysis of Variance
DF Sum of Squares Mean Square
Regression 19 163.58853 8.60992
Residual 1504 145.03792 .09643
F = 89.28233 Signif F = .0000
♦ ♦♦♦MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGTAX log (base 10) of taxes
Variables in the Equation
Variable B SEB Beta T SigT
REFORM .247954 .017496 .399130 14.172 .0000
POP90 -9.87583E-07 7.0057E-07 -.055212 -1.410 .1588
DENSITY 2.18478E-04 3.1307E-04 .021276 .698 .4854
POPCH98 -.004258 6.5723E-04 -.144531 -6.478 .0000
INCOME 2.69779E-06 3.9574E-06 .017178 .682 .4955
POVERTY -.001427 .001599 -.023773 -.892 .3724
SCOPE .045468 .003662 .267508 12.415 .0000
MDFUNC .004923 .001033 .113048 4.766 .0000
PRTAX -.097102 .020413 -.093434 -4.757 .0000
DEBT -.105239 .023748 -.090938 -4.431 .0000
NEWENG -1.676838 .097755 -.367851 -17.154 .0000
NOEAST -.261288 .065091 -.086974 -4.014 .0001
NOCENT -.114526 .02% 17 -.122556 -3.867 .0001
SOUTH -.360454 .034778 -.397500 -10.364 .0000
AGE -8.64266E-04 .006099 -.004072 -.142 .8873
HRULE -.466388 .074246 -.120817 -6.282 .0000
UNIN .005935 4.985 IE-04 .262939 11.906 .0000
ROWOFF -.018547 .005122 -.083031 -3.621 .0003
DISTRICT .001074 8.3145E-04 .035922 1.292 .1967
(Constant) 1.514658 .116160 13.039 .0000
T o ta l C a s e s = 1 5 2 4
1 6 3
T O T A L T A X E S W I T H T H R E E R E F O R M I T E M S
Multiple R .72441
R Square .52476
Adjusted R Square .51813
Standard Error .31123
Analysis of Variance
Regression
Residual
F = 79.08302 Signif F = .0000
DF
21
1504
Sum of Squares
160.87106
145.68781
Mean Square
7.66053
.09687
♦♦♦♦MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGTAX log (base 10) of taxes
Variables in the Equation
Variable B SEB Beta T SigT
EXECU .396151 .031179 .279511 12.706 .0000
ADMI .234264 .023707 .230600 9.881 .0000
ATLARG .054047 .018241 .057895 2.963 .0031
POP90 -8.26056E-07 7.1079E-07 -.046384 -1.162 .2454
DENSITY 2.47478E-04 3.1466E-04 .024207 .786 .4317
POPCH98 -.004274 6.7538E-04 -.142669 -6.329 .0000
INCOME 3.51780E-06 3.968 IE-06 .022484 .887 .3755
POVERTY -8.06028E-04 .001605 -.013488 -.502 .6156
SCOPE .053581 .003734 .315978 14.350 .0000
MDFUNC .006572 .001040 .151457 6.321 .0000
PRTAX -.065381 .020452 -.063192 -3.197 .0014
DEBT -.128496 .023772 -.111273 -5.405 .0000
NEWENG -1.417540 .100666 -.301537 -14.082 .0000
NOEAST -.268169 .065294 -.089567 -4.107 .0000
NOCENT -.073336 .030672 -.078832 -2.391 .0169
SOUTH -.317358 .035392 -.351447 -8.967 .0000
AGE -6.76973E-04 .006137 -.003201 -.110 .9122
HRULE -.273693 .073049 -.071139 -3.747 .0002
UNIN .006320 4.9520E-04 .281075 12.763 .0000
ROWOFF -.046228 .004921 -.207169 -9.394 .0000
DISTRICT .001344 8.3343E-04 .045120 1.613 .1070
(Constant) 1.455458 .118523 12.280 .0000
T o ta l C a s e s = 1 5 2 6
1 6 4
U R B A N E X P E N D I T U R E S W I T H R E F O R M IN D E X
Multiple R .78914
R Square .62274
Adjusted R Square .61761
Standard Error .38864
Analysis of Variance
Regression
Residual
F = 121.28249
DF Sum of Squares Mean Square
19 348.05934 18.31891
1396 210.85650 .15104
Signif F = .0000
****MULTIPLE REGRESSION ****
Equation Number 1 Dependent Variable.. LOGUEX log (base 10) of uex
------------------Variables in the Equation------------------
Variable B SEB
REFORM .106309 .022410
POP90 2.73668E-07 8.9301E-07
DENSITY -4.64259E-04 4.0879E-04
POPCH98 .001137 8.3272E-04
INCOME 2.34640E-05 5.2914E-06
POVERTY -.002906 .0020%
SCOPE .148254 .004920
