PUBLIC CHOICE AND PUBLIC POLICY: THE CASE OF
REFORMING THE FINANCING 0F PUBIJC EDUCMION
IN MICHIGAN

Disseflatim for the Degree of Ph. D.
MICHIGAN STATE UNIVERSI‘IY
WILLIAM A. SEDERBURG
1974

 

WW '1’
I

  
 

LIBRA.
M 9 My _"

{J‘z’rivem I

f”

This is to certify that the
thesis entitled
Public Choice And Public Policy: The Case Of
Reforming The Financing Of Public
Education In Michigan

presented by

William A. Sederburg

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Political Science

 

 

M74 2\

Major professor

Date 3/7/74

 

0-7639

 

 

02
20
P‘ 0 9

ISE

'9"

I A

j ’.
I) /.

(‘ ~' .

£1 ‘1

 

 

ABSTRACT

PUBLIC CHOICE AND PUBLIC POLICY: THE CASE OF
REFORMING THE FINANCING OF PUBLIC
EDUCATION IN MICHIGAN

By

William A. Sederburg

On November 7, I972, the Michigan electorate voted on a con-
stitutional amendment which, if passed, would have accomplished two
objectives. First, it would have greatly reduced property taxes. In
effect, its passage would have meant substantial ”property tax relief”'
for Michigan homeowners. Second, the amendment would have shifted the
primary responsibility for financing public education from the local
school district to the state. The amendment, known as Proposal C, was
defeated by a margin of 5A to A6 percent. The purpose of this disser-
tation is to analyze the financial and political circumstances leading
up to the vote on Proposal C and to compare two explanations for the
public's choice on school finance.

The two objectives of the proposal suggested to the author two
basic explanations for the electorate's vote. One explanation was
that the vote ought to be related to the property tax burden incurred
by residents in different school districts in the state. If voters
desired property tax relief, they should have voted in favor of Pro-

posal C. Three economic variables were used to determine whether or

a.

\
I?!

ah
I}
f

I". i
r‘-
=-

s

a
William A. Sederburg

not a relationship existed between the desire for property tax relief
and the vote on Proposal C. The three variables were tax burden on
income, tax burden on homeowners, and total tax effort.

The proposed shift in the level of government which would be
primarily responsible for financing public education led to a second
explanation. This explanation was that the vote was related to the
degree of ”social distance” among school districts in the seven metro-
politan areas in Michigan. Three social distance variables were used
to test this theory. They were segregation, life style, and social
rank. The social distance explanation was used in an attempt to ”tap”
the public's concern over preserving local control of education and
preventing cross district busing of school children.

The economic and social distance explanations were tested in
seven metropolitan areas in Michigan. Multiple regression analysis
was used to determine which set of variables “best explained” the
variance in the vote on Proposal C. In five of Huaseven areas, the
social distance variables explained more of the variance than did the
economic variables. In two areas, the economic variables best ex-
plained the variance.

The reason for the difference in the explanatory power of each
set of variables was determined by conducting a content analysis of
newspapers in each area for three months prior to the election. A
direct relationship was found between the power of the social distance
variables and newspaper coverage given to racial integration in the

schools.

William A. Sederburg

The importance of the social distance variables and the affect
of newspaper coverage about racial integration in the schools indicates
that voters were concerned with sociological as well as economic issues.
Social Scientists have recently been very concerned with the relation-
ship between the distribution of economic costs and benefits and public
policy. This research leads one to suggest that the reform of public
school finance in Michigan was as much affected by sociological factors
as by economic benefit. It points to the need to include both types

of explanations in analyzing public choice and public policy.

PUBLIC CHOICE AND PUBLIC POLICY: THE CASE OF
REFORMING THE FINANCING OF PUBLIC

EDUCATION IN MICHIGAN

By

*r

\‘0 " ‘-
William A. Sederburg

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Political Science

I974

ACKNOWLEDGMENTS

I would like to thank the members of my dissertation committee
for the time and effort expended in their thoughful and constructive
criticisms of the original draft. Kent J. Chabotar and Leroy Ferguson
contributed greatly to the development of the hypotheses and the presen-
tation of the findings.

A special thanks is given to the chairman of the committee,
Frank A. Pinner, for his personal and intellectual contribution to the
study. His theoretical concerns and practical solutions to difficult
methodological problems were extremely valuable in conducting the analy-
sis of the vote on Proposal C. With the rare combination of humanistic
concern and academic sophistication, his guidance has been well beyond
the call of duty. I am forever grateful.

I thank the Department of Political Science for a research
grant that enabled me to collect and analyze the needed data. The
entire department staff proved very helpful and constructive. I also
wish to thank my wife, Joyce, and my parents, for their constant under-
standing and support. This dissertation is dedicated to them.

All errors in judgment and analysis are my own.

/’\ TABLE OF CONTENTS

Page

LIST OF TABLES . . . . . . . . . . . . . . . . V
LIST OF FIGURES . . . . . . . . . . . . . . . viii

Chapter

I. INTRODUCTION . . . . . . . . . . . . . . I

II. TWO EXPLANATIONS OF PBULIC CHOICE AND PUBLIC POLICY . IA
The Economic Explanation . . . . . . . . . l5

Social Distance Explanation . . . . . . . . 26

III. RESEARCH DESIGN . . . . . . . . . . . . . A0
Hypotheses . . . . . . . . . . . . . . #0
Methodology . . . . . . . . . . . . . Ah

Unit of Analysis . . . . . . . . . . . . A6

The Universe . . . . . . . . . . . . . SI

Sources of Data . . . . . . . . . . . . SA

Dependent Variable . . . . . . . . . . . 56
Independent Variables . . . . . . . . . . 56

Summary . . . . . . . . . . . . . . . 6l

IV. FINANCING PUBLIC EDUCATION IN MICHIGAN . . . . . 6A
Rising Costs . . . . . . . . . 66
Exhaustion of the Property Tax . . . . . . . 68
Inequities in School Finance . . . . . . . . 75.

V. THE POLITICS AND HISTORY OF PROPOSAL C . . . . . 84
History . . . . . . . . . . . . . . . 86

Campaign . . . . . . . . . . . . . . l02

Election . . . . . . . . . . . . . . llO

Summary . . . . . . . . . . . . . . . II“

VI. ANALYSIS OF DATA . . . . . . . . . . . . 118

Over- All Picture . . . . . . . Il9
Comparison of Two Explanations . . . . . . . l25

Page

Metropolitan Differences . . . . . . . I35
Social Differences Within School Districts . . . IAB
Conclusions . . . . . . . . . . . . . lh7
VII. CONCLUSIONS . . . . . . . . . . . . . . ISI

APPENDICES . . . . . . . . . . . . . . . . . lS6

BIBLIOGRAPHY . . . . . . . . . . . . . . . . l59

Table

LIST OF TABLE

Revenue and Percentages by Governmental Level for
Public Elementary and Secondary Schools .

Trends in Total Expenditures for the Public Schools

l930-l970 .

Variations in Local Ability, Per Pupil, to Support
Education . . . . . . . .

Classification of Variables Employed in Factorial
Ecology

Varimax Rotated Factor Matrix after Rotation with
Kaiser Normalization for IA Variables in l83 School
Districts . . . . . . . .

Vote on Proposal C in Eleven Counties and the Re-
mainder of the State of Michigan

School District Classification for 183 Districts in
II Counties

Varimax Rotated Factor Matrix after Rotation with
Kaiser Normalization for IA Demographic Variables
for A9 Subcommunities of Detroit

Percent of Michigan Public School Systems Indicating
Methods Used for Curtailing Expesnse .

School Revenues by Source for All Public School
Districts in Michigan for Selected Years

Zero-order and Partial Correlations between Per
Pupil Expenditures and Selected Variables in l82
School Districts .

Financial Data for Selected School Districts, l97l-

I972

Page

A3

A5

52

53

6O

65

67

73

78

Table

4.5

SEV in Three Major Cities and Surrounding County

Districts l966/67 - l970/7l

Public Support for Proposal C in the Detroit Metro-
politan Area . .

Support of Proposal C and Attitudes Toward Busing
Reasons Given for Opposing Proposal C

Vote on Proposal C by Percent Rural Population in
Michigan's 83 Counties

Zero-order Correlations Between l3 Demographic Varia-
bles and the Percentage Favoring Adoption of Proposal
C in l83 School Districts

Zero-order Correlations Between Nine Demographic
Variables and the Percentage Favoring Adoption of
Proposal C in l82 Census Tracts in Six Central
Cities

Zero-order Correlations for Selected Variables from
A9 Subcommunities of Detroit

Partial Correlations for Selected Variables from A9
Subcommunities of Detroit with Percentage Nonwhite
Controlled for

Multiple Coefficients of Determination, Simple
Correlation Coefficients, and Beta Scores for l83
School Districts on Two Explanations of Public
Choice and Public Policy .

Zero-order Correlations Between Independent Variables
in Seven Regions and the Entire Universe

Multiple Coefficients of Determination, Corrected
Multiple Regression Coefficients of Determination,
and Multiple Partial Correlations of Determination
for Two Explanations of Public Choice and Public
Policy .

Multiple Coefficients of Determination, Zero-order
Correlations, and Beta Weights for three Social Dis-
tance Variables, Three Economic Variables and the
Two Most Powerful Economic Variables in Metropolitan
Areas where Multi-Collinearity Exceeds .79 .

vi

Page

8]

llO
lll

ll2

ll3

l20

l22

l23

12A

I27

l29

I3]

I33

Table

6.9

Multiple Partial Coefficients of Determination for

Two Explanations of Public Polity and Public Choice
in Seven Metropolitan Areas and the Relative Rank-

ing of Each Explanation .

Mean, Ranking of the Mean, Standard Deviation, Rank-
ing of the Standard Deviation, and Spearman Rank-
Order Correlations for the Mean Ranking on Six In-
dependent Variables

Mean Values and Rankings of Eight Wealth Variables
in Seven Metropolitan Areas

iRatios of the Importance of School Integration in

Seven Metropolitan Areas

Comparative Rankings of the Social Distance Explana-
tion and the School Integration Ratio

Multiple Regression Coefficients of Determination
and Beta Weights for Three Social Distance Varia-
bles in Census Tracts or Subcommunities in Seven
Metropolitan Cities

Percentage of Variance Explained by the Social

Distance Variables in Metropolitan Areas and
Central Cities .

vii

Page

l3A

I36

I38

IAO

IA3

IAS

IA6

LIST OF FIGURES

Figure

l. Changes in the Yields of Three State Taxes, I959-
l97I -

2. Changes in Property Tax Revenues for Three Purposes .

3. Percentage of Millage Elections Won and of Requested
Mills Passed, September, I968 to August, I972 .

viii

Page

69

7O

72

CHAPTER I

INTRODUCTION

Throughout the last half of the I960's and into the l970's,
‘many school districts throughout the United States were facing a
severe financial strain. During the l970-7l school year, for example,
the financial problems manifested themselves in a variety of ways.
In California 30 school districts went bankrupt; the number of teach-
ers throughout the state decreased by 9,000, while the number of stud-
ents increased by l00,000. Teachers were laid off in Cincinnati,
Chicago, New York, Los Angeles, Detroit, and in many other cities.
Libraries were closed in Cincinnati. In New York, teachers were paid
from the I972 budget.l

The financial problems of the public schools in the United
States stimulated a number of studies on the causes and solutions to
the problems of school finance. The National Educational Finance Com-
mission, Advisory Commission on Intergovernmental Relations, Urban In-
stitute, Michigan's Citizen Research Council, and other sources
published studies of the underlying crises in financing public educa-
tion.

The consensus of these studies was that the crisis is the re-
sult of the inadequacy and inequity built into the present method of

funding education. Currently, most school districts are funded from

a combination of local, state, and federal sources. Table l.l shows

the relative contribution of each source.

TABLE l.l.--Revenue and Percentages by Governmental Level for Public
Elementary and Secondary Schools (in Billions).

 

School Local Per- State Per- Federal Per- Per-

Total
Year Revenue cent Revenue cent Revenue cent cent

 

I96I-62 $l0.0 56.9 $ 6.8 38.7 $ .76 A.3 $l7.5 IOO
l965-66 l3.A 53.0 9.9 39.l l.9 7.9 25.A IOO
l970-7l 2l.8 52.0 l7.2 Al.l 2.9 6.9 Al.9 IOO

 

Source: Tables 29, 30 in Financial Status of the Public Schools,
Washington, D. C., Committee on Educational Finance, NEA, I970.

The backbone of the financial system is the revenue raised from
the local property tax. According to the National Educational Finance
Commission, 52 percent of all educational funds are derived from local
taxes. Of the amount, 98 percent is derived from the property tax.3
Consequently, increasing attention is being paid to the relative merits
or demerits of the property tax.

The crisis in public school finance takes the form of the
classic ”cost revenue squeeze,” in which educational costs far exceed
the capacity of the financial system to provide needed revenue.“ The
cost revenue squeeze can be easily demonstrated. Educational costs
have risen at an annual rate of 9.7 percent for the past decade (l96l-

l97l). This corresponds to an annual increase in the Gross National

Product of only 6.8 percent. Teacher salaries have increased by 78

5

percent, from $5,AA9 per year to $9,689 per year. Trends in total
expenditures for public schools are shown in Table l.2.

Although the costs have increased sharply (302 percent), the
means for obtaining the needed revenue have been strained to the point
where the public has refused to support increased revenues for public
education. The strain on the methods used for financing public educa-
tion is most evident in the often-discussed ”taypayers' revolt” against
increased millage for local districts.6 The method of allowing the
public to vote on educational millages and school bonds has opened the
public education system to the public's negative reaction to more taxes.
In I960 voters rejected ll percent of all school bond issues placed
before them. In I965 the number increased to 33 percent. In I970 the
percentage of bond election failures was 52 percent.7 Many states,
such as Michigan, also require local school districts to obtain voter
approval for basic operating millage, thereby greatly compounding the
problem of inadequate funds.

The causes of the taxpayers' revolt are typically attributed
to the nature of the property tax, rather than a reaction against public
education. An advisory Commission on Intergovernmental Relations survey
found that the property tax is two and one-half times more unpopular
than the federal income tax and three and one-half times more unpopular
than the sales tax.8

Burkehead, Johns, Rossmiller, and others have pointed to defects
in the property tax that have had a particulary important impact on

the public attitude toward it.9 One of the defects is the method of

._mm_ .mc_co_u .m___>mc_mu .umeOcd oucmc_m .mco_umuaum _mco_umz Hooc30m

 

 

_m Nam mm com :__ ~_o._ oe-mem.

on om-m~m_

ommocuc_

ucoocud

~.: 0mm mm_.~: mws.mm N.~N_ mm:.mm ou-mmm.

~.m moo ems.~m .so.m_ m._o_ m_e.m_ oe-mmm.

~.~ me sm~.- Nfim.w c.mm mme.m omIm:m_

m.~ mew mso.- ms_.m s.ms .mm.~ osImmm_

~.~ MNNW mo_._~ mmm.¢ m ~.wm mo~.~ w om-m~m_

m_oo;um u__n:a mtm__oo Am.ooo_v mtm__oo xmec_

co acoam mmm. <o< oucmtcmuu< mom. c. ou.cm muc:u_ccmaxN cmo>

mzo mo c_ __a:m cum >__mo moc:u_ncuaxW cesamDOU .muOF

Havocom moc:u_ccuaxm ommLo><

 

.Ammocvuuaonc_ _oo;um co “mocouc_ new >m_u:o .mu_amu .umcuaxN unoccau cow
meou_ __m moon—uc_v cum—Iomm_ .m_oo;om u__n:d ego to» moc:u_vcoaxm .muOh c_ mucochuu.~._ o_m<h

administering the tax. The Advisory Commission on Intergovernmental
Relations observed that IImuch more action toward improved administra-
tion of the property tax is still urgently needed."10 Their indictment
is based on allegedly inequitable assessing practices within and among
localities and the regressive nature of the prOperty tax.

The regressive nature of the prOperty tax is shown in the in-
creased percentage of income the poorer property owner must pay, com-
pared to the wealthier property owner. For example, in I970, a home-
owner with an income of $l0,000 paid about $3AO or 3.A percent of his
income for property taxes. Homeowners with an income of less than
$7,000 paid an average tax of 6.6 percent. Elderly homeowners paid,
on the average, 8.l percent of their annual income.11 Coupled with the
problem of regressivity is the fact that the prOperty tax has risen
much faster than average income. In I953, a family with an income of
$5,000 paid $ll0 or 2.2 percent in property taxes. The same family
in l97l earned $l2,000 and paid a little over SAOO, or 3.A percent of
their annual income in property taxes.

A'Uwhxjfeature of the property tax is the uneven distribution
of property value among school districts. Property value varies ac-
cording to the value of homes in the district and the degree of in-
dustrialization. School districts with either residential property
with high assessed valuations or large industrial plants within their
boundaries tend to have a strong tax base on which to draw funds. The
impact of such a variation is that two districts, each levying the

same amount in taxes, will not raise the same amount of money. The

unequal tax base among school districts has created wide differences
in tax rates and in the economic burden carried by districts with low
property value. The unequal tax base from one district to another, and

from one state to another, is indicated in Table l.3.

TABLE I.3.--Variations in Local Ability, Per Pupil, to Support Educa-

 

 

 

tion.

Property Value . .

State Unit Analyzed Year Per Pupil RatIo Of HIgh
. to Low

Low HIgh
Mass. City, Town 65-66 $ 5,000 $3,335,398 66
Kentucky School Dist. 6A-65 A,868 9A,l29 l9
Colorado County 63-6A 8:339, A8,672 ll
Oregon County 62-63 l7,lOA 73.l0A A
Maryland County 6A-65 7,7A2 20,06A 3

 

Source: Joel Berke, “The Current Crisis in School Finance: Inadequacy
and Inequity,“ in Phi Delta Kappan, September, 7l.

 

A final weakness of the property tax is the fact that revenues
are not related to educational need. Many districts, wealthy in pro-
perty value, do not have the great educational problems faced by the
inner city school district. Conversely, in the ”poorer“ districts,
where educational need is often the greatest, the revenue sources are
commonly the weakest.

Coupled with the increasing costs and the strain on the present
method of raising revenue is the need to achieve equity in school fi-

nancing. Already noted were the uneven distribution of property tax

base between districts and the lack of a relationship between educa-
tional need and property wealth. These disparities have caused tre-
mendous differences in the amount of money spent per pupil each school
year. For example, in Michigan the dollars spent for each pupil
varied between $580 and $l,6OO in l97l.l3

Variations in the amount of money spent by school districts
for each child have been the subject of recent court cases.'h In cases
brought before the Supreme Courts of Michigan, California, Texas,
Minnesota, and New Jersey, the plaintiffs have argued that the acci-
dents of a child's birthplace and of his parents' residence should not
be cause for denying him equality of educational opportunity, as
measured by dollars Spent.

The California Supreme Court, in Serrano v. Priest (l97l),
found the method of funding public education in California unconstitu-
tional. In their ruling, the court stated:

This system conditions the full entitlement to such in-

terest [quality education] on wealth, classified its recipients

on the basis of their collective affluence and makes the qual-

ity of a child's education depend upon the resources of his

school district and ultimately on the pocketbook of his par-

ents.‘5
In ruling California's system of financing education unconstitutional,
the court established the criterion of equal educational opportunity
based on equal dollars and equal tax base. Other state supreme courts
have followed California in declaring the present method of funding
public education unconstitutional. Recently, the United States Su-

preme Court, in Rodriguez v. State of Texas, I973, upheld the current

system of financing public schools.

School finance and the property tax are inextricably linked. Ap-
proximately one-half of all funds spent are derived from the local prop-
erty tax, yet this tax has been the primary cause of the “taxpayers'
revolt,” inequities in expenditures per student, and the inability of
school districts to raise additional revenue. The weaknesses and in-
equities in the property tax, outlined above, have helped create
political pressure to reform public school finance.

The property tax has become the primary target of political
leaders seeking to reform school finance. The obvious solution to the
financial crisis is to provide homeowners with property tax relief,
public schools with adequate funds, and students with equal educational
opportunity. This could be accomplished by shifting from reliance on
the local property tax to reliance on state taxes. However, to alter
the method of financing public schools, state constitutions would have
to be amended to restrict local funding. In many states the amendment
process requires approval of the state electorate.

On November 7, I972, the ballots of four states--Oregon, Michi-
gan, California, and Colorado--included proposed constitutional amend-
ments that would have restricted the use of local prOperty taxes and
shifted the primary source of education funds to the state. The voters
in all four states rejected the amendments. In Michigan, the amendment
was defeated by a margin of l,32A,702 in favor to l,8l5,l26 against
(A6-5A percent). The results were similar in the other three states.

The primary purpose of the present study is to determine why
the constitutional amendment was defeated in Michigan. What caused the

voters to reject the change in the way schools are financed? Two models

or explanations of public choice and public policy are proposed. The
two explanations are compared by using multiple regression analysis.
The percentage of variance explained should offer clues about the ex-
planatory power of each model. The status of school finance in Michi-
gan prior to the election and the political history of the amendment
are also discussed. Through this case study approach, the underlying
dimensions of the vote should be apparent.

The two explanations of public choice and public policy to be
explored are the economic and social distance models. The economic
interpretation hypothesizes that the economic benefits of property
tax relief would best explain the vote. It is assumed that voters
would want to shift from the highly disliked property tax to one
that would be less disliked. Three independent variables make up the
economic explanation: tax burden on homeowners, tax burden on income,
and the total tax effort of the district. All three variables are
measured by school district.

The social distance explanation assumes that voters wanted to
keep the property tax because of a desire to maintain the role of the
local school district in financing education. Approval of the amend-
ment would have removed the opportunity for voters to vote on educa-
tional millages. The model hypothesizes that the amendment was in-
terpreted as an integrative mechanism: the shifting from local to
state funding would have meant the social and political consolidation
of the state school system. If voters interpreted the amendment as a

threat to local political and social boundaries, they may have been

l0

willing to accept the burden of the property tax in order to maintain
social boundaries.

The concept of social distance is used to test this hypothesis.
Social distance is defined as ”life style differences between aggre-
gates of people living in proximity."'6 Life style refers to general
ways of living. Shevky and Bell used three measures-~segregation, ur-
banization, and social rank--to differentiate social areas according
to life styles. Segregation is the degree of racial separation. Ur-
banization is a measure of the urban, as compared to suburban, mode
of life: low familism. Social rank is a measure of social class:
education and occupation. Social distance between areas is computed
by taking the difference between the suburban and inner-city district
on each measure.

The two models of public choice and public Policy are compared
in the seven Standard Metropolitan Statistical Areas of Michigan. The
analysis is limited to the seven metropolitan areas so that the concept
of social distance can be operationalized. Chapter III, Research De-
sign, explains this choice in greater detail. The seven areas incor-
porate over 70 percent of the total vote on the constitutional amend-
ment. An explanation of the vote in these metropolitan areas provides
good clues about the entire vote.

This study is organized into seven chapters. In Chapter II
the two models of public choice and public policy are presented. Chap-
ter III contains a description of the research design used to analyze

the vote. The financial situation facing public education in Michigan

ll

prior to the constitutional amendment is reviewed in Chapter IV. In
Chapter V the political history of the amendment is discussed. Chap-
ter VI contains the findings and analysis of the data. The final chap-

ter is a summary and conclusion chapter.

FOOTNOTES--CHAPTER ONE

IJoel Berke, ”The Current Crises in School Finance: Inadequacy
and Inequity,” Phi Delta Kappan, September, I97], p. 3.

 

2Recent studies on educational finance include eleven smaller
studies and six volumes (Gainsville: National Educational Finance
Project, l97l), The Advisory Commission on Intergovernmental Relations
has published Financing Schools and Property Tax Relief - A State Re-
sponsibility (Washington: Advisory Commission on Intergovernmental
Relations, I973), the Urban Coalition has published Public School
Finance: Who Pays and Who Benefits (Washington: Urban Coalition, I972),
and in Michigan the Michigan Citizens Research Council has published a
series of background papers on educational finance.

 

 

 

 

3Committee on Educational Finance, National Education Associa-
tion (Washington: NEA, I970), Table 30.

“Berke, op. cit., p. 2.
Slbid.

6Karen DeW. Lewis, ”Educational Report/Federal Policy Toward
School Finance Faces Review by Administration, Courts,” National Journal,

 

December, I972, p. I95A.

7Ibid., p. I955.

Advisory Commission on Intergovernmental Relations, ”Financ-
ing Schools and Property Tax Relied‘A State Responsibility,“ Report
A-AO, January, l973, p. 26.

9Richard Rossmiller, Fiscal Capacity and Educational Finance,
(Gainsville: National Educational Finance Project, Special Study
No. IO, l97l), Roe I. Johns and E. L. Morphet, Problems and Issues in
Public School Finance (New York: Teachers College, Columbia University,
I952), and Jesse Burkehead, Public School Finance: Economics and
Politics (Syracuse: Syracuse: Syracuse University Press, I963TT

loAdvisory Commission on Intergovernmental Relations, op. cit.
p. 69.

 

’

l'Lewis, op. cit., p. I958.

I3

Ibid.

'3Michigan Department of Education, Bulletin lOl2, l97l-72,
Lansing, Michigan, I972.

'hFor an excellent discussion of the impact of the court cases
see Joel Berke and John Callahan, “Serrano v. Priest: Milestone or
Millstone for School Finance,” Journal of Public Law, Vol. 2I, No. l,

1972: pp° 23-7]-

lSCalifornia Supreme Court, Serrano v. Priest (5 Cal. 3d 58A,
A87 P 2d IZAI, 96 Cal. Rptr. 6OI, I97I), p. SA.

l6Brett W. Hawkins, ”Life Style, Demographic Distance, and
Voter Support of City-County Consolidation,” Southwestern Social Science
Quarterly, Vol. 58, No. 2, September, I972, p. 325.

 

CHAPTER II

TWO EXPLANATIONS OF PUBLIC CHOICE AND PUBLIC POLICY

On November 7, I972, the voters of Michigan had an opportunity
to vote on a constitutional amendment which would have accomplished two
purposes. First, the amendment would have greatly reduced reliance on
the property tax as the primary source of educational revenue. This
change would have provided many citizens throughout the state with sub-
stantial ”property tax relief.” Secondly, the amendment would have
shifted the primary responsibility for funding public education from
the local school district to the state. This change would have meant
a fundamental restructuring of the process through which decisions con-
cerning the financing of public education are made.

The vote on the amendment, better known as Proposal C, was ob-
viously a policy-making decision. Citizens, both individually and
collectively, had an opportunity to accept the changes outlined in the
amendment or to maintain the status quo. Consequently, the vote on
Proposal C can be, and is, approached theoretically in terms of the
relationship between public choice and public policy.

In recent years, the literature on the relationship between
public choice and public Policy primarily has taken two directions.

One direction can be termed “economic,” in that it emphasizes the

economic conditions of the unit of analysis and the economic costs or

IA

l5

benefits accrued to the public if a certain policy is adopted. The
second direction stresses the importance of the structural character-
istics of the policy process and the political environment. Research
on metropolitan consolidation and on public referenda suggests that
the maintenance of social, economic, and political differences is
important in determining how the public will vote on different policy

alternatives.

The Economic Explanation

 

The economic explanation or model of public choice and public
policy has two facets. One method of analyzing the impact of economics
on policy is to study the economic conditions of the unit of analysis
and correlate environmental variables with different policies. Gener-
ally, this method has concentrated on the relationship between the
economic condition of American states and the level of expenditure in
different policy areas. The second method of studying the relationship
between economics and policy is to analyze the distribution of costs
and benefits caused by different policies.

The literature on the systematic analysis of the effect of econ-
omic conditions on public policy dates back to at least I952, when Fab-

ricant published The Trend of Government Activity in the United States.

 

He found that more than 70 percent of the variance in state and local
expenditures could be attributed to differences in per capita income,
urbanization, and population density. On education policy, Fabricant
identified per capita income as the most important variable in deter-

mining expenditures per pupil.

l6

Subsequent research has supported Fabricant's identification
of the importance of the economic condition of the state in determin-
ing expenditure levels. Fisher, Shapiro, and Hirsh also concluded from
their research that expenditure levels for education were determined
largely by the economic wealth of the state or locality.2

Dawson and Robinson expanded the Fabricant, Fisher, Shapiro
research design to include political dimensions of policy outcomes.
In exploring Key's hypothesis that interparty competition and welfare
policy were related, Dawson and Robinson found that party competition
had little effect on the ”welfare orientation” of the state when per
capita income was controlled for.3 Hofferbert redefined ”welfare orien-
tation” and added other political variables including party control,
malapportionment, and regional controls. Hofferbert's conclusion was
that political variables had no independent impact on ”welfare orienta-
tion.”h

Politics, Economics, and the Public: Policy Outcomes in the

 

American States by Dye was the appex of this line of research relating

 

socio-economic characteristics of political units to expenditure levels
in different policy areas.5 Dye expanded the previous research by in-
cluding additional socio-economic and political variables, and found
that only economic development variables had an independent effect on
educational policy. Educational policy was defined primarily as the
amount of money spent on education in the state. Economic deVelopment
consisted of education, income, urbanization, and industrialization;

of the four, per capita income was of greatest influence in affecting

l7

educational policy.6 Dye found that political variables (constitu-
tional framework, electoral system, party system, interest group
structure, power structure, and political style) were not significantly
correlated with educational policy. After analyzing five different
policy areas, he concluded:

Political system characteristics have relatively little

effect on political outcomes of the states. Economic develop-
ment shapes both the political system and the political out-
comes, and most of the association between system character-
istics and policy outcomes can be attributed to the influence
of economic development.7

Dye, Sharkansky, Hofferbert, and other policy analysts have been
criticized on two grounds.8 One criticism is that the dependent varia-
ble, public Policy, has been defined primarily as an expenditure level.
It seems almost axiomatic to say that states with more moeny will
spend more. The second criticism is that Dye and others have not
offered an adequate explanation of how economic development gets trans-
lated into higher expenditures or into other types of policy.

Students of public policy have attempted to meet the first
criticism by conceptualizing different dimensions of public policy.9
For example, Sharkansky and Hofferbert distinguished between public
policy and policy outputs.l0 Public policy was defined as expendi-
tures or other indicators of official concern. Policy outputs were
defined as the effect of the policy. For education, the policy was
the amount of money spent in the state for education. Educational
policy outputs included the percentage of ninth grade students who

went on to complete high school and the percentage of candidates

passing the minimum academic standards for the selective service.

l8

Sharkansky and Hofferbert concluded that socio-economic variables were
more highly correlated with public policy (expenditures) than with
policy output (impact). Policy output was most highly correlated with
policy.

Fry and Winters introduced a new dimension to public policy
analysis that may prove fruitful. Their study was concerned with ans-
wering two basic questions: (I) Does politics make a difference in
public policy outcomes? and (2) How are certain factors, socio-economic
and political variables, specifically related to policy outcomes?"
They contended that the allocation of benefits and burdens of state
revenue and expenditure policy was a more fruitful area of analysis
than studying only the level of expenditure.

We have taken as our dependent variable the net redistribu-

rtive impact of revenues and expenditures as represented by the
ratio (fl: expenditure benefits to revenue burdens for the three
lowest income classes in each state. The major hypothesis of
our study is that, in regard to the allocation of the burdens
and benefits of state government revenues and expenditures,
political variables will have a stronger influence on policy
outcomes than will socio-economic variables.

The authors operationalized the redistributive impact by select-
ing a number of allocation bases that appeared to describe the incidence
of revenue burdens and expenditure benefits across income classes.

For instance, elementary and secondary education expendi-

ture benefits were assumed to be distributed according to the
number of children under 18 in families in each income class,
so that if 20% of children under l8 were in families with an
income of $A,OOO to $A,999, 20% of expenditures on elementary
and secondary education were assigned as benefits to that
income class.'
They then multiplied this percentage by the total amount of money spent

in that category, i.e., elementary and secondary education. The totals

were summed for all categories for both expenditures and revenues. A

l9

redistributive ratio Vfifii computed by dividing the expenditure benefits
for each income class by the revenues paid by that class.

Using multiple regression analysis, Fry and Winters computed
the percentage of the variance, in redistributive impact between states,
explained by both political and socio-economic variables. Multiple
partial correlations were computed to determine which set of variables
best explained the redistributive ratio. They found that political
characteristics explained more of the variance than did socio-economic
variables. Political variables included five mass political behavior
variables, four governmental institution variables, and four elite be-
havior variables. Socio-economic variables included Dye's economic
development variables (education, urbanization, industrialization, and
income), a Gini index of equality, and the percentage of families earn-
ing less than $3,000 per year. They concluded:

...Previous studies of policy outcomes in the states have

been hard pressed to find an independent impact for the political
variables considered, and where the relative impact of political
and socio-economic variables has been examined, the socio-
economic variables have predominated. In the present analysis,
these finding are reversed.Hi

Sullivan attacked Fry and Winter's findings on methodological

I5

grounds. He pointed out that in order to use multiple regression
analysis to compare different models each model must have the same
number of independent variables. Although Fry and Winters used five
variables for each group of variables, they selected the most power-
ful political variables from a set of nine, while selecting five socio-

economic variables from a set of six. Sullivan recomputed the multiple

regression coefficients and the multiple partial correlations. He

20

found that Fry and Winter's findings were reversed when proper statis-
tical procedures were used. Sullivan concluded that although politics
may make a difference in the amount of redistribution, it by no means

overshadows the role of socio-economic factors.

Booms and Haldorson expanded Fry and Winter's search by refin-
ing the dependent variable.'6 Using six redistributive equations, Booms
and Haldorson changed the rankings of many of the states on the redis-
tributive ratio. This method increased the percentage of variance ex-
plained by the socio-economic variables. Their study still found that
political variables explained more of the variance than did socio-
economic variables. However, they did not restrict the number of inde-
pendent variables in each group of variables. This opens their analy-
sis to the same criticisms made of Fry and Winter's research.

Research on the impact of socio-economic characteristics of the
state or locality under study has pointed to the importance of economic
conditions in affecting public policy. Policy has been defined as the
level and distribution of public resources. In both cases, the economic
environment has had a direct effect on public policy.

If Proposal C had passed, the economic effect would have been
a reduction of property tax burdens. Supporters of Proposal C attempted
to sell it on grounds that the proposal offered substantial ”property
tax relief.” If, as policy analysts have shown, the economic condition
of the environment is related to policy outcomes, it is conceivable
that economic variables measuring the extent of property tax relief
offered by Proposal C ought to be correlated with the vote. Theoreti-

cally, in school districts with a heavy property tax burden, support

2]

for Proposal C should have been greater than in districts with a less
heavy tax burden.

The second criticism raised against policy analysts is that
they have not provided a clear explanation of how environmental condi-
tions are translated into public Policy. Dye referred to the economic
development literature in explaining how higher governmental expendi-
tures come about.'7 As governments become more urbanized and tech-
nologically developed, increased demands for governmental survices are
created. The government is responsible for providing new industrial
services as well as meeting the social problems created by urbanization.
The new responsibilities lead to increased expenditures and increased
revenues. This description explains how higher levels of expenditures
may come about. It does not explain other policy preferences such as
the redistribution of public resources.

If we accept the conclusion that socio-economic conditions are
positively associated with expenditure levels, redistribution, or other
policies, the question of how the environment is linked with different
policy choices becomes important.

Most policy analysts have explained the linkage in terms of
systems theory, which creates a model of public Policy in which the
socio-economic environment creates inputs for the political system. The
political system translates these inputs into public policy. The ef-
fects of the public policy change the environment; this, in turn,
creates new inputs. Inputs are defined as demands and supports.

Demands ”occur when individuals or groups, in response to perceived

22

environmental conditions, act to promote goals, interests, or actions.
Support is rendered when individuals or groups accept the outcome of
elections, obey the laws and pay their taxes.”'

In this conceptualization, the socio-economic environment creates
the political demands for policy change. Dye's environmental measures
of economic development (urbanization, industrialization, education,
and income) are not demands in themselves, but are representative of
demands. Demands are individual or group responses, not economic con-
ditions. In this analysis, the economic conditions of the school dis-
trict are seen to create political preferences among individuals and
groups. These preferences are indicated by support or opposition to
the proposed policy change. It is hypothesized that the extent of prop-
erty tax relief offered by Proposal C in the local district should be
positively related to the vote on Proposal C. Support for Proposal C
should be greater in districts with high property tax burden than in
districts with low tax burden.

If this hypothesis is plausible, it is important to ask how
individuals select among different policy choices. This question has
been studied by political economists who have looked at the distribu-
tion of costs and benefits caused by different policies and how this
distribution influences policy preferences.

The conceptualization of economic distribution as being at the
heart of the political process is the core of the political economy
approach to public policy. Jacob and Lipsky, in a review of the litera-

ture on policy analysis, identified this core:

23

The distribution of benefits and sanctions is perhaps
the most significant output dimension for political scientists,
since most of the conflict preceding the adoption of a pro-
gram is not about whether it should be embarked upon but who
will pay and who will benefit.'9
The study of policy in terms of distribution of resources has
generated considerable research.20 In education there has been inter-
est in the distribution of costs and benefits in the method of financing
public education. The claim that ”equal educational opportunity” is
dependent on local wealth and social status has encouraged courts to
demand that expenditures per pupil be based on the wealth of the entire
state rather than the local school district. In the most well-known
court decision, Serrano v. Priest, the California Supreme Court ruled
that the current method of funding public education is inherently un-
equal and therefore unconstitutional. Indeed, the current controversy
in financing public education is based primarily on the analysis of
public policy in terms of the distribution of costs and benefits.
Various authors have used the concept of distribution of costs
and benefits in developing economic models of how policy decisions are
made.2' For example, Ross identified the main components of the politi-
cal economy model of public choice and public policy as consisting of
the analysis of costs and benefits; the concepts of markets, exchange,
and bargaining; and the use of strategies to gain relative advantage.22
Wade and Curry offered a seven-element model of the policy process.
Their model consisted primarily of the government acting as the mechan-
ism for distributing or redistributing economic resources.23
Although these models are useful, Mitchell provided probably

the best summary of how to decide which policy to accept. He summarized

the political economy model as follows:

2A

Underlying the above listed questions is the fundamental
conception of politics as being essentially an exchange phenom-
enon, not totally different from economic exchange. In this view
of politics, the economists are inclined to emphasize rational
choice on the part of individuals and organizations as they engage
in various types of exchange among themselves and with political
parties and governments in pursuit of their subjective self in-
terest.... Governments want support, compliance, and resources,
while the individual citizens want to improve their shares of the
benefits and/or reduce costs.2

Mitchell's elaboration of the political economy model of policy choice
implied that choice is based on a ”balance shed,” by which the citizen
calculates the economic advantage or disadvantage of the policy being
considered. It seems logical that the political economy model ascribes
to Down's statement that “rational men are not interested in policies
. . . . . ”25
per se but In theIr own utIlIty Income.
In essence, the economic explanation used in this research is
a synthesis of the two facets of policy research. It is accepted that
environmental conditions affect policy preferences. However, rather
than defining environment as gross economic development, we define en-
vironment as the distributive impact of current school finance policy.
According to systems theory, the output of current school finance policy
feeds back into the environment, creating demands from individuals and
groups. Current school finance policy has created differences in taxa-
tion burden between school districts. In turn, these differences gen-
erate political demands for shifting to another means of funding public
education. Conceivably, these demands and policy preferences are
based on a cost/benefit analysis of the prOposed policy.

Fry and Winters and Booms and Haldoson used the distributive

impact of state policy as their dependent variable. We suggest here

25

that the distributive impact of current policy can also be used as a
measure of the economic environment. The dependent variable becomes
the percentage of peOple favoring the adoption of a change in public
policy.

Thus, we hypothesize that Proposal C will be supported in school
districts that would receive substantial ”property tax relief” with the
passage of the proposal. Three variables of economic benefit ought to
be related to the vote. First, some districts tax themselves at a much
higher rate than other districts. However, the tax rate, in itself,
does not adequately reflect financial effort since districts vary as
to tax base. Consequently, a better measure of tax effort is to divide
the tax rate for education by the state equalized valuation for each
school district. This is the first economic variable. The second
variable measures the tax burden on homeowners in the district. Dis-
tricts vary greatly according to the percentage of property tax reven-
ues raised from business property, thereby putting less strain on resi-
dential property owners. The variable used to measure this tax effort
is computed by dividing the total number of dollars raised from resi-
dentail property in the district by the number of homeowners. A final
economic variable is tax effort in comparison to total income. People
paying a smaller percentage of their income in prOperty taxes would be
less likely to support a change in the method of financing education
than individuals paying a larger percentage of their income in property
taxes. This variable was measured by dividing the money raised from

residential property taxes by the total family income in the district.

26

These three variables provide a good indication of the distributive

impact of current policy and whether or not economic benefit was corre-

lated with the vote on Proposal C.

Social Distance Explanation

A second approach to public choice and public policy is to study
the relationship between policy preferences and the social environment.
One of the key questions to answer when discussing the policy prefer-
ences of different electorates is whether individual or household char-
acteristics are sufficient for the prediction of behavior, or whether
it is also essential to study the social environment as well. Bell,
in reviewing the work on this question, concluded that there was ”con-
vincing evidence that the social character of local areas within a city
as defined by economic, family, and ethnic characteristics is an im-
portant factor in predicting individual attitudes and behavior, sub-
cultural patterns and social organization.“26 The characteristics of
the social area are seen to shape both attitudes and social interaction.
These attitudes and interactions are reflected in the public's support
of or opposition to a change of policy.

Passage of Proposal C would have shifted the primary burden of
financing public schools from the local school district to the state.
In effect, it would have financially integrated the public school sys-
tem in Michigan. Many opponents to Proposal C argued that its passage
would have meant the loss of local control of public education. If
this argument had an impact, the vote on Proposal C might best be ex-

plained as a vote against social and economic integration.

27

Social scientists have identified socio-economic differences
within metropolitan areas as important determinants of public policy.
The pioneer work done in identifying socio-economic differences was
Shevky and Bell's The Social Areas of Los Angeles.27 The authors iden-
tified three measures useful for clarifying social areas within a
metropolitan region: social rank, urbanization, and segregation.
Social rank is an index made up of education and occupation. Urbani-
zation is an index measuring life style or family status characteris-
tics. It consists of the proportion of women in the labor force,
fertility ratio, and the percentage of single family dwellings.
Segregation is the proportion of minority ethnic groups in the census
tract. Although developed primarily for methodological purposes,
Shevky and Bell suggested their measures of urban differentiation could
be related to policy preferences.

The early uses of Shevky and Bell's social measures correlated
policy preferences with the three indices within a single metropolitan
community. Greer and Kaufman used the three indices to explain the
vote for Stevenson in the I952 Presidential election and support for
a metropolitan sewage district. As their universe they used 2I6 cen-
sus tracts in the St. Louis metropolitan area. Generally, they con-
cluded that ”the population types as described by the Shevky-Bell
typology produce significant differentiation in voting behavior."28

One conclusion is of particular interest for this research.
Greer and Kaufman discovered that the percentage of people supporting

a metropolitan sewage district was very low in familistic areas (low

28

on the urbanization index). This was in direct conflict with the voters'
economic self-interest.
Were one to assume a strictly rational model of
voting, in which the material interests of the voters
would be clearly calculated and made manifest in choice
patterns, the familistic neighborhoods of low urbanization
should have voted for the district.
The explanation was that familistic areas were interested in maintain-
ing local control over sewage diSposal and the corresponding taxation
policy. The familistic census tracts tended to be loCated in the sub-
urban fringe surrounding the city of St. Louis. Consequently, Greer
and Kaufman concluded that the vote on the sewage district was inter-
preted as an integrative decision in which the suburban areas, with high
familistic scores, wanted to maintain their individual role in sewage
control.
Local control was also identified as a major influence on policy

outputs in the Philadelphia metropolitan area.30

Dye, Herman, Williams,
and Liebman used urban differentiation to explain policy differences
among governmental units. They combined the social rank index of
Shevky and Bell with measures of the size, age of the population,
wealth, religion, race, property value, property use, and political
party preference. Unfortunately, they did not empirically demonstrate
that these measures sufficiently differentiated urban areas.

The major conclusion reached by Dye et_313 was that urban
differentiation and the desire to maintain differences shaped both
the politics and the policies of governments in the metropolitan area.

Dye, et. al. concluded that suburban school districts and municipal-'

ities were engaged in the politics of protecting and preserving

29

suburban values. “There is no doubt that local control does have mean-
ing for suburban municipalities.”3'

Dye foresaw future conflict between efforts to integrate public
services and the desire to maintain local control. He hypothesized that
when suburban areas perceive a policy to be a threat to ”local control,”
there will be little support for the policy, regardless of other conse-
quences.

The identification of local control as a major factor in policy
choices in metropolitan areas leads to the hypothesis that issues
viewed as integrative in nature will be opposed by suburban units of
government. Students of international politics have suggested that
viable political integration is a function of cultural or social dis-
tance. As Karl Deutsch pointed out, culturally similar societies are
more likely to share enough values and to enjoy sufficient facility of
communication to provide the necessary policy consensus for political

32

integration. Dissimilar societies are less likely to achieve the
consensus necessary for such integration. The related hypothesis in
this research is that social distance, as measured by differences in
social rank, urbanization, and segregation is related to the vote on
Proposal C.

A number of studies have used the concept of social or “demo-
graphic” distance as the major independent variable to explain policy
preferences;33 Dye §£_al. hypothesized that cooperation between differ-
3A

ent political units was an effect of social and economic distance.

The authors measured cooperation as the existence of an agreement_

30

between municipalities or school districts in the Philadelphia, Penn-
sylvania; area jointly to finance a service facility. Sewers, schools,
and police radio networks were the only policy areas in which a substan-
tial number of municipalities had a joint agreement. Distance was com-
puted as the difference between contiguous Infits on measures of social
rank (as measured by the procedures of a Shevky and Bell), market value
of property per capita, and party identification. Using this method,
53A pairs of contiguous municipalities or districts were subdivided
according to whether or not they had a joint agreement. The coopera-
tive and noncooperative pairs were then measured on social and economic
distance.

Dye §£_al, found that cooperative agreements between school dis-
tricts were more common among districts with similar social rank. When
they controlled for urbanization, they found a rural district was more
likely to have an agreement with another rural district than with an
adjacent urban district. This was explained by the ”lesser social
distance between rural districts,“ than between rural and urban dis-

35

tricts. Differences in social rank were also related to coopera-
tive agreements for sewage disposal and radio networks.

Along the same line of research, Dye found that differences in
education policy between central cities and their suburbs were related

36

to differences in social status and life style. Education policy was
defined as total educational expenditures in the school district and
the municipal and school tax rate. Life style and social status were

composed of variables similar to those included in Shevky and Bell's

3l

indices. Life style consisted of a fertility ratio. percentage of
women in the labor force, percentage of single family dwellings, and
percentage of lA-l7 year olds in school. Social status was composed
of median years of education for adults, percentage of population hav-
ing graduated from high school, percentage in white collar jobs, and
income. Dye did not combine these variables to make two indices. In-
stead, he subtracted the value for each variable for the central city
from the average value for each variable for the suburban school dis-
tricts. His study used five metropolitan areas of Wisconsin.. Thus,
he was able to determine whether life style and social status scores
were higher for the city or the suburbs in five metropolitan areas.
Dye concluded that expenditure levels for education are re-
lated to the direction of social distance.
In larger urban areas, where social and life style
differences between city and suburb are more pronounced,
public expenditures reflect these differences; but in
smaller urbanized areas, where social and life style differences
are slight, differences between city and suburb in educational
expenditures are slight or tend to run opposite to the expected
direction.3
Suburbanites tax themselves more for education than for other munici-
pal services. Dye attributed this to the fact that education is the
means whereby suburbanites maintain their social status relative to
the central city.
Wood supported the same conclusions reached by Dye: Education
and suburban life style are closely associated; and maintaining the

role of the locality is an important concern for suburbanites. In his

study, Wood theorized that suburbia is a result of the ideology of

32

faith in a community of limited size and a belief in the condition of

38

intimacy. To Wood, suburbia has become the heroic defender of de-
mocracy because of its faith in representation and smallness. Central

to this faith is the fiscal independence of the suburban municipal
government and the local school district.

The independence of both municipal governments and school dis-
tricts is based on an independent tax base--traditionally, the local
property tax. As Wood viewed it, any attack on the property tax will
be interpreted by suburbanites as an attack on the suburban life style.39
If Wood's theories are correct, Proposal C would have been viewed by
suburbanites as a threat to their life style and fiscal independence.

Oliver Williams used Shevky and Bell's life style measure as
the key for explaining the proliferation of metropolitan governments.
In a theoretical essay, he suggested that ”those life style values
which depend upon location for their realization are the major sources
of metropolitan politics."l'O He hypothesized that local political
units will resist integrative efforts for the sake of life style serv-
ices such as education, but will accept integration of system mainte-
nance services such as sewage control, highways, etcetera. Consequently,
if Proposal C was interpreted as an integrative mechanism, life style
characteristics ought to help explain the vote.

Wirt and Walter also studied policy preferences in both sub-
urban and central cities.h' They attempted to determine whether or not
demographic differences between the central city and suburbs are related

to Republican or Democratic voting behavior, Congressional voting be-

havior of suburban legislators, and voting on public referenda.

33

Wirt and Walter factor analyzed demographic variables for A07
communities throughout the country. Four factors emerged from the
analysis. The first three corresponded to the dimensions found by
Shevky and Bell. The four factors were labeled affluence (income, edu-
cation, and occupation), young families (age, fertility, women in the
work force, owner-occupied housing, etc.), ethnicity (percentage
foreign born or nonwhite), and recent growth (percentage of single
family dwellings). These factors were then correlated with Republican
voting, voting by the elected Congressmen, and voting on public refer-
enda.

The most interesting conclusion reached by Wirt and Walter,
at least as it relates to the vote on Proposal C, was that differences
in social status ku12 more important in affecting voting behavior on
public referenda than differences in place of residence (i.e., city or
suburb). Suburbs tended to vary across the spectrum on measures of
affluence (social status), young families, ethnicity, and recent growth.
Policy preferences also varied across the same spectrum. Consequently,
rather than just studying the difference between the central city and
suburb, it is important to study the differences among suburbs as
measured by demographic characteristics and policy preferences.

In contrast to Wirt and Walter, Hawley and Zimmer contended
that the place of residence (either suburban or central city) has been
a key factor in voter support of school consolidation plans in metro-
politan areas. Hawley and Zimmer conducted a public opinion survey of

why voters in metropolitan districts opposed or supported school

3A

consolidation. They asked whether there was more variation in opin-
ions about school consolidation between the central city and the sub-
urban communities than there was between groups of people whether they
lived in the central city or suburban community. They concluded that
”there is much more variation in the views held by place of residence
than by type of resident."li2 The preservation of the suburban life
style was seen as playing a major role in affecting voter support of
school consolidation. While residents of the central city favored con-
solidation, suburban residents opposed consolidation on grounds that
they would lose ”local control” of public education.

Students of metropolitan politics have pointed to the impor-
tance of urban differentiation in determining both politics and policy
preferences. The original indices of urban differentiation developed
by Shevky and Bell have been used to identify the differences between
suburban and central city policy preferences. A common theme in the
literature is that suburban communities will resist integrative policies
that are perceived to be a threat to local control. These policies are
opposed because IDf the impact they might have on suburban life style
and social status.

The social distance explanation presented here suggests that
the vote on Proposal C can be interpreted better in terms of social
integration than as a referendum on property tax relief. The vote
may have been a decision by the public on whether or not to shift the
primary responsibility for funding public education from the local school

district to the state. In this view, support of Proposal C would

35

indicate a desire to alleviate differences among school districts and
among communities. Opposition to Proposal C was indicative of the de-
sire to maintain social, economic, and political differences.

Most studies of urban differentiation have used a derivative
of Shevky and Bell's three dimensions (social rank, urbanization, and
segregation). In Chapter III the research design used in this re-
search is presented. The design uses these three measures to differ-
entiate school districts in seven metr0politan areas of Michigan. In-
dices of social rank, life style (urbanization), and segregation are
computed for each district. The value for each index for the central
city school district is subtracted from the value for each index for
the suburban school district. This gives a measure of the social dis-
tance between the central city and the suburb.

It is hypothesized that the greater the social distance between
the suburban districts and the central city district, the less support
there will be for Proposal C. Testing this hypothesis will give an
indication of the importance of the local control arguments presented
during the campaign for Proposal C. If Proposal C was interpreted as
an integrative mechanism and a threat to local control, social distance
variables should explain the vote on Proposal C better than economic

variables.

CHAPTER II--FOOTNOTES

ISolomon Fabricant, The Trend of Government Activity in the
United States Since I900 (New York: National Bureau of Economic Re-
search, l952), Chapter 6.

 

2Glenn Fisher, ”Determinants of State and Local Government
Expenditures: A Preliminary Analysis,“ National Tax Journal, XIV
(December, l96l), pp. 3A9-355. Werner Z. Hirsch, Analysis of the
Rising Cost of Public Education (Washington: Joint Economic Com-
mittee, I959), and Sherman Shapiro, ”Some Socio-Economic Determinants
of Expenditures for Education,” Comparitive Education Review, VI,

(I962), pp. l69-l66.

 

 

 

 

3Richard E. Dawson and James A. Robinson, ”Inter-Party Compe-
tition, Economic Variables and Welfare Policies in the American States,”
The Journal of Politics, XXV (I963), pp. 265-289.

 

“Richard I. Hofferbert, ”The Relation Between Public Policy
and Some Structural and Environmental Variables in the American States,”
American Political Science Review, LX (March, I966), pp. 73-82.

5Thomas R. Dye, Politics, Economics, and the Public: Policy
Outcomes in the States (Chicago: Rnad McNally, I966).

 

 

 

Ibid.

7Ibid., p. 293.

8See Herbert Jacob and Michael Lipsky, ”Outputs, Structure,
and Power: An Assessment of Changes in the Study of State and Local
Politics,“ in Hofferbert and Sharkansky, State and Urban Politics
(Boston: Little, Brown, and company, I97l), pp. IA-AO.

 

9For an excellent review of how policy has been defined in
different ways see Lewis A. Froman Jr., “An Analysis of Public Policies
in Cities,” Journal of Politics, Vol. 29 (I967), pp. 9A-l08.

'OIra Sharkansky and Richard I. Hofferbert, “Dimensions of
State Politics, Economics and Public Policy,” American Political
Science Review, Vol. 63, September, I969, pp. 880-899.

l'Brian Fry and Richard F. Winters, ”The Politics of Redis-

tribution,‘l American Political Science Review, Vol. 6A, June, I970,
pp. 508-522.

 

 

 

 

36

37

'2Ibid., p. 508.

'3IbId., p. Sll.

'“Ibid., p. 52I.

'5John L. Sullivan, ” A Note on Redistributive Politics,‘I un-
published critique of Fry and Winter's ”The Politics of Redistribution,”
Iowa State University, I972.

6Bernard H. Booms and James R. Halldorson, ”The Politics of
Redistribution: A Reformulation,” American Political Science Review,
Vol. 67, No. 3, September, l973, pp. 92A-933.

 

17Thomas R. Dye, Understanding Public Policy (Englewood Cliffs:
Prentice Hall, lnc., I972), p. 232.

'8Thomas R. Dye, ”A Model for the Analysis of Policy Outcomes,”
in Ira Sharkansky (ed.), Policy Analysis in Political Science (Chicago:
Markham Publishing, I970), p. 23.

 

 

'9Jacob and Lipsky, 0p. cit., p. l9.

20The distribution of resources as the ”heart” of politics has
been the focal point of the new ”political economy“ or rational calcu-
lus approach to politics. This approach includes James M. Buchanan
and Gordon Tullock, The Calculus of Consent: Logical Foundations of
Constitutional Democracy (Ann Arbor: University of Michigan Press,
I962), Anthony Downs, An Economic Theory of Democracy (New York:
Harper and Row, I957),R. L. Curry and L. L. Wade, A Theory of Politi-
cal Exchange (Englewood Cliffs: Prentice-Hall, I968), Mancur Olson,
The Logic of Collective Action (Cambridge: Harvard University Press,
I965), William Riker, The Theory of Political Coalitions (New Haven:
Yale University Press, I962); and many others.

2'For an excellent review of the literature on equal educational
opportunity see: Frederick Mosteller and Daniel P. Moynihan, On Equal-
ity of Educational Opportunity (New York: Vintage Books, I972 . '

22

 

 

 

 

 

 

 

 

Robert Ross, “Public Choice in American Politics: An Intro-
duction,” Introductory Readings in American Government: A Public
Choice Perspective (Chicago: Markham Publishing, l97l), pp. l-ll.

 

 

23L. L. Wade and R. L. Curry, A Logic of Public Policy: Aspects

 

of Political Economy (Belmont, California: Wadsworth Publishing Com-
pany, 1970).

2[William C. Mitchell, ”The Shape of Political Theory to Come:
From Political Sociology to Political Economy,” The American Behavioral

Scientist, Vol. II, No. 2 (Nov-Dec., I967), pp. 8-2I.

 

 

38

25Anthony Downs, An Economic Theory of Democracy (New York:
Harper and Row, l957). P. l8.

26Wendell Bell, “Social Areas: Typology of Urban Neighborhoods,‘l
Marvin Shussman (ed.) in Community Structure and Analysis (New York:
Thomas Y. Crowell, I959). PP. 62-92.

 

 

27Eshref Shevky and Wendell Bell, Social Area Analysis: Theory,
Illustrative Application and Computational Procedures (Westport, Connecti-
cut: Greenwood Press, l972). This is an update of their work on the
Social Areas of Los Angeles (Berkeley: University of California Press,

l9A9I-

 

 

 

28Walter Kaufman and Scott Greer, ”Voting in a MetrOpolitan Com-
munity: An Application of Social Area Analysis,” Social Forces, Vol. 38,
March, I960, pp. l96-22A.

 

291bIcI., p. 201.

30Oliver Williams, Harold Herman, Charles Liebman, and Thomas
Dye, Suburban Differences and Metropolitan Policies (Philadelphia:
University of Pennsylvania Press, I965).

 

3'Ibid., p. 297.

32Karl Deutsch et. al., Political Community in the North Atlantic
Area (Princeton: Princeton University Press, I957).

33Demographic Distance has been defined most clearly by Brett
Hawkins as indicators of life style differences between aggregates of
peOple living in proximity and is trated as indicative of separatist,
or life-style protective values. See Brett Hawkins, “Life Style,
Demographic Distance and Voter Support of City-County Consolidation,”
Social Science Quarterly, Vol. 58, No. 2, Sept., I972, pp. 325-337.

3“Thomas R. Dye, et. al., ”Differentiation and C00peration in
a Metropolitan Area,” Midwest Journal of Political Science, Vol. 7,
No. 2, May, I963, pp. lA5-l55.

 

 

 

35Ibid., p. 1A6.

36Thomas R. Dye, ”City Suburban Social Distance and Public
Policy,” in Thomas R. Dye and Brett Hawkins (eds.), Politics in the

Metropolis (Columbus: Charles Merrill Books, Inc., I967), pp. ISA-
I6A.

 

 

”Ibid., p. 161.

38Robert C. Wood, Suburbia: Its People and Their Politics (Cam-
bridge: Riverside Press, I958).

 

39

39Ibid.

“OOIiver Williams, ”Life Style Values and Political Decentrali-
zation in Metropolitan Areas,” Social Science Quarterly, Vol. A8, Dec.,
I967. P- 303.

“'Benjamin Walter and Frederick Wirt, I'The Political Consequences
of Suburban Variety,” Social Science Quarterly, Vol. 58, No. 3, Dec.,
I972. pp- 7A6-767.

thasil G. Zimmer and Amos H. Hawley, Resistance to Reorganiza-
tion of School Districts and Government in Metropolitan Areas (Provi-
dence: Brown University Press, I966).

 

 

 

CHAPTER III

RESEARCH DESIGN

In Chapter II, two explanations of public choice and public
policy were developed. The economic explanation of public choice sug-
gested the hypothesis that the economic benefit of property tax relief
would best explain the vote on Proposal C. The social distance explana-
tion suggested that the vote would be associated with differences among
social areas as measured by social rank, segregation, and life style.
The purpose of this study is to compare the two explanations of public

choice and public policy.

Hypotheses

 

The hypotheses from the economic explanation state that if prop-
erty tax relief had been the most important criterion determining voter
support of Proposal C, the communities with a heavy tax burden would
have been most interested in tax relief and therefore most likely to
support Proposal C. Conversely, if property tax relief had not been
an important criterion, tax burden should not be related to the vote.

Three variables measure the tax burden of a school district.

The first variable is a ratio of the property tax rate to the value
of property in the school district. This variable measures the tax
effort exerted by the district. If districts voted on the need for

tax relief, communities with greater tax effort should have supported

AO

AI

Proposal C more than communities with low tax effort. The second varia-
ble in the economic explanation is tax burden on income. Support for
Proposal C should have been greater in districts where families pay

a larger share of their income for property taxes than in districts
where families spend a smaller share of their income for property taxes.
The third independent variable is tax burden on homeowners. Many of

the revenues raised for public education come from business property.
Consequently, the burden of the property tax on homeowners varies from
one district to another. Therefore, support for Proposal C should have
been greater in districts where taxes on residential property pay most
of the school costs than in districts where taxes on business prOperty
pay most of the costs.

The social distance explanation theorizes that differences in
social areas, or in this case school districts, can best explain the
vote on Proposal C. The difference or ”distance” between districts is
measured by differences in social rank, segregation, and life style.
The original analysis of social areas by Shevky and Bell identified
these three measures as the key indices of urban differentiation.I
Later research found that these variables were also related to differ-
ences in policy preferences.

The original social area analysis by Bell and Shevky has been
criticized on two accounts. One of the criticisms is that the indices
used to represent the three dimensions were not empirically justified.2
Social rank, life style, and segregation were composities of different

measures from the Federal Census. Social rank consisted of the average

A2

between an education and occupation ratio. Life style, or familism,
was an average of a fertility ratio, the percentage of women in the
labor force, and the percentage of homeowners. Segregation was the
percentage of nonwhite and foreign-born residents in each census tract.
The second criticism is that the number of dimensions used to differ-
entiate urban areas was not justified empirically.3 The question is
why three rather than four, five, or more dimensions were used.

5

Bell“ and later Van Arsdol used factor analysis to determine
whether the component variables were grouped prOperly into the three
dimensions. Bell's analysis confirmed the original dimensions of seg-
regation, life style, and social rank. Van Arsdol found that in six
of the ten cities he studied, the factor structures were in accordance
with the Shevky and Bell conceptualization. In the other four cities,
the factor structures deviated from the three dimensions, with the
component variables not falling into the postulated factors.

The Van Arsdol analysis suggested that perhaps more than three
dimensions are needed to differentiate urban areas. Rees reviewed most
of the recent studies using cluster and factor analysis to construct
indices of urban differentiation.6 He found seven dimensions in the
studies reviewed. In the order of their strength, as measured by how
often they appeared in the research as significant factors, the seven
were socio-economic status, family status, ethnicity or minority group
status, change and mobility (population change and movement), scale
variables (population and population density), health and social prob-

lems, and others. Table 3.l shows what variables are typically associa-

ted with the various factors. The three most often cited factors are

A3

TABLE 3.l.--Classification of Variables Enploycd in Factorial Ecology.

VI.

VII.

.-...._.—_.- .—

Socioeconomic Status Variables

 

l. Population Variables
l.l. Education
l.2. Occupation
l.3. Income
2. Housing Variables
2.I. Quality
2.2. Value of Rent
Household Material Possessions
Mixed Population and Housing Variables (for instance, the
degree of overcrowding)

#W

Family Status or Life Cycle Stage Variables
I. Population Variables

I.l. Age

I.2. Size

l.3- Fertility

l.A. Marital Status
2. Housing Variables

2.i. Type

2.2. Age

 

Ethnicity or Minority Group Status
I. Racial Group

2. Nativity Group

3. Linguistic Group

A. Regional Group (Migrants)

 

Change and Mobility Variables
l. Mobility

l.l. Movement Rates

l.2. Movement classified by origin or destination
2. Population change '

 

Scale Variables

l. Population

2. Area

3. Population Density

A. Locational Measures (such as distance from inner city)

 

Health, Welfare, and Social Problems

 

l. Mental Health

2. Physical Health

3. Welfare

A. Crime and Delinquency

5. School Population Statistics
Others

(A number of other variables such as commuting statistics or
land use measures have been included in factorial ecologies.)

 

Source: Philip Rees, “Problems of Classifying Sub-areas Within Cities,”

In Brian Berry, City Classification, p. 285.

 

AA

consistent with the original dimensions identified by Shevky and
Bell.

To determine whether or not the Shevky and Bell indices were
compatible with the analysis of social areas as determined by school
district boundaries, IA variables for l83 school districts were factor
analyzed. The factor analysis is presented in Table 3.2. Five factors
account for 80.7 percent of the variance and have an eigenvalue above
l.O. They can be labeled social status, family status or life style,
urbanization index, stability and segregation. The first three are
consistent with Shevky and Bell's indices of urban differentiation,
thereby allowing the use of their indices to measure differences in

school districts.

Methodology

 

A multiple regression design is used to compare the two models
of public choice and public policy. Multiple partial correlations
are used to control for the interaction between models. Multiple
partial correlations allow the researcher to determine the relative
success of each model in explaining the vote on Proposal C.

The use of multiple regression analysis to compare two models
has caused some debate in political science. Fry and Winters used
multiple regression analysis to compare an economic model of policy

7

analysis with a political model. Sullivan criticized Fry and Winter's
approach on grounds that they selected their variables in favor of the
political model.8 Fry and Winters used five independent variables for

each model. However, they selected the five most powerful political

A5

TABLE 3.2.--Varimax Rotated Factor Matrix after Rotation with Kaiser
Normalization for IA Variables in l83 Districts.

 

Factor Factor Factor Factor Factor Comm.

 

Variable One Two Three Four Five
Percent l8 .l97 -.772* -.lO6 .277 .090 .733
Percent 65 .l96 -.788* I92 .285 .l75 .8lO

PerCent Families

with children -.OlO -.8A3* -.055 .OA6 -.03O .7l8
Percent staying

in location

5 Years .223 -.059 -.OI8 .IIZ .696* .SAS
Average Rent -.873* .lA5 -.O3O -.I2I -.O67 .8OA
Average Income -.9S3* .02l -.029 .I95 .005 .9A7
Urban Index .396 .029 -.A66 .A78* -.l89 .6Al
People/House .073 -.85A* -.2l2 .280 .ll3 .872
Percent Black .l86 -.l8O .698* ' -.lOl -.O68 .570
Fertility Ratio .36A .A3l* .l82 .2l5 .032 .369
Percent Female

Laborers .OA9 -.637* -.O79 -.I7O -.l63 .A7O
Ed. Ratio -.708* -.I23 .335 .I95 .2AA .727
Occupation '

Ratio -.7A8* .3OI -.O37 -.022 .I29 .669
Percent Home

Owners -.III -.520* -.23l .067 .AAS .530

 

*Indicates what factor the variable is closest to.

 

 

 

Factor Eigenvalue Percent of Variance Cumulative Variance
I 3.96 28.3 28.3%
2 3.62 25.9 5A.2%
3 I-57 ll.2 65.A%
A l.l2 8.0 73.A%
5 l.02 7.3 80.7%

Factor Names

 

Factor: I Social Status

2. Life Style
3. Segregation

A Urbanization Index

5

Stability

A6

variables from a set of nine variables, while selecting five out of
six economic variables. This procedure weighted the research in
favor of Unspolitical variables. Sullivan claimed that in order to
use multiple regression analysis to compare different models, identi-
cal numbers of independent variables must be used. Furthermore,
these variables must be derived from the theory or from empirical
analysis so that the power of each variable in each model is similar.
Rao and Miller stated two conditions for the use of this type

9

of research design. First, the dependent variable must be the same
for each model. Second, each regression equation, or model, must have
an equal number of independent variables. This stipulation is similar
to Sullivan's criticism of Fry and Winters approach. A third condition
made by Sullivan, is that the variables in each model be derived from
theory rather than by selecting the strongest variables from a larger
sample. The dependent variable in this study is the percentage of
voters favoring adoption of Proposal C in l83 school districts; it

is the same for each model. Each model contains three independent

variables derived from theory. Therefore, the stipulations made by

Sullivan and by Rao and Miller have been met.

Unit of Analysis

 

The main unit of analysis is the school district within the
seven Standard Metropolitan Statistical Areas of Michigan. School
districts are used because property tax rates vary more among school
districts than among other governmental units or individuals, and data

were available and were relatively inexpensive.

A7

A second unit of analysis is the census tract within the central
city school districts. Central city school districts have a wide va-
riety of subcommunities within their boundaries. It is important to
determine if the social distance model is as applicable within school
districts as it is among districts. Therefore, a multiple regression
equation is computed for each central city school district using dif-
ferences among census tracts to measure social distance. The economic
model could not be used because of the uniformity of tax rates and the
lack of information on revenues raised in each tract.

The choice of school districts and census tracts as the units
of analysis implies the use of aggregate rather than survey data. Ag-
gregate data are used because the lack of financial resources pro-
hibited an extensive survey of public opinion.

The difficulty in using aggregate data, as W. S. Robinson

l . . .
0 ''Is that one cannot necessarIly Infer

showed in the early fifties,
the correlations between variables, taking people as the unit of
analysis, on the basis of correlations between the same variables based
on groups of people as units.“" Since the Robinson article, further
logical and statistical studies have attempted either to prove or to
mitigate the danger of the I'ecological fallacy.“'2
In a rejoinder to Robinson's article, Menzel argued that not
all aggregate data are used to infer individual characteristics or
relationships. Menzefl believed that territorial units are viable units
of analysis in their own right.'3 Ranney expanded Menzel's argument

by defending the use of aggregate data to explain the behavior of dif-

ferent electorates.

A8

Accordingly, if aggregate data studies carefully and

thoroughly identify and describe recurring patterns of

preferences and turnout characteristics of particular elec-

torates over time, and by ecological correlations, relate

those patterns to other traits of the electorate and their

environments, they can be valuable allies to the sample sur-

veys in the investigation of electoral behavior.
The debate over the ”ecological fallacy“ centers on the level of analy-
sis conducted and the inferences made across different levels. Although
inferences are made about individual behavior, this research design
is primarily concerned with the territorial unit as the level of
analysis.

l5 l6 . .

Scheuch and Alker dISplayed other types of ecologIcal falla-
cies in addition to those found by Robinson. They found that aggregat-
ing individual characteristics and opinions to describe group or terri-
tory behavior often leads to the ”individualistic fallacy.” Specifically,
one cannot infer correlations between variables, taking groups of
people as the unit of analysis, on the basis of correlations between
the same variables based on individuals as the unit of analysis. Sur-
vey research, then, is suspect of misleading the researcher when he
is attempting to describe the behavior of groups or territorial units.

Ecological correlations may tell something about territorial units
that ”can be used as contextual properties explaining the variations
in the correlated variables."l7 They may also provide clues to the
behavior and motivation of individuals within the unit. As Allardt
pointed out:

One may contend that the analysis of ecological data and

correlations may indeed for some researchers in some circum-
stances be a powerful tool in making statements about indi-
vidual behavior. It may be that ecological data facilitate

fruitful causal interpretation betteg than corresponding indi-
vidual data, if such are available.I

A9

The use of aggregate data in this research is defended on grounds
that social areas have an impact on voting behavior and that the use
of ”social distance“ between areas requires either ecological data or
aggregating survey data. Both methods result in problems of inferring
behavior. To answer the question of why differences in social areas
affected the vote on Proposal C, inferences are made about the motiva-
tion of individuals. Naturally, care must be taken in inferring the
psychological reasons for a person supporting or opposing a constitu-
tional amendment. However, as Allardt said, it may be that such in-
ferences will “facilitate causal interpretation better than correspond-
ing individual data.”'9

One method of determining the validity of using aggregate data
is to compare the correlations between variables, based on aggregate
data, with the relationship between the same variables based on survey
data. If the relationships are in the same direction, the researcher
can be more confident in using ecological correlations to infer indi-
vidual behavior.

A survey of attitudes of Detroit-area voters toward Proposal
C showed the same direction of relationships between variables as
those indicated by the ecological correlations. The survey, done by
Market Opinion Research for the Michigan Education Association, was
conducted two months prior to the election. The correlations are
between the same variables as those used in the survey and the vote
on Proposal C (percentage yes). Consequently, there is a two-month

time difference that may have had some effect on the findings. The

50

unit of analysis for the ecological correlations is the school dis-

trict within the metropolitan Detroit area.

 

 

Market Opinion Research Summary Ecological Correlation
”Blacks support C in better proportion +.A3 [Percent Black with per-
than whites (63 percent blacks, A8 cent yes vote on C]
percent white).”

”Support drops as income rises (6l -.A9 [Average family income
percent lowest income group; A9 with yes vote on C]

percent highest).“

”Best support among those 60 and over +.28 [Percent over the age

(6i percent)." of 65 with percent yes vote
on C]

”City voters give better support -.22 [Urban index (high is

(6i percent) than suburban voters rural) with percent yes vote

(A6 percent).“ on C]

”Support rises with educational +.02 [Median educational

level.” level with percent yes vote
on C]

N = A50 individuals N = 8A school districts
Significance Level = .2l7 at
.05 level.

Although only five variables were comparable, all five are in
agreement about the relationship between variables. The link between
education level and support of Proposal C is not significant in the
ecological correlation. However, the percentage of support for Pro-
posal C in the higher educational levels was not given in the survey
report. Consequently, the strength of the relationship between voter
support and education is unknown and open to question. The comparison
of survey data with ecological correlations gives confidence to the

researcher in using aggregate or ecological data.

5l

The Universe

 

The social distance and economic models of public choice are
analyzed in the seven Standard Metropolitan Statistical Areas (SMSA)
of Michigan as defined by the I970 Federal Census. The metropolitan
areas are Detroit, Grand Rapids, Lansing, Jackson, Flint, Muskegon,
and Saginaw. The school districts in these seven metropolitan areas
constitute the universe under study. They were selected because of
two considerations. The first was that to operationalize the concept
of social distance, some school district had to be used as the refer-
ence point.20 In the urban area this decision was easy to make. The
central city school district was the obvious nominee. Many people have
moved from the central city to suburban areas because of the different
school districts involved. Suburbia has become symbolic of a certain
”life style” in which the education system plays an integral role of
maintaining social status and independence.2' If suburban residents
were concerned with maintaining social distance, part of that concern
would most likely be related to their impression of the condition of
the central city district. Thus, the central city district is used
as the reference point in the urban areas. In rural areas the concept
of social distance is less meaningful. With whom do residents in rural
school districts compare their district or life style? It is much
less likely that rural residents are concerned with maintaining social
distance, since the creation of rural districts is more related to
historical precedent than urban differentiation.

The second reason for studying the seven metropolitan areas

IS that Proposal C was defeated primarily because of the vote in the

52

ll counties that make up the seven SMSA's. Table 3.3 gives the vote
on Proposal C in the seven metropolitan areas (ll counties), compared
with the remainder of the state. Analysis of the vote on Proposal C
in the metropolitan areas gives an excellent indication of why the

proposal was defeated state-wide.

TABLE 3.3.--Vote on Proposal C in Eleven Counties and the Remainder
of the State of Michigan.

 

State minus

ll CountIes ll Counties

Entire State

 

Yes Vote 79A,A77 530,225 I,32A,702
No Vote I,25I,608 563,5I8 l,8l5,l26
Percent Yes 38.3 ‘ A8.5 A2.3

 

The counties are sufficiently heterogeneous to include a vari-
ety of school districts ranging from urban to suburban to rural. Table
3.A lists the metropolitan area, counties, and types of school districts
in each county. School districts are classified according to the Michi-
gan Department of Education's classification of school districts into
metropolitan core, city, suburban, small town, and rural.22 School
district boundaries are not coterminous with county boundaries. Con-
sequently, a method of determining whether or not a school district
should be considered part of the county had to be determined. The
decision rule was that if a school district had at least three-

fourths of its land area within the county it was considered to be

53

TABLE 3.A.--School District Classification for l83 Districts in II

 

 

 

Counties
County Core Cities IYP? 0f D'Str'Ct Urban Fringe Rural Total
CItIes Towns
Lansing Area

Ingham l O l A 5 ll

Clinton 0 O l A 3 6

Eaton O O 2 2 3 _;7

2A
Detroit Area

Wayne 3 l A 28 O 36

Oakland l l 6 IS A 27

Macomb O l 3 I6 I 21_

8A
Saginaw I O l 5 6 l3
Muskegon 2 O ' 2 5 3 l2
Flint '

(Genessee) l O 2 l2 5 20
Jackson l O O 3 8 l2
Grand Rapids I O 3 ll 3 l8
Total ll 3 25 l03 Al l83

 

Source: Michigan Department of Education, Local District Report:
Explanatory Materials, I970.

Metropolitan Core: One or more adjacent cities with a population of
50,000 or more which serve as the economic focal point of

their environs.

City: Community of I0,000 to 50,000 that serves as the economic focal
point of its environs.

Town: Community of 2,500 to I0,000 that serves as the economic focal
point of its environs.

Urban Fringe: A Community of any population size that has as its
economic focal point a metropolitan core or a city.

Rural Community: A community less than 2,500.

5A

of the universe under study. Using this rule, the universe consists
of l83 school districts.

A second part of this research is an analysis of the effective-
ness of the social distance model in explaining the vote on Proposal C
in the seven central city school districts. In this part of the re-
search the universe under study consists of all the census tracts or
subcommunities within the central city school district. A composite
index of segregation, life style, and social rank was used to identify
the census tract or subcommunity with the most ghetto-like character-
istics: low social rank, high degree of segregation, and an urban life
style. The identified census tract or subcommunity was used as a refer-
ence point to determine social distance within the city. Census tracts
were used because of their availability at low cost and to make analy-
sis easier. In Detroit, the large number of census tracts (800) made
it economically impossible to use them as the unit of analysis. In-
stead, A9 subcommunities, as identified by the Detroit Mayor's Committee
for Community Renewal were used. The subcommunities were composites
of census tracts with similar geographic, social, and economic char-

23

acteristics; these census tracts had to be geographically contiguous.

Sources of Data

 

The use of school districts, census tracts, and subcommunities
made it possible to use a variety of data. The Michigan Department of
Education provided measures of taxation rates, state equalized valua-
tion of property, and the number of students in each district. The

Federal Census for I970 provided the socio-economic data on educational

55

levels, percentage of craftsmen or laborers, percentage of women in
the labor force, percentage of people owning or buying their own home,
number of children per I,OOO women of child-bearing age, and the per-
centage of Blacks. The census data were placed in school districts by
first determining if the census tract was geographically in the dis-
trict and then apportioning the data if the census tract was in more
than one school district. The census data were apportioned according
to the percentage of land area within the tract residing in the school
district. For example, if one-half of the tract was in one district
and the other half was in another district, the census data were di-
vided in half and distributed equally to each district. The overlap
was estimated as either one-fourth, one-half or three-fourths. Pre-
cise proportions were not computed because of the difficulty in
measuring land area. This estimation procedure undoubtedly resulted
in some error in the final measures. Consequently, some explanatory
power may have been lost.

A third source of data was a survey mailed to local and inter-
mediate school district superintendents.2h The survey asked the super-
intendents to estimate the percentage of property tax revenue raised
from local business pr0perty. lSubtracting the estimates from IOO gave
the percentage of property tax revenue raised from residential prop-
erty. When the local superintendent's estimate was not available, the
intermediate district superintendent's estimate was used. Through
this procedure, a return of IOO percent was accomplished. To test the

validity of the estimates made by the superintendents, a correlation

56

(Pearson's) was computed between the local superintendent's estimate
and the intermediate district superintendent's estimate. The correla-
tion was +.79, which is significant at the .OOl level, indicating a
high degree of confidence in the validity of the estimates.

The final source of data was the Detroit Mayor's Committee for
Community Renewal. The committee provided census data for A9 sub-
communities in Detroit. Their report included 20 demographic variables
that were drawn from the I970 Federal Census and the I965 Michigan

Census.

Dependent Variable

 

The dependent variable for both models is the percentage of
votes in favor of Proposal C in the schoOl districts, census tracts,
and subcommunities. The election results were obtained from the county
clerks' offices in the II counties. Precinct maps were obtained from
the Republican State Headquarters and city clerk's offices.

Placing the vote totals of precincts into school districts,
census tracts, and subcommunities required the same proportioning
procedure as was used in placing census data into the different units.

Again, this procedure undoubtedly clouded the results to some degree.

Independent Variables

 

There are six independent variables, three for each explanation.
The three variables for the economic model are tax effort, tax burden
on homeowners, and tax burden on income. The three independent varia-
bles for the social distance model are social rank differences, life

style differences, and segregation differences.

57

The tax effort of the school district is computed by dividing
the dollars raised from residential property by the total equalized
valuation of property in the district. The dollar amounts raised from
residential property were determined by multiplying the total yield of
local property taxes levied for school operations by the proportion
of the yield raised from residential property as estimated by the
local or intermediate school district superintendent. The equalized
valuation of property is 50 percent of the total value of property
within the school district, equalized throughout the state. In effect,
tax effort is the total tax burden a community is willing to accept
in order to pay for its schools. A

Tax burden on homeowners is computed by dividing the number of
dollars raised from residential property in the district by the number
of homeowners. This provides a measure of the dollars paid for school
taxes per home.

Tax burden on income was computed by dividing the total family
income in the district (income per family multiplied by the number of
families in the district) by the number of dollars raised from resi-
dential property. This measures the proportion of family income paid
in school taxes for the district.

The three independent variables for the social distance model
are the difference between social areas on measures of social rank,
life style, and segregation. Segregation was measured by subtracting
the percentage of black residents in the inner-city school district

or census tract from the percentage of black residents in the suburban

58

districts. In most cases this resulted in a negative value. The
difference was then assigned as the segregation value of the suburban
district. The central city district was given the value of zero.
This provided a measure of the racial distance or separateness be-
tween two districts.

The life style variable was measured by averaging the standard-
ized values of the number of children below the age of five per l,000
women between the ages of IS and AA, the number of women in the labor
force per l,OOO women between the ages of IS and 65, and the percent-
age owning or buying their own home multiplied by l,000. The values
were standardized within metropolitan areas as well as between areas.
The formula used for standardizing the ratios was [(X-X)/S.D.]*IO + 50
where X is the value of the ration being standardized, X'is the mean
score for the ratio, 5.0. is the standard deviation of the ratio, and
* indicates multiplication. This formula standardizes the ratio with
50 being the mean of the standardized scores. The life style index
was computed by adding the three standardized ratios and dividing the
sum by three. The life style index of the central city school dis-
trict or census tract was subtracted from the suburban district's core
to measure the distance between life styles in the two areas.

Social rank is the average of the standardized scores of an
occupation and education ratio. The education ratio is the number of
people with less than eight years of education per l,OOO adults. The
occupation ratio is the number of craftsmen or laborers per I,OOO em-

ployed people. Both ratios were standardized within the metropolitan

59

area. The social rank index was computed by adding the standard scores
for each ratio and dividing by two. The index score of the central
city district or census tract was subtracted from the suburban or

outer census tract to give a measure of the difference in social rank
between areas.

A second part of the analysis consisted of determining whether
or not the social distance explanation was applicable within the central
city school district as it was within the metropolitan area. Social
rank, segregation, and life style indices were computed for each
census tract. The ratios making up the different indices were stand-
ardized within the city according to the formula given above. The in-
dices of social rank, segregation, and life style were then subtracted
from the index scores of the census tract with the highest average
score. The census tract with the highest average score was determined
by standardizing the percentage of black residents within the city,
then adding the three indices and dividing by three. In this manner
the census tract with the highest score--high percentage of Blacks,
low social rank, and an urban life style--was used as the reference
point from which to compute social distance.

Detroit proved to be a specialproblem. The data that were
available for the A9 subcommunities were not identical to the census
data used for the census tracts in the other six communities. There-
fore, a factor analysis of the 20 demographic variables available by
subcommunity was conducted. Table 3.5 presents the analysis. Three

factors, life style, segregation, and social rank, account for 7A.2

60

TABLE 3.5.--Varimax Rotated FaquA Matrix After Rotation with Kaiser
Normalization for IA Demographic Vairables for A9 Sub-
communities of Detroit.

 

 

Variable Factor One Factor Two Factor Three Comm.
Percent Black -.Al2 .6Al -.20l .620
Percent 65 or older -.025 -.89l* .l38 .8lA
Percent l7 or younger -.l23 .733* -.l8l .586
Average Income .6AA .030 .702* .907
Percent Home Owners .888* .Ol5 .379 .93A
Average Rent -.A6l* .030 .063 .2l7
Percent Single Family

Dwellings .9l9* -.020 .3A2 .96l
Female Heads of the

Household -.6AA* -.029 -.390 .568
Education .076 .076 .792* .638
High Socio-Economic

Status .215 -.262 .79A* .7A5
Percent White Collar -.O6A -.A57 .78A* .828
Percent Families with

Children .8IA* .33I .092 .775
People per House .A82 .660* .033 .670
Length of Residence .A60* .lO7 -.05l .226

 

*lndicates what factor the variable is closest to.

 

 

 

Factor Eigenvalue Percent of Variance Cumulative Variance
l 5.09 36.A 36.A%
2 3.7l 26.5 62.8%
3 l.58 Il.3 7A.2%

Factor Names

 

Factor: I. Life Style
2. Segregation
3. Social Rank

6l

percent of the variance and have an eigenvalue of 1.0. They are very
similar to the three dimensions used by Shevky and Bell. Substitute
variables were selected from the three factors to make up for the dif-
ferences in the two sets of data. The substitute variables were
chosen on whether or not they measured the same characteristic as the
original variable and if it was in the same factor. Median years of
education were substituted for the education ratio, percentage of
people employed in white-collar occupations was substituted for the
occupation ratio, the percentage of the population below the age of l8
was substituted for the fertility ratio, and the percentage of women
in the labor force was dropped from the index. These changes will
give a picture of social rank, segregation, and life style for the

Detroit subcommunities that is similar to that for the other cities.

Summary

This research design proposes to compare two models of public
choice and public policy. The social distance and economic models
will be compared by computing multiple regression equations for each
model in each of the seven metropolitan areas in Michigan. The multi-
ple partial correlations between the vote on Proposal C and the two
models will provide an answer to the question of which model best
”explained” the vote. A multiple regression equation will also be
computed for the social distance model in each central city school
district. This is done so that the strength of the model within a
school district can be compared to the strength of the model among

school districts.

CHAPTER III--FO0TNOTES

lEshref Shevky and Wendell Bell, Social Area Analysis: Theory
Illustrative Application and Computational Procedures (Westport,
Connecticut: Greenwood Press, 19727.

 

2See Otis 0. Duncan, ”Review of Social Area Analysis by Shevky
and Bell,” American Journal of Sociology, Vol. 6l, I955, pp. 84-95.

 

3See Phillip Rees, ”Problems of Classifying Sub-areas within
Cities,” in Brian Berry (ed.), City Classification Handbook of Methods
and Applications (New York: Wiley lnc., I972), pp. 265-330.

 

 

AWendell Bell, ”Economic, Family and Ethnic Status: An Empiri-
cal Test,” American Sociological Review, Vol. 20, I955, pp. A5-52.

 

5Maurice Van Arsdol, Santo F. Camilleri, and Calvin F. Schmid,
”The Generality of Urban Social Area lndices,” American Sociological
Review, Vol. 23, I958, pp. 277-28“.

 

6Phillip Rees, op. cit., p. 285.

7Brian R» Fry and Richard F. Winters, ”The Politics of Redis-

tribution,” American Political Science Review, Vol. 6“, June I970,
pp. 508-522.

 

John L. Sullivan, ”A Note on Redistributive Politics,” un-
published paper (Iowa State University, Ames, Iowa), I972.

9Tohluri Rao and Roger LeRoy Miller, Applied Econometrics
(Belmont, California: Wadsworth Publishing, I97l), p. I8.

loW. S. Robinson, I'Ecological Correlations and the Behavior of
Individuals,” American Sociological Review, Vol. I5, June, I950,

PP- 351-357.

IIW. Phillips Shively, “Ecological Inference: The Use of Aggre-
gate Data to Study Individuals,” American Political Science Review,
Vol. 62, December, I969, pp. ll83-ll96}

l2For an excellent collection of essays on the topic see:
Mattel Dogan and Stein Rokkan, Quantitative Ecological Analysis in the
Social Sciences (Cambridge: M.I.T. Press, I969).

 

 

 

 

 

62

63

13Herbert Menzel, ”Comment on Robinson's Ecological Correla-

tions and the Behavior of lndividuals,‘I American Sociological Review,
Vol. 15 (October, 1950), pp. 67h-675.

 

MAustin Ranney, I'The Utility and Limitations of Aggregate Data
in the Study of Electoral Behavior,“ in Austin Ranney (ed.), Essays on
the Behaviorlal Study of Politics (Urbana: Univeristy of Illinois
Press, I962), pp. 91-102.

 

ISErwin K. Scheuch, ”Social Context and lndivdual Behavior,”
in Dogan and Rokkan, 0p. cit., pp. 69-86.

‘6Hayward R. Alker, Jr., ”A Typology of Ecological Fallacies,”
in Dogan and Rokkan, Op. cit., pp. 69-86.

l7Erik Allardt, ”Aggregate Analysis: The Problem of Its Informa-
tive Value,” in Dogan and Rokkan, op. cit., p. #2.

'8lbld., p. till

'9lbid., p. #2.

20A second method of operationalizing the concept of social
distance would have been to use Dye's method of using all possible
pairs of communities. Two problems arise when using this strategy.
First, it would have been very difficult to determine the dependent
variable and assign a value to each pair of communities. Secondly,
the large number of pairs would have made using this method a very
cumbersome task.

2lFor a good description of the suburban phenomenon see Robert
Wood, Suburbia: Its People and Their Politics (Boston: Houghton
leflin Co., 1958).

 

2Michigan Department of Education, Local District Report:
Explanatory Materials for Assessment Report No. 6, 1971.

 

 

23Mayor's Committee for Community Renewal, A Profile of
Detroit: 1969, City of Detroit, Michigan, 1970.

 

 

2MA copy of the survey is included as Appendix A.

CHAPTER IV

FINANCING PUBLIC EDUCATION IN MICHIGAN

Michigan has not been immune to the financial crisis facing
public education throughout the United States. In a survey of school
superintendents, Kllngele found that 94 percent of the school dis-
tricts in Michigan experienced financial problems during the 1972-73
school year which required the curtailment of expenses.I Fifty-five
school districts operated with deficit budgets.2 One district was
forced to reduce the hours of education below the legal limit.3 In
Detroit, the financial situation forced larger class sizes, less ex-
tensive extra-curricular programs, fewer teachers, teachers being
hired as full-time substitutes, and other budget-reducing tactics.
The situation deteriorated to the point where a federal court judge
ordered the Detroit Public Schools to maintain their 1971-72 pro-
gram during the 1972-73 school year. In the spring of 1973 the
Michigan Legislature had to enact special legislation to ensure ade-
quate funds for Detroit.

The means by which school districts curtailed expenditures
varied. The Kllngele study identified the 15 most frequently used
methods of limiting expenses. Table “.1 gives the percentage of school

districts that used that particular method of limiting expenditures.

6h

65

TABLE h.l.--Percentage of Michigan Public School Systems Indicating
Methods Used for Curtailing Expenses.

 

 

1...... 22:22?
Tightening purchasing and accounting controls 72
Purchasing less equipment, materials, and supplies 59
Enlarging class size 45
Cutting travel expenses (to conferences, etc.) Ah
Cutting maintenance A0
Reducing the number of faculty 34
Freezing the development and implementation of new

programs 33
Reducing administrative staff 30
Reducing academic programs (P. E., music,

kindergarten) 27
Reducing transportation expenses (busing) 23
Not filling positions caused by attrition 23
Cutting athletic programs 12
Halting all new programs 10
Terminating co-curricular activities 10
Freezing salaries 6

 

Source: William E. Kllngele, “Curtailing Expenses in Michigan Public
Schools,” in Michigan School Board Journal, Vol. XX, April,
1973. P- 23-

 

N=l65

66

As can be seen, most school districts in Michigan were directly af-
fected by the financial crisis facing public education in the United
States. The crisis in Michigan was precipitated by rising costs, the

exhaustion of the property tax, and inequities in school finance.

RisingfiCosts

 

The first component of the financial crisis in Michigan is the
sharp rise in the cost of education over the last ten years. The state's
expenditures for kindergarten through twelfth grade, exclusive of
locally raised funds, went from $299 million in 1961-62 to $813 million
in l97l-72. During the same period, per pupil expenditures rose from
$375.66 to $875.26. Total operating expenses for the entire educational
system increased by 166 percent, from $611.7 million to $1,019 million.‘4

During the same decade, the average teacher salary rose from
$5,898 to $11,685. This dramatic increase could be partially attribu-
ted to the organization of teaches into bargaining units. 0f Michigan's
601 school districts, 570 recognize a teacher organization as the bar-
gaining agent for their professional employees. This figure includes
virtually all districts of any appreciable size.5 The recognition of
bargaining agents in so many school districts has greatly increased
the influence of the two teacher organizations in Michgan: the Michi-
gan Education Association and the Michigan Federation of Teachers.

The rising costs of public education in Michigan are given in

Table h.2 for the time period from l9h9-50 to 1970-71. The consumer

price index is also given to account for inflation.

67

TABLE 0.2.--School Revenues by Source for All Public School Districts
in Michigan for Selected Years (in Thousands).

 

 

Percent
Source 1909-50 1950-55 1959-60 1965-66 1970-71 Change
Property Tax 113,569 206,363 307,701 510,363 1,027,002 + 905
(35%) (35%) (00%) (95%) (49%) + 14
State Aid 128,000 200,153 236,295 072,389 811,971 + 632
(39%) (30%) (27%) (01%) (38%) - 1
Federal 5,830 6,903 8,725 18,833 23,861 + 008
( 2%) ( 1%) ( 1%) ( 1%) ( 1%) - 1
Others 78,821 175,736 282,599 137,626 200,622 + 305
(26%) (30%) (32%) (13%) (12%) - 10
Total 326,668 589,155 875,360 1,103,212 2,113,096 .+ 607
(100%) (100%) (100%) (100%) (100%)
Consumer
Price Index 83.0 101.5 127.7 + 150

 

Source: Bulletin No. 1011, Michigan Department of Education and the

National Educational Finance Project.

68

Increasing costs, in themselves, are not a threat to the education
system. However, when rising costs are combined with a limited tax

structure, problems result.

Exhaustion of the Property Tax

 

The second component of the financial crisis is the ”exhaustion
of the property tax.” This expression can be interpreted in a number
of ways. First, it can be seen as the unwillingness of the citizens
to support increased taxation. Secondly, it can be interpreted as the
effect of constitutional property tax limitations. The Michigan Consti-
tution provides that the total number of mills levied on any property
in the state may not exceed 50 mills. Charter cities and townships
may, however, write different limitations into their charters. This
limitation was partially negated in 1971 by a decision of the Michigan
Supreme Court. The court ruled, in Butcher v. Grosse Isle, that there
was no limit on millage used for bonded indebtedness as long as it was
voted by the people. As of fall, 1972, there was no way of knowing
how many districts were close to the constitutional limitation. Fi-
nally, property tax exhaustion can be interpreted as reaching a maxi-
mum yield in comparison to other tax alternatives.

Property tax revenues, prticularly those used to support edu-
cation, have risen more rapidly than revenues derived from any other
source. Figure 1 compares the rate of increase in property tax yield
with other taxes, with 1959 serving as the base year. Much of the

recent increase in property taxes is clearly due to the rising cost

of education. Figure 2 shows that the educational property tax has

2.00.

1.750

........ Corporate Income Tax
——-—- Sales Tax
————- Property Tax

   

 

1959

Fflgure l.--Changes

 

 

19650
1970-

in the Yields of Three State Taxes, 1959‘1971.

SSCDurce: Citizen Research Council of Michigan, Legislative Handbook.

69

 

....... County Taxes /'
City Taxes '
3.00 -..—- School Taxes ./
/
2.75. /
/

 

 

 

1959
1965
1970

F:IGURE 2.--Changes in Property Tax Revenues for Three Purposes 1959-1971
U9S9==I-0)-

$<3Larce: Citizen Research Council of Michigan, Legislative Handbook.

70

71

increased at a much faster rate than have property taxes for other
purposes.

If property taxes are to be held down to a level at which the
public will support public education, and if, in particular, they are
not to “bump” against the constitutional limitation, the educational
property tax is the most obvious candidate for trimming or elimination.
This is the context in which political leaders, particularly the
Governor, would see the problem and the need to offer a constitutional
amendment to limit the property tax.

As mentioned earlier, the much-noted ”taxpayers' revolt” is
also an indication of the ”exhaustion of the property tax.” Unfortu-
nately, the Michigan Department of Education did not start collecting
data on local millage elections until 1968. The curves in Figure 3
show the proportions of millage elections won, as well as of requested
mills passed, during the period for which data exist. The curves,
although smoothed by the use of three—month moving averages, are very
uneven. In spite of vigorous up and down swings, there is no evidence
of a major trend related to the years between 1968 and 1972. If any
trend exists, it is toward increasing acceptance of millage proposals.
It may be that millage defeats were less common prior to than follow-
ing 1968. However, this information was not available.

In Pinner, Collins, and Sederburg's study of financial reform

3" Michigan education, a number of superintendents argued that millages
”VEBre more difficult to pass than in previous years.6 However, defeat
()rlly meant that the superintendents had to return to the voters with

an(other millage request. It was argued that in Spite of all the trials

90%‘

85%‘

3:: f‘ (1.1

70%,

\\
\c
\‘k“

 

65 % , = 1 . . \ ' a. ," .
I /II/ \I \j / \‘1 \ ' ,‘
I : 1‘ 1 / f: ,
‘ I [/1 \\ I g I \, i ‘, .1 .

 

 

 

 

 

.
9, i, I f ' l !I
600 4 ‘3 f.' N I I I If \v 1I
*= 1 ’ l 1 . 1 41.
J i' ' I )5—._ l I l " I'll \E'
I " , ‘ J I 1 1 11 W
55% -. :1 f I ‘ / 'l 1' ‘/'I
I \1 ‘ \ ‘ - I 0 1 I
II b! l . 1 '. ,
I I/\I J B“ V,» I I I i I
1'- \I -- ., 1 :
50% . i. f i) ' ~ I I
1' : ‘
hi! U 2“ I} .
H ‘. 1;
45% .1 11/ ‘1“ 1
I 0
9.
40° “I '7 '—'T T“? T r 'r‘ T“ j n L fi—T] ‘fi T‘ 1 . I T It lyfi f“?
N
HA A Q ’\ A Oh ’\ rs r-{A A
r—loo srcn. Mow <10 r—IC) '33‘1—1 cor-l Ir—I \ON
1 o I o 1 o I n I n IP~ IV\ 01\ I h
oxv NV cov NV gov NV \0" av vv
A _ Total Villages Passed B = Number of Millage Elections Won
Total millages Requested Number uf Elections Held

Figure 3.--Percentage of Millage Eelctions Won and of Requested Mills
Passed, September, 1968 to August, 1972.

SOurce: Michigan Department of Education, Operating Millage and Bond

Issue Reports.

72

73

and tribulations, most districts managed to obtain the millage they
needed.

To test this theory Pinner et_al: related the instructional
dollars per pupil to the frequency of the millage defeats. Table 0.6
shows the zero order and the partial correlations between per pupil in
structional expenditures and various factors, including percentage of

millage defeats.

TABLE 0.6.--Zero-Order and Partial Correlations between Per Pupil
Expenditures and Selected Variables in 182 School Districts.

 

Correlations
Zero-Order Partial

 

 

Percent of millage elections lost .100 .053
Number of pupils in district .565 .307
SEV per pupil .501 .098
Urbanization -.569 -.303
Average Socio-economic status of families in

district .166 .196
Percent white -.322 -.232

 

Source: Frank Pinner et al., The State and Education: Decision Making
on the Reform of Educational Finances in Michigan, a report
to the Urban Institute.

 

The correlations are based on all 182 school districts for which millage
elections were reported by the Michigan Department of Education during
the period from September, 1968, to April, 1971. The average SEV
(Estate Equalized Valuation) of these districts was slightly below the
state average--$lS,073 per pupil as against $16,321 for the state as
5’ \whole. Their average socio-economic status was 50.35, which is very

(:liose to the state average of 50.00 (the mean of the distribution of

70

standard scores used by the Department of Education). The socio-
economic status scores were obtained from the 1970 Michigan educational
assessment program. The assessment tests asked a series of questions
related to socio-economic status of fourth and seventh grade students
in Michigan school districts. The closeness of the state average and
the average of the 182 school districts makes it relatively safe to say
that the 182 districts are representative of all the districts in
Michigan.

These data show virtually no effect of millage defeats upon the
level of instructional expenditures, particularly when all other major
variables for which data were available had been controlled for. This
approach, however, does not answer the question of whether or not
school districts had to limit their programs during the past few years.

To answer this question, a brief poll of the superintendents
in the seven metropolitan areas used in the analysis presented in
Chapter VI was conducted. The superintendents were asked if they had
been forced, because of financial pressure, to reduce their educational
programs during the past two years. Most superintendents responded by
replying no, they had not had to reduce their programs, but they could
not expand them. Typical of the comments received from local super-
intendents were:

On the other hand, we have had to maintain a no improvement
status quo for the past three years and will continue to do

in 1973'70.

We have not been able to expand programs, and this is so
badly needed.

()f: the superintendents in the 183 districts surveyed, 53 percent indi-

<:*3‘ted they had not had to reduce their programs during the past two

75

years. On the other hand, 07 percent of the districts actually were
forced to reduce the educational program in their districts.7
The perceived ”taxpayers' revolt” may be less severe than imag-
ined by some. Wirt and Kirst, in a study of California referenda, found
that the public was voting against any change in the property tax level
rather than against education.8 They concluded that ”it seems as if
change of any kind is more opposed than the status quo."9 Thus, al-
though there may be an increasing tendency for voters to reject addi-
tional millages for education, the voters also rejected lower millages
for education. If Wirt and Kirst are correct, it appears reasonable to
say that the general attitude of the public is to maintain the status
quo on property tax rates. The cost-revenue squeeze in public education

may be due to increasing costs coupled with a tax structure that has

become frozen at the status quo.

lnequity in School Finance

 

Increased sensitivity to inequalities in funding education con-
stitutes the third component of the financial crisis. These inequalities
are inherent in any system that relies heavily on local property values
as a tax base. To reduce the degree of inequality among districts (and
the burdens that increasing educational costs impose on local taxpayers)
the State of Michigan supplements local tax revenues by two types of

state aid. The first, based on the number of pupils in a district,
atitempts to compensate for variations in local property wealth. The
Second type of state aid attempts to compensate, by flat grants, for the
Special needs of various categories of pupils, such as the physically,

'jweirwtally, or emotionally handicapped and the culturally deprived.

76

The first form of state aid is embedded in the general aid form-
ula that allocates state funds to local districts under the provisions
of the School Aid Act of 1957. To meet changing financial requirements,
the act is revised annually. 0f crucial importance in applying the
formula to any given district is the concept of State Equalized Valua-
tion (SEV) per pupil. To compensate for differences in local assessment
practices, the total assessed property valuation in a district is ad-
justed (equalized) by the state so that the resuting amount represents
as closely as possible 50 percent of the current market value of all
real property. This figure, when divided by the number of |'members"
(full-time equivalent students), yields the district SEV per member,
or, for short, ”SEV.“ The law specified a gross allowance per member.
To obtain the state's per student contribution, this allowance is re-
duced by a ”deductible millage” (also specified by law) multiplied by
the SEV. In most years since this method for computing state aid was
enacted, Michigan has operated under a two-part formula, with different
gross allowances and deductible millages applying to districts above
and below an SEV level specified by law. Both the gross allowance and
the SEV level defining the breakpoint for the application of the two
alternative formulae have been revised upward from year to year. The
”deductible millages” have also been changed from time to time. For
fiScal year 1972-73, the legislature again adopted a two-part formula

‘Vi‘th the break point set at the SEV of $17,750. For high SEV districts,
tifiea deductible gross allowance was $600, and the deductible millage was
$ 16 for each $1000 of SEV. For low SEV districts, the corresponding

F i gures were $715 and 20 mills.

77

A simple illustration will Show that this formula, while it
reduces disparities among districts, does not come close to eliminat-
ing inequalities in funds available for each child in districts of
varying SEV levels. Assume that two districts have SEV's of $10,000
and $20,000, respectively. Assume further that both districts tax
themselves at the rate of 26 mills, which is the state average. In
the low SEV district, the combination of locally raised property tax
income and of state aid will produce $775 per pupil ($515 state aid
and $260 local tax), whereas in the high SEV district, $800 will be
available for each child ($320 state aid and $520 local tax). More-
over, in a low SEV district, a 26 mill rate may constitute more of a
hardship than in a district with higher property values because the
residents in districts with low property values typically have
smaller incomes.

In addition to general state aid, school districts receive
various types of grants for transportation, vocational education,
remedial reading and other programs. Further, a provision makes it
possible to reduce the SEV of a district whose overall property tax
rate for all purposes (except school operating costs) exceeds the
state average by 25 percent or more. The purpose of this disposition
is to bring relief to the taxpayers in cities suffering from “munici-
pal overburden“--essentially the large cities with high operating costs
that must provide services for large numbers of nonresidents working
and doing business in town.

The joint operation of the various state aid, grant, and relief

provisions produces results that are often discriminatory or irrational,

78

usually both. The number of educational resources available to a
child depends on the social status characteristics of his community,
the amount of business and industrial property located in his school
district, the density of the school population, and the ability and
aggressiveness of the school administrators in pursuing additional
funds under various grant programs. The inequities are illustrated in
Table 0.7. Data are taken from the l97l-72 school aid formula be-
cause Proposal C was written, and debated, about eliminating these

inequities.

TABLE 0.7.--Financial Data for Selected School District 1971-1972.

 

 

W 1111:. 12:11:93.2}. 12:31.31? i135???

Salary
River Rouge 66,800 20.9 $1,027.63 $1,500.36 $10,895
Inkster 8,556 25.9 661.13 936.01 11,729
Detroit 19,872 20.80 700.35 1,003.35 13,707
Ann Arbor 31,608 32.55 972.60 1,296.23 12,763
Lansing 20,515 30.0 800.77 1,131.26 11,093
East Lansing 26,688 30.50 893.87 1,277.01 11,000
Redford Union 13,935 35.90 701.90 952.15 12.093

 

River Rouge is a working-class suburb of Detroit that benefits
from the presence of several large industrial plants. At a moderate

cost to the taxpayer, it can afford one of the most expensive school

79

systems in the state. Inkster is an entirely residential, working-
class suburb (predominantly black) that, in spite of a much higher tax
effort, has much less to offer its children. Redford Union is an ex-
ample of a district relatively low in property wealth that, at consid-
erable cost in taxes, can achieve only mediocre results. These, as
well as the other figures in Table 0.7, are only indicative of the
problem of inequality, which has been amply studied and discussed in

several important documents. (Thomas Report, 1968, and Schools and

 

 

Inequality, 1969).lo

 

The increased sensitivities to the inequalities built into the
'method of financing education are probably a result of the national
attention given to court cases, research studies, tax pressures, and
the dramatic accentuation of inequalities during the past few years.
In 1966-67, the average SEV for districts in Michigan was $13,988.
Six years later, in 1971-72, it had risen to $18,883, an increase of
38 percent. The standard deviation in district SEV's increased by 20
percent from $7,011 to $9,207. The amount of variability in district
SEV's, as measured by the standard deviation, was already quite high
in 1966-67, and it increased even more during the next five years.
This indicates that some districts experienced increases in their prop-
erty wealth well above the state average, others remained stationary,
and a few experienced losses.

Similar comparisons based on the funds available for each pupil
complete the picture of inequitable changes. During the same six-year
period, average instructional expenditures per pupil rose by 91 percent

from $328.31 to $627.00. The spread in per pupil expenditures also

80

increased, but owing to the “gross allowancs” and other features of the
state aid act, it was not nearly as sharp as the cost increase. The
standard deviation increased by 71 percent. Even so, differences in the
dollars available for the education of children in different districts
had augmented further.

The most dramatic shifts, and those likely to have the greatest
impact on officials as well as voters, occurred in the dollar amounts
per pupil derived from local sources. 0n the average these increased
by 35 percent from $320.02 to $032.00. At the same time, the amount of
disparity in locally raised dollars increased by 92 percent as the
standard deviation went from $89.89 to $172.85. The rate of increase
in locally raised dollars was nearly four times the rate of increase in
SEV levels. The disparities obviously reflect rapidly rising millage
rates. But thisincrease had to be quite uneven to produce a 92 percent
increase in the standard deviation. The better situated districts must
have increased their millages much more drastically than the districts
with less favorable property conditions. The latter, in all likelihood,
were unable to keep up with inflation.

The shifts in prOperty values have had a particularly adverse
effect on the major cities in the state, as can be seen in Table 0.8.
The SEV of core cities has grown more slowly than the state average,
whereas that of the surrounding suburban areas has increased more rap-
idly.

There can be little doubt that the issue of inequality had been
brought to the attention of all groups and people involved in the edu-

cational policy-making process. It was one of the major concerns in

81

TABLE 0.8.-~SEV in Three Major Cities and Surrounding County Districts.
l966/67 ' 1970/71.

 

 

Percent

1966/67 1967/68 1968/69 1969/70 1970/71 Change
Michigan
Average $13,898 $10,059 $15,198 $16,321 $17,760 22
Wayne County
except Detroit 16,096 20,550 19,300 20,872 22,815 28
Detroit 16,665 16,261 16,808 17,720 18,307 9
Kent County
except
Grand Rapids 11,262 12,065 12,938 13,730 10,703 20
Grand Rapids 18,928 19,727 20,191 20,303 21,500 12
Genessee County
except Flint 9,722 10,299 11,068 12,106 13,796 30
Flint 17,350 17,005 18,285 19,087 20,063 15

 

the Thomas and Guthrie studies. It has figured prominently in Governor
Milliken's several messages to the legislature, and its reduction was
the first objective listed by the Governor's Commission on Education
Reform, where it was stated in the form that became a cliché. The
Commission's aim was: ”To assure each child, no matter where he may
live or what his circumstances may be, an equal educational opportu-
nity.”

The combination of increased expenditures, the exhaustion of the
property tax, and the inherent problems of inequality resulting from

reliance on the local property tax created an atmosphere throughout

Michigan that was hOSpitable to increased pressure to reform the

82

financing of public education. The financial pressures were reflected
in political strategies aimed at removing the burden of the property tax
from homeowners. The key to these strategies was the need to change the
Michigan Constitution to allow greater state funding. The result of the
political strategies was a constitutional amendment presented to the
Michigan electorate at the November, 1972, general election. The pro-
posed amendment, known as Proposal C, was the primary effort to reform

school finance in Michigan during 1972.

CHAPTER lV--FOOTNOTES

1William E. Kllngele, ”Curtailing Expenses in Michigan Public
chools,” Michigan School Board Journal, Vol. 20, April, 1973, p. 23.

Interview with Mr. Robert McKerr, Associate Superintendent for
Finance - State Department of Education, April 5, 1973.

31bid.

“Michigan Department of Education, 1012 Bulletins 1960-1971,
Lansing, Michigan.

 

5Interview with Kai Erickson, Director of Public Relations,
Michigan Education Association, November 6, 1973.

Frank Pinner et. al., Education and the State: Decision Making
on the Reform of Educational Finances in Michigan, report to the Urban
Coalition, 1971. Chapter Three.

7Copy of the survey is found in Appendix A.

8Frederick Wirt and Michael Kirst, The Political Web of Schools
(Boston: Little Brown and Company, 1972).

Ibid., p. 105.

loJ. Alan Thomas, School Finance and Educational Opportunity in

Michigan (Lansing, Michigan: Michigan Department of Education, 1968).
James Guthrie, George Kleindorfer, Henry Levin and Robert Stout,
Schools and Inequality (Washington: The Urban Coalition, I969).

IlOffice of Planning and Coordination, Bureau of Policies and

Programs, A Chronology of Educational Reform in Michigan, Lansing,
Michigan, 1970, p. 6.

83

CHAPTER V

THE POLITICS AND HISTORY OF PROPOSAL C

The financial crisis faced by public education in Michigan
provided the impetus for increased political pressure to reform the
method of funding education. The political history of financial re-
form prior to the vote on Proposal C showed the interaction Of the
economic and social distance models of public choice and public policy.
It is clear from an analysis of the political history of Proposal C
that most state policy makers debated the issue in terms of the
economic effects of shifting from a property tax to an income tax. This
is particularly interesting because the vote is best explained by the
social distance explanation.

Students of the politics of education at the state level have
concentrated on the differences between ”closed” and ”open” systems of
educational politics. Iannaconne, Bailey, Masters, Kirst, Usdan,
Bowles, and others have discussed the relative merits of the education
community presenting an united front or being "fragmented."I The level
of educational expenditures in a state, according to these authors, is
dependent on the ”ability of the schoolmen to present a cohesive
front.”2 They did not discuss the politics of different financing plans
or the distributive impact different plans have on school districts

within states.

80

85

According to Masters and Salisbury, and Iannacconne, the over-
riding characteristic of educational politics in Michigan is the ”lack
of consensus'i among the players of the political game.3 The lack of
consensus noted by Masters could also be seen in the political history
of financial reform and the debate over Proposal C. The history of
Proposal C appeared to confirm Master's hypothesis about the red flag
of school finance.

Issues that involve basic changes in the revenue structure or
involve substantially increased expenditures for educational
purposes, can activate groups that are capable of exerting
strong counter pressures and manipulating strong consensus
building symbols in Opposition to the proposed changes or
increases in expenditure levels.
The debate over financial reform showed the division within the educa-
tion community as well as the appearance of noneducation groups that
were activated because of the taxation changes involved in educational
finance reform.

Pinner, Collins, and Sederburg, in a study of financial reform
in Michigan education, found that the political coalitions involved in
taxation questions were different from those formed with respect to
school aid distribution.5 The development of Proposal C, and of finan-
cial reform generally, activated both types of coalitions and involved
a large share of the political community in Michigan.

A brief review of this activity shows the dynamic of reform-
ing school finance in Michigan and how this dynamic can be discussed
in terms of the economic and social distance explanations. First, a

history of Proposal C is given. Emphasis is given to the struggle

between the Governor and the Legislature, the resulting stalemate in

86

the legislature, and the initiative petition drives conducted by the
Governor, the Democratic Party, and the Michigan Education Association.
Second, a brief description of the campaign conducted for Proposal C
and the issues brought out in the campaign are presented. Last, a gen-

eral review is made of the results of the vote on Proposal C.

History

The history of Proposal C dates back to at least 1966. In 1966
the Michigan State Board of Education and the Michigan Legislature
spent $200,000 for a thorough analysis of educational opportunity in
Michigan. The study was conducted by Dr. J. Alan Thomas. Completed
in the fall of 1967, the Thomas report concluded that educational oppor-
tunity varied greatly throughout Michigan. The report suggested that
the state play a greater role in equalizing the financing of public
education and offered four alternative plans that would make for
greater equality.6

The report was commissioned primarily on the initiative of the
Democratically controlled State Board of Edcucation and legislature.
However, their goal of equal educational opportunity was embraced by
Republican Lt. Governor William G. Milliken when he became Governor in
1969. When Milliken took office in 1969, the political climate was
generally favorable to the need for educational reform. Milliken soon
announced his intention of appointing a special commission of non-
educators to study legislative needs for educational reform.

The Governor's proposal to create a special commission began

the debate over school finance reform. The proposal was criticized

87

as ”a delaying tactic,” “paralysis through analysis,” and ”a study of

a study."7

The commission was primarily an attempt by the Republican
Governor to gain the initiative in education reform. Governor Milliken
had long been an active participant in educational reform as Chairman
of the Senate Education Committee. It was natural that he selected
education as his main priority during his first term in office.

In October of 1969 Governor Milliken presented to the legisla-
ture his special message on educational reform. Basing his recommenda-
tions on the commission's report, Milliken outlined the basic objectives
for which he would strive during the 1969 and 1970 legislative sessions.
These included the reduction, if not elimination, of inequality in
school expenditures among school districts.. This objective was to be
attained by eliminating the local property tax, except for a locally
adopted three-mill ”enrichment” tax, and by guaranteeing a fixed basic
allowance for each pupil. Each mill of the ”enrichment” tax was to be
equalized, so that a mill levied anywhere in the state would raise the
same amount of money. Monies were to be distributed to the different
districts according to a plan developed by the Michigan Association of
Professors of Educational Administration called the Classroom Unit
Plan. This plan would provide equal funds for every classroom of 27
children in any district in the state.

Second, Milliken called for the elimination of the inequality
of tax burden among school districts. This would be done by shifting
the financing of public education away from the local property tax to
a statewide property tax. Property taxes would be collected by the

state and then redistributed to the local districts.

88

Another objective was the ”rationalization” of the governance
of public education to insure accountability. This was to be accomp-
lished by establishing regional superintendents of public instruction,
appointed by the Governor, an appointed State Superintendent of Public
Instruction, a state assessment to test achievement of school children,
and aid to nonpublic schools.

The last objective of the Governor's message became the most
controversial. In his message to the legislature, Milliken asked for
their approval of public aid to parochial schools. Parochiaid, as it
became known, was added to the reform package to gain the support of
Democratic Speaker of the House, William Ryan. Ryan, a strong Catholic
from the inner city of Detroit, was very interested in aiding the fi-
nancially depressed parochial schools in Detroit. Although undoubtedly
interested in aid to nonpublic schools, Milliken also could see the
need to get Speaker Ryan's help in passing the reform proposals through
the Democratically controlled House of Representatives.

The Fall, 1969, and Winter, 1970, sessions of the legislature
saw only limited action on Milliken's proposals. Only aid to non-
public schools and state assessment were fully debated. The debate on
parochiaid clouded the major issues Milliken wanted to stress. Paro-
chiaid and assessment were the only proposals passed by the legislature.
The Michigan Education Association, along with over 20 other organiza-
tions, organized a petition drive to place the question of parochiaid
on the November, 1970, ballot in the form of a constitutional amendment.

The electorate overruled the legislature and the Governor and voted to

89

prohibit aid to nonpublic schools. Thus, the net results in reforming
education and educational finance were meager. The legislature passed
a state aid bill for public education that was similar to the 1969-70
school aid bill.

The governor's original proposals for educational reform if
passed, would have established a state system of education. His plan
called for a statewide means of collecting and distributing funds,
regional superintendents appointed by the Governor, a state superintend-
ent appointed by the Governor, and aid to nonpublic education. In
pushing his proposals, the Governor stressed the economic arguments
of equality of tax burden and state aid distribution. He ignored the
importance and strength of the arguments used in favor of ”local con-
trol” of public education. The proposals were opposed by most members
of the educational community because of the threat they posed to the
role played by the local and intermediate school districts. The oppo-
sition to his proposals by the public and educators meant the failure
to achieve any significant financial reform. In effect, Milliken's
proposal had within it elements of both the social distance and econ-
omic explanations. First, he stressed economic arguments in favor of
a state system of educational finance. Second, he challenged the im-
portance of the social distance explanation by (l) ignoring the role
played by the local district in financing and administering the schools
and (2) opposing the political strength of the local districts.

With the legislative stalemate over his initial proposals,
Milliken again addressed the legislature on educational reform. In

April, 1971, Milliken spoke of the educational crises.

90

We are now in our second year of the battle for educational
reform. The crisis is still with us, and has grown worse.
Many school districts are bankrupt; others teeter on the brink
of financial disaster.8
In the context of his second message, Milliken outlined a number of
proposals for educational reform. Many were changed from the 1969
recommendations.
Although Milliken kept alive the objectives he had identified
in 1969, with the exception of aid to nonpublic schools, he made a
change in strategy. He now called for a constitutional amendment to
eliminate the property tax as the primary source of revenue for public
education. Thus, he abandoned his previous suggestion of a statewide
property tax, replacing it with a proposal to rely on the personal flat
rate income tax and a value added tax for businesses. The decision to
change from a state property tax to an income tax was outlined as
follows:
1. Frequently, because of taxpayers resistance, school
operating millages fail in elections, thereby deny-

ing needed support to the local educational system.

2. The property tax, while very stable, does not grow
as quickly as the economy or educational needs.

3. The varying property wealth of different districts,
regardless of the level of the property tax levied,
produces varying resources for education.

0. The property tax falls particularly heavily on senior
citizens, small farmers, and low income persons who are
buying homes.9

For these reasons, Milliken proposed a constitutional amendment that

would be equitable and adequate in funding education. This recommenda-

tion began a discussion of possible constitutional amendments, a

91

discussion that resulted in Proposal C being presented to the public
on November 7, 1972.

The 1971 message on educational reform also recommended a
variety of other reforms. Each local district could pass an “enrich-
ment” tax of up to six mills for additional educational programs. The
property tax would be equalized by the state so that a mill levied in
one district would raise the same amount of money as a mill levied in
another district. The concept of regionalization was dropped because
of the vocal and effective criticism of local and intermediate school
district superintendents. Their strong influence in the legislature
limited any serious effort to create regional districts with regional
superintendents appointed by the Governor. ‘Milliken added the recom-
mendation to approve a constitutional amendment to make the State Board
of Education appointed by the Governor, rather than elected by the
people. The Classroom Unit Plan of distributing state aid was dropped.
It was replaced with a set of criteria the legislature should use in
setting up a distribution formula.

In effect, Milliken's 1971 prOposals indicated a shift of em-
phasis from creating a total state system of public education to con-
centrating on a means of equalizing tax burden and school aid. Milliken
dropped many of the proposals that generated Opposition from people
interested in maintaining local school district boundaries and the
power of the local district. He was forced to concentrate on reforming
public education through economic reform. Milliken stressed the need

for finding a more equitable and adequate means of financing education.

92

In doing so, he shifted from a set of proposals based on both the econ-
omic model and social distance model to a one that was limited to the
economic model of public choice.

The emphasis placed on the economic model was shown by the
rationale Milliken used in proposing a constitutional amendment: tax-
payers' resistance to the property tax, elasticity of the property tax,
variation in tax burden, and distribution of property wealth. In the
1971 message Milliken attempted to meet the expected criticism that a
shift to state funding would mean a loss of local control.

Long standing tradition in the state requires that local commu-
nities and school districts retain control over important matters
of educational concern such as curriculum and personnel. The
difficult problem of raising educational revenue would be re-
moved from local districts, so they can concentrate on educa-
tional quality.
Thus, with the proposal for a constitutional amendment to reform school
finance, the Governor became aware of the conflict between the two
models of public choice and public policy. The 1971 proposals em-
phasized the economic model.

The Democrats' response to Milliken's two proposals (constitu-
tional amendment and a flat rate income tax) combined basic agreement
on goals with disagreement about means. Speaker Ryan stated that he
”would go along with Milliken's proposals if Milliken would consider
adding a graduated income tax to the constitutional amendment."ll
Since 1963, the Democratic Party in Michigan has supported deleting
the prohibition of the graduated income tax in the Michigan Constitu-

tion. Democrats in the legislature saw this as an excellent opportunity

to achieve that goal.

93

The issue in the legislature became one of whether to support
a flat rate or graduated income tax and whether there should be one or
two constitutional amendments--one to limit the property tax and the
other to permit a graduated income tax. The Democrats wanted to include
a specific rate of graduation within a property tax relief amendment.
Apparently, their hope was to have prOperty tax relief sell the gradu-
ated income tax to the public. The Republicans, on the other hand,
wanted separate amendments. Since constitutional amendments to allow
a graduated income tax had twice been defeated, Republicans did not
want to risk the success of property tax relief.‘2 Further, two amend-
ments would most likely mean passage of property tax relief but defeat
of the graduated income tax, thereby achieving the Republican goal of
retaining the flat rate income tax.

The constitutional amendment that received the most attention
in the legislature was House Joint Resolution (HJR) GG, which was
written by a coalition of Democrats and rural Republicans. It elimi-
nated all millage for operating expenses of local schools and allowed
13.25 mills for county, township, community colleges, and special edu-
cation programs. An additional 1.75 mills could be added by a vote
of the people in the local school district. The resolution also sought
to repeal the graduated income tax prohibition and limited the rate of
the value added tax to 2.5 percent.

Milliken, although previously opposed to a combined amendment,
decided to support HJR GG. The change was largely due to Democrat

Speaker Ryan's agreement to drop his previous demand of including

90

specific rates of graduation in the amendment. With this agreement,
Ryan and Milliken were successful in gaining passage of the resolution
in the Michigan House of Representatives.

When GG reached the Senate it was sent to the ”unfriendly”
Senate Judiciary Committee, where it died. The Michigan Senate, con-
trolled by Republicans, was very inhospitable to the graduated income
tax portion of the resolution. On September I, 1971, Senate Republican
Leader Robert Vanderlaan reported that the Senate Republican caucus had
voted not to bring GG out of committee. He commented that ”Republicans
would continue to work for meaningful tax reform but could not accept
the graduated income tax."'3

Resolution GG would have accomplished the same basic goal of
shifting the primary responsibility for funding public education from
the local property tax to a state income tax that Proposal C attempted
to accomplish. Consequently, the debate over resolution G0 is of con-
siderable interest in understanding the debate over Proposal C.

The debate over Resolution GG showed the predominance of the
economic model of public choice and public policy. Property tax re-
lief was the primary reason given by legislators for supporting the
resolution. Typical of the comments from proponents of HJR GG was
Republican Floor Leader Roy Spencer's statement:

I support this resolution because the Legislature of the State
of Michigan is willing to place before the electorate of the
state a proposal that will allow the people to make a decision

as to whether or not they should limit the property tax structure
of this state.1

95

Spencer, representing a primarily rural district, had long been an ad-
vocate of limiting the property tax. Other legislators, such as
Speaker Ryan, supported GG Because ”it is the only way we can get fis-
l
cal reform and education reform in the state.” S
Opponents of GG stressed their opposition to combining property
tax relief with the graduated income tax. Representative Crampton
summed up much of this opposition to G0:
Believing that property tax reduction to be our urgent priority,
I was prepared to vote for this imperfect means of achieving it.
Amending it to kill the prohibition against the imposing of
graduated income tax on the people of this state has killed any
chance of my supporting the resolution.'

Of the 13 legislators recording statements in the House of Representa-

tives Journal, only Representative Bryant objected to the resolution

 

because of the effect of shifting the primary responsibility for fund-
ing education to the state and thereby limiting local control.

I voted no on HJR GG because both aspects of it are a fraud

on the people. The property tax aspects are badly conceived

and would end local control. Amending it to eliminate the

prohibition of the graduated income tax opens unlimited state

government.1

The voting on GO in the House of Representatives followed party

lines, except for a few Republicans whose districts would have gained
from property tax relief (primarily rural districts). The problem of
supporting or opposing HJR GG was evidently interpreted by most legis-
lators as a question of the distribution of resources and tax burden.
Debate centered around the economic model of public choice. Little

debate was devoted to the question of the importance of the state

assuming greater control over educational finance. Thus, at the

96

legislative level, little importance was given to a social distance
model of public choice.

The second phase of the history of Proposal C started with the
death of HJR GG in the Michigan Senate. Its death made it clear to
both the Governor and the Democratic leaders that the legislature was
not likely to act on a constitutional amendment in the near future.
Consequently, both Milliken and the Democrats initiated plans to con-
duct petition drives to place the question on the ballot.

Governor Milliken's proposed constitutional amendment would
have abolished local millages for school operating costs, except for
a six-mill enrichment program, with retention of local millages for
construction, continuation of property taxes for county and township
operations, and reduction of total property levels by one-half. It
did not contain any specific statement about a graduated income tax.

In September, 1971, State Democratic Chairman James McNeely
announced a petition drive to get an amendment similar to HJR GG on the
ballot. The DemocratWSconstitutional amendment limited property taxes,
allowed a graduated income tax, specified the rates of graduation of
the income tax, and stipulated that renters as well as homeowners
would receive tax relief.

The Governor's petition drive was ”torpedoed” in early March,
1972, when the Mlchigan Supreme Court ruled that the Michigan Constitu-
tion did not limit the bonded indebtedness of Michigan political units.
In Butcher v. Grosse Isle the court ruled:

All governmental units with power to tax, including those speci-
fied in paragraph 2, clause 2, ”city, village, charter authority

97

or other authority,” as well as unchartered units, are not

limited, either as to rate or amount, as to tax imposed for

capital outlay expenditures or bonded indebtedness, which

is approved by the voters.18
The effect of the court's decision was that unless the proposed con-
stitutional amendment limited bonded indebtedness, a school district
could exceed the prescribed taxation level by using allowed millage
for operating purposes and using property taxes above the prescribed
levels for bonded indebtedness or capital outlay. This would mean
that districts would still vary on property taxation and tax burden.
The Governor's proposed amendment did not limit the bonded indebtedness
of local school districts or intermediate school districts. Conse-
quently, it would not have guaranteed equal property tax relief.

A further complication was encountered by the Governor's proposal
when State Senator Harry Demaso asked the Attorney General for a rul-
ing on one phrase of the amendment that inadvertently might have
allowed a graduated income tax. In view of these problems, the Governor
decided to withdraw the petitions. The Democratic Party had not yet
printed their petitions. Consequently, they were able to rewrite
their amendment to compensate for the Butcher decision.

With only the Democratic Petition drive alive, the Michigan Educa-
tion Association (MEA) endorsed the concept of two constitutional
amendments to appear on the November 7, 1972, ballot. The MEA's Board
of Directors had previously debated whether to support the Democrats'
proposal or to support the Governor's proposal, adding to it an addi-

tional amendment to droptimeprohibition against the graduated income

tax. Basically, the MEA felt the graduated income tax proposal might

98

defeat the prOperty tax relief amendment. The two amendments, later
to be known as Proposals C and D, (1) limited the use of the property
tax to no more than 26 mills, and (2) removed the prohibition against
the graduated income tax.

It is interesting to conjecture why the MEA put their consid-
erable organizational strength behind a petition drive to reform
school finance. Undoubtedly, part of the reason was a sincere commit-
ment to the goals of equity in school funding. However, a more com-
pelling reason might have been the need to find new sources for in-
creased expenditures for education. Millage defeats and local school
boards' subsequent refusal to give teachers additional economic bene-
fits added to the need to find a new source of funds. Shifting to the
state income tax offered the opportunity to find such new funds.

Two compelling reasons exist for teachers to support these
two proposals (C and 0).

One reason involves job security and working conditions.
Passing property tax millages becomes increasingly more difficult.
In the wake of millage defeats come teacher lay-offs and for the
teachers who remain on the staff, bad working conditions because
of reductions in programs, staff, supplies, and services. Even
affluent school districts are discovering that they are no
longer exempt from such problems.

Secondly, teachers are property-owning taxpayers or may be-
come property owners in the future and they have friends and
relatives who are property owners and who are threatened by a
tax that doesn't have to maintain any relation to their ability
or non-ability to pay.'9

Adoption of the amendment would also mean that funding levels would be
determined in the legislature rather than at the local level. Appar-

ently, the MEA was confident that they could be effective in securing

their demands from the legislature.

99

Proposal C, the property tax relief amendment, reduced the
maximum levy from 50 mills to 26 mills. The 26 mills were divided as
follows: 8 for counties, 1.5 for townships, 0.5 for intermediate
school districts, special compensatory or vocational education, 6
(voted) for county and township purposes, and 6 (voted) for school en-
richment. The enrichment millages were to be equalized so that one
mill levied in one district would raise the same number of dollars as
one mill levied in another district. The proposed amendment also lim-
ited the millage to be used for bonded indebtedness. This provision
overcame the obstacles presented in the Butcher decision.

The State Chamber of Commerce, Michigan Farm Bureau, and the
Michigan Association of Elementary School Principles quickly endorsed
the proposal and joined the petition drive. The Chamber of Commerce
and Farm Bureau were particularly interested in the tax relief aspects
of Proposal C. The Michigan Chamber of Commerce took the following
position:

The State Chamber of Commerce policy recommends elimination of
the property tax as a primary source of support for public
school operations. Revision of the property tax laws as they
relate to school operations is inevitable in view of court
rulings in other states and pending court action in Michigan.
This present proposal is the beginning of such necessary revis-
ion. Alternatives to which the public is being subjected are
much less desirable. The State Chamber has extended its sup-
port to the property tax amendment proposed by the M.E.A.20

The Michigan Association of Elementary Principles was primarily
interested in finding an equitable source of school funds. Their Board
of Directors took the position that full state funding was inevitable;

it was only a question of how the funds were to be raised.

The issue then before us is not whether or not the state
will fully fund public education; the courts are already making

('3
"a

L1

100

that decision. The issue we face is whether or not that collection

of funds will be based upon property taxes or some other means of

revenue colleCtion, logically some type of income tax.2'
By assuming that the state would be responsible for funding education,
the MAESP argued that ”local control” had nothing to do with Proposal C.

...The matter of local control, then, is not an issue in the

question of property tax relief. The sole issue is how will

our schools be funded, and the choice, ho efully, will be

made by voters in the November election.23
The MAESP position on Proposal C showed a handy division between the
two models of public choice and public Policy. The decision on whether
to support Proposal C had to be based on the economic model, since the
social distance model was not relevant because of current court cases.

The Michigan Association of School Administrators (MASA) and
the Michigan Association of School Boards (MASB) opposed the original
petition drive on grounds that the legislature should be responsible
for offering constitutional amendments. The MASA and MASB were concerned
about the possibility that both the Democratic and MEA amendments might
appear on the ballot, thereby confusing the voters. Once the petition
drive was certified as Proposal C and was set to appear on the ballot,
the MASB Board of Directors urged their membership to endorse the
proposal. However, a resolution endorsing the amendment was soundly
defeated at the MASB state convention the first week in November. The
debate on the floor of the convention showed the concern over the loss
of local control that adoption of PrOposal C would cause.
Following the MEA's announcement of a petition drive to place

two questions on the ballot, Governor Milliken made the following state-

ment:

101

The MEA's announcement today for a statewide coalition drive

represents a major breakthrough toward our mutual objective

of achieving not only property tax relief but also assuring

the quality and equality in the education of every child in

Michigan.23
With this statement the Governor teamed with the MEA and other organi-
zations in a coalition effort to place on the ballot the question of
property tax relief and shifting funding from the local district to the
state.

The process of getting enough signatures to place the proposals
on the ballot showed the debate between property tax relief (the econ-
omic model) and local control (the social distance model). The MEA
and the Governor were both pleased with the pubic's reaction to the
petition drive. However, the drive did indicate the MEA had to deal
with the issue of local control. The Exectuive Director of the MEA
stated on May 30, 1972, that:

Public acceptance of the petition drive has been ”absolutely
overwhelming.“ People are obviously sick and tired of fi-

nancing schools through the outdated and inequitable property
tax. This is certainly true of retired and elderly citizens.

An analysis of the public's reaction to the petition drive by the MEA's

Teacher's Voice showed three basic concerns expressed by citizens in

 

explaining their opposition to the petition.

Local Control. Many citizens argue that because the MEA
proposes to remove the local property tax as a basic means
of school support, it will destroy ”local control“ of schools.

Ability to control the local property tax. Taxpayers point
out that the local property tax is the one tax that must be
passed upon by themselves and hence is a tax over which they
exercise some control.

Property tax replaced by a more burdensome tax. Higher-income
families with modest property ownership argue correctly that

102

they are apt to pay more through an income tax than they would
through a property tax.

Of the three objections encountered by the MEA, two are directly re-
lated to the desire to maintain the role of the local district. The
social distance model is most useful in describing the impact of these
arguments on the vote on Proposal C.

Through the combined efforts of the MEA, the Governor's office,
and affiliated groups, enough signatures were obtained to place the
two proposals on the ballot in November. The Democratic Party's initia-
tive petition drive was unsuccessful. Consequently, on July 8 only the
two proposals made by the MEA, Proposals C and D, were certified by

the State Board of Canvassers to appear on the ballot.

Campaign

The campaign for and against Proposal C also showed how economic
and social distance variables can be used to interpret public choice
and public policy. The campaign in support of Proposal C emphasized
the economic aspects of property tax relief. In its support, the MEA
and the Governor organized campaigns to acquaint the voters with the
inequities built into the present method of funding education. They
presented their case in public appearances: television, radio, and
newspaper advertisements; and through state organizations. The only
organized opposition, at the state level, came from the Michigan AFL-
CIO. They made their opposition known by distributing information
about Proposal C to all their members. They did not conduct a general

campaign aimed at informing all voters in Michigan about Proposal C.

103

Their opposition to Proposal C was based primarily on an economic
argument stressing the uncertainty of what the changes would mean to
the average working man. They also pointed to the possible loss of
local control if the proposal was adopted. The impact of the statewide
campaigns was an emphasis on economic concerns. The voter was given
information about the economic aspects of property tax relief. He was
not as well informed, through a campaign effort, of the social distance
aspects of Porposal C.

The MEA spent over $250,000 in the campaign for both proposals.
The bulk of these funds was raised by a voluntary contribution of $0
by over 60,000 MEA members. The money was used to buy advertising time
on television and radio, publish brochures, advertise in the major
newspapers, gain the support of interested organizations, and staff
an additional office in Detroit.26

The Governor's assistance to Proposal C was evident in both
the petition drive and the campaign itself. During the petition drive
the Governor's office hired three individuals to coordinate the drive
in different regions of the state. The total budget was approximately
$18,000, which was raised from contributions from individuals person-
ally supporting the Governor. The Governor and his staff were success-
ful in obtaining nearly 80,000 of the required 216,000 signatures; they
mere obtained primarily from Republican party members. The staff per-
sonnel worked closely with party members in soliciting volunteers to
distribute petitions and solicit signatures.27

The campaign for Proposal C, once it was certified, was less

well financed by the Governor's office. The major effort consisted in

100

speaking for the Proposal whenever the Governor or any of his staff
made a speech in Michigan. The Governor's press section also solicited
the editorial endorsements of Proposal C from major newspapers in the
state.

The only organized effort against Proposal C at the state level
came from the Michigan AFL-CIO. Three weeks before the election the
AFL-CIO sent a news release to all affiliates and newspapers in Michi-
gan urging members to work for the defeat of Proposal C. It is impos-
sible to test the effectiveness of this last-minute campaign. However,
it undoubtedly raised a number of questions in people's minds about
the effectiveness of the amendment.

The issues outlined in the AFL-CIO news release give an indica-
tion of the major reasons used by people opposing Proposal C. In con-
trast to the MEA's analysis, the AFL-CIO concentrated on economic
reasons why they opposed Proposal C. The news release outlined five
major reasons for opposing C:

1. Proposal C would grant $500 million tax break to business
by eliminating property taxes on business property.

2. Proposal C would force the legislature into increasing the
flat rate income tax to at least 7 percent just to replace the lost
revenue, thus placing an intolerable burden on the workers and their
families.

3. Proposal C, by prohibiting a state-wide property tax on
business would force the legislature into the enactment of a so-called
value added tax which is actually a hidden sales tax falling hardest

on those least able to pay.

105

0. Proposal C, while maintaining silence on the issue of
local control, leaves this important area open to the interpretation
by the courts.

5. Proposal C can only be labeled irresponsible because it
fails to guarantee adequate state financing while eliminating the
local tax base.2
The issue of the loss of I'local control'I if Proposal C had passed was
a major point of contention with other opponents of PrOposal C. Dr.
Michael Deeb, a Democratic member of the State Board of Education, ob-
jected to the Board's support of Proposal C because:

There has been virtually no public discussion as to what this

method would do to the local control of schools. It is all

part of a compromise to deprive the people of their check on

how well schools are run in the local area.2
Deeb's position concurred with the Democratic Party's resolution urg-
ing the defeat of Proposal C. The Democratic Party opposed the pro-
posal on grounds that it did not guarantee tax relief for renters or
local control of school finance.

The debate over Proposal C was reflected in the editorials of
major newspapers throughout Michigan. Their discussion shows the pre-
dominance of economic arguments used in gaining support for the pro-
posal.

In Detroit, the two major newspapers split over the support of

Proposal C. The Detroit Free Press called for its support because:

 

...The overriding appeal of Proposal C is that it would offer
a replacement for a millage system that has already broken down.
In short, it is a means of keeping the schools open.30

106

The Detroit News wrote against Proposal C because “it cannot deliver

 

property tax relief as it claims.“ Further, the Ngws_referred to the

fear of some educators that Proposal C would ”significantly decrease

local control over education.” Consequently, the Ngw§_wrote that Pro-

posal C contained ”clear and present dangers to the financing and ad-

ministration of public schools in Michigan. It should be defeated."3'
In the other large newspapers in Michigan, the reaction was

mostly favorable towards adoption of Proposal C. The Muskegon Chroni-

 

cal supported the amendment:
Clearly, approval of the amendment will bring a measure of
property tax relief and, we are convinced, enough meaningful

reform to warrant voter support.32

The Saginaw News justified its support of Proposal C as being

 

...based on the realization that it offers the opportunity
to end the incessant and acrimonious local school operating
millage votes. That it represents truly meaningful property
tax relief for legions of older, retired persons living on
fixed incomes. That it guarantees a better standard of liv-
ing in 60% of the state's hardest pressed school districts
...and that it is a fairer means of financing public school
operations and providing a vehicle for equity in educational
opportunity for all children in Michigan.3

The Lansing State Journal urged voters to vote yes on Proposal C be-

 

cause it remedied the heavy tax burden on property owners for support
of local schools and provided ”a uniform method of distributing school

revenues so that all schools get an even break.”3h The Jackson Citi-

 

zen Patriot said that whether or not people vote for Proposal C

 

depended on the degree of their confidence in the legislature:

If voters have enough faith in the Legislature to write a tax
program that will sufficiently recover the lost revenue from
property taxes while funding education equitably then they
should vote ”yes” on Proposal C....We trust the lawmakers in
Lansing; we'll vote yes on PrOposal C.3S

107

Other newspapers throughout the state made the same types of argu-
ments in favor of Proposal C. A minority of newspapers opposed the
amendment because of the loss of local control. The conservative Grand

Rapids Interpreter wrote:

 

The voters of Michigan should wake up before it is too late,
and vote ”No” with an overwhelming tide of emphasis, to let
these schemers (the Governor and the MEA) know that we are not
yet ready to allow them to take our home, or other property, in
the misguided belief that they are doing it for our children's
education. Most of them couldn't care less about the education
of children, what they want is complete control of the schools,
and your property.3

In summary, most of the larger newspapers in the state spoke in favor
of Proposal C, basing their support on the economic model. The Detroit

News and Grand Rapids Interpreter opposed Proposal C because of the

 

implications of losing local control.

The debate over Proposal C was clouded by the issue of school
integration and “busing.” The busing issue tended to dominate the
political campaigns in Michigan in 1972.37

In September, 1971, Federal District Court Judge Stephen Roth
found the State of Michigan guilty of de jure segregation of the Detroit
Public Schools. Roth ordered the Michigan State Board of Education to
prepare different proposals to combat segregation. The proposal ac-
cepted by Roth as best meeting the goal of integrating the schools
required cross-district busing among 53 school districts in the Detroit
area. In effect, this threatened the local integrity of most of the
suburban districts surrounding Detroit. Roth's decision was appealed
to the United States Supreme Court by Attorney General Frank Kelley

and Governor William Milliken. Final determination of the case is

Still pending at the time of this writing.

108

Roth's order triggered the highly emotional and divisive de-
bate over the merits or demerits of “busing” to attain racial balance
and equal educational opportunity. Eventually, the issue of busing
would complicate greatly the issue of financial reform. Speaking of
the impact of the Roth decision, Milliken said:

Even the suggestion of cross district busing to achieve racial
integration has stirred a storm of emotions and uncertainty
that threatens to tear the very fabric of our whole public edu-

cation system.3

The Macomb Daily newspaper agreed with Milliken's assessment of the

 

impact of the Roth decision. In analyzing the election returns follow-
ing the November election, the Daily wrote:

No assessment of last Tuesday's election can ignore how exten-
sive was the scope of the cross busing issue in the Michigan

balloting.

Its range was not confined to southeastern Michigan as
many political leaders expected. It cut a wide statewide
swath.39

State Republican Chairman William McLaughlin attributed the

defeat of Proposal C to the impact of busing and the threat