MDFUNC .002253 .001323
PRTAX 4.34697E-04 .026156
DEBT .262617 .031151
NEWENG -.347685 .136462
NOEAST .082677 .082283
NOCENT -.171099 .037964
SOUTH .012060 .044630
AGE .008764 .008003
HRULE -.027812 ,092%7
UNIN .002267 6.3990E-04
ROWOFF -.021375 .006675
DISTRICT -.002324 .001076
(Constant) -.936188 .151478
Total Cases = 1416
Beta T SigT
.124824 4.744 .0000
.011037 .306 .7593
-.032088 -1.136 .2563
.028146 1.366 .1722
.104121 4.434 .0000
-.034530 -1.387 .1658
.593286 30.132 .0000
.037538 1.703 .0888
3.007E-04 .017 .9867
.161355 8.430 .0000
-.048586 -2.548 .0109
.020432 1.005 .3152
-.129173 -4.507 .0000
.009555 .270 .7870
.029890 1.095 .2736
-.005351 -.299 .7649
.072098 3.543 .0004
-.068749 -3.202 .0014
-.055747 -2.160 .0309
-6.180 .0000
165
U R B A N E X P E N D I T U R E S W I T H T H R E E R E F O R M I T E M S
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 109.79888
.78934
.62305
.61738
.38963
DF Sum of Squares
21 350.05141
1395 211.78189
Signif F = .0000
Mean Square
16.66911
.15181
**** m u l t i p l e REGRESSION ****
Equation Number 1 Dependent Variable.. LOGUEX
------------------Variables in the Equation —
log (base 10) of uex
Variable B SE B Beta T SigT
EXECU .154590 .040509 .078748 3.816 .0001
ADMI .116150 .030094 .083153 3.860 .0001
ATLARG -.014619 .023748 -.011185 -.616 .5383
POP90 9.34055E-08 9.0552E-07 .003758 .103 .9179
DENSITY -3.83150E-04 4.0962E-04 -.026414 -.935 .3498
POPCH98 .001194 8.3803E-04 .029484 1.425 .1543
INCOME 2.37014E-05 5.3138E-06 .104944 4.460 .0000
POVERTY -.003122 .002118 -.037005 -1.474 .1406
SCOPE .150966 .005074 .602664 29.754 .0000
MDFUNC .002789 .001332 .046382 2.094 .0365
PRTAX .007576 .026333 .005228 .288 .7736
DEBT .256924 .031162 .157459 8.245 .0000
NEWENG -.277346 .137437 -.038656 -2.018 .0438
NOEAST .073936 .082514 .018224 .896 .3704
NOCENT -.171420 .039397 -.129167 -4.351 .0000
SOUTH .021006 .045330 .016604 .463 .6432
AGE .010051 .008062 .034190 1.247 .2127
HRULE .063354 .091337 .012157 .694 .4880
UNIN .002317 6.3792E-04 .073517 3.632 .0003
ROWOFF -.035393 .006296 -.113625 -5.622 .0000
DISTRICT -.002139 .001079 -.051167 -1.982 .0477
(Constant)
Total Cases =
-.893775
1417
.155241 -5.757 .0000
166
FORM OF 1978 IS EQUAL TO FORM OF 1989
TOTAL EXPENDITURES WITH TWO REFORM ITEMS
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 160.72461
.81099
.65770
.65361
.19568
DF
20
1673
Sum of Squares
123.09075
64.06325
Mean Square
6.15454
.03829
Signif F = .0000
**** m u l t i p l e REGRESSION ****
Equation Number 1 Dependent Variable.. LOGEX
------------------Variables in the Equation
log (base 10) of expenditure
Variable B SE B Beta T SigT
EXECU .031519 .025166 .021817 1.252 .2106
ADMI .058758 .015328 .076757 3.833 .0001
POP90 2.74897E-08 2.3100E-08 .026488 1.190 .2342
DENSITY 1.25323E-05 1.2378E-05 .018563 1.012 .3115
POPCH98 -.002239 3.5152E-04 -.113348 -6.369 .0000
INCOME 1.38595E-05 2.0664E-06 .142023 6.707 .0000
POVERTY 6.54358E-04 9.4040E-04 .014426 .696 .4866
SCOPE .066709 .002317 .572678 28.785 .0000
MDFUNC .004831 6.5375E-04 .150384 7.390 .0000
PRTAX .071292 .013705 .087204 5.202 .0000
DEBT .029209 .016561 .034063 1.764 .0780
NEWENG -.480063 .054873 -.163520 -8.749 .0000
NOEAST -.120426 .031921 -.074528 -3.773 .0002
NOCENT -.035030 .017281 -.051464 -2.027 .0428
SOUTH -.070051 .021151 -.103333 -3.312 .0009
AGE -.005927 .002001 -.060586 -2.962 .0031
HRULE -.024248 .030041 -.014520 -.807 .4197
UNIN .001494 2.7760E-04 .097724 5.380 .0000
ROWOFF -.013896 .002830 -.091753 -4.910 .0000
DISTRICT -.001629 3.4843E-04 -.111897 -4.674 .0000
(Constant)
Total Cases =
1.603911
1694
.069802 22.978 .0000
1 6 7
T O T A L T A X E S W I T H T W O R E F O R M I T E M S
Multiple R
R Square
Adjusted R Square
Standard Error
Analysis of Variance
Regression
Residual
F = 121.49312
.76891
.59123
.58636
.29546
DF
20
1680
Sum of Squares Mean Square
212.11245 10.60562
146.65395 .08729
Signif F = .0000
*** * m u l t i p l e REGRESSION ****
Equation Number 1 Dependent Variable.. LOGTAX
------------------Variables in the Equation —
log (base 10) of taxes
Variable B SEB Beta T SigT
EXECU .113301 .037982 .056651 2.983 .0029
ADM1 .278462 .023033 .263114 12.090 .0000
POP90 -4.02440E-09 3.4932E-08 -.002859 -.115 .9083
DENSITY 7.01353E-06 1.8685E-05 .007522 .375 .7074
POPCH98 -.005331 5.2927E-04 -.195594 -10.072 .0000
INCOME -3.28232E-06 3.1083E-06 -.024465 -1.056 .2911
POVERTY -.002868 .001414 -.045909 -2.028 .0427
SCOPE .041692 .003482 .257149 11.974 .0000
MDFUNC .010881 9.7936E-04 .245045 11.110 .0000
PRTAX .056507 .020611 .049897 2.742 .0062
DEBT .080534 .024923 .067354 3.231 .0013
NEWENG -1.030106 .086263 -.235729 -11.941 .0000
NOEAST -.294546 .048072 -.131670 -6.127 .0000
NOCENT -.028179 .025789 -.029980 -1.093 .2747
SOUTH -.173751 .031589 -.185280 -5.500 .0000
AGE -3.19491E-05 .003023 -2.356E-04 -.011 .9916
HRULE -.089742 .045238 -.038816 -1.984 .0474
UNIN .005405 4.1940E-04 .255389 12.887 .0000
ROWOFF -.044167 .004269 -.211205 -10.347 .0000
DISTRICT 3.86649E-04 4.9438E-04 .020252 .782 .4343
(Constant)
Total Cases =
1.281361
1701
.105017 12.201 .0000
1 6 8
U R B A N E X P E N D I T U R E S W I T H T W O R E F O R M I T E M S
Multiple R .82864
R Square .68665
Adjusted R Square .68260
Standard Error .37366
Analysis of Variance
Regression
Residual
F = 169.38918
DF Sum of Squares
20 473.01030
1546 215.85616
Signif F = .0000
Mean Square
23.65052
.13962
* * **MULTIPLE REGRESSION ***
Equation Number 1 Dependent Variable.. LOGUEX log (base 10) of uex
------------------Variables in the Equation------------------
Variable B SEB
EXECU .190076 .048745
ADMI .074740 .029685
POP90 -2.34121E-08 4.4654E-08
DENSITY 6.O6871E-05 2.3731E-05
POPCH98 .001336 6.6545E-04
INCOME 1.19922E-05 4.1630E-06
POVERTY -.003137 .001861
SCOPE .151149 .004855
MDFUNC .004900 .001286
PRTAX .055093 .027035
DEBT .251130 .033989
NEWENG -.202532 .130825
NOEAST -.032367 .061540
NOCENT -.086147 .033471
SOUTH .075781 .041231
AGE .013865 .003969
HRULE -.040969 .057556
UNIN .002424 5.5028E-04
ROWOFF -.037142 .005518
DISTRICT -.001889 6.3948E-04
(Constant) -.903554 .138484
Total Cases = 1567
Beta T SigT
.068075 3.899 .0001
.050116 2.518 .0119
-.011910 -.524 .6001
.046645 2.557 .0106
.035138 2.007 .0449
.061376 2.881 .0040
-.034724 -1.686 .0921
.608789 31.132 .0000
.077039 3.810 .0001
.033613 2.038 .0417
.143870 7.389 .0000
-.026628 -1.548 .1218
-.010421
-.526 .5990
-.062735 -2.574 .0102
.056248 1.838 .0663
.071270
3.493 .0005
-.012765 -.712 .4767
.080400 4.405 .0000
-.123790 -6.731 .0000
-.070205 -2.954 .0032
-6.525 .0000
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
Asset Metadata
Core Title
00001.tif
Tag
OAI-PMH Harvest
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC11256527
Unique identifier
UC11256527
Legacy Identifier
9617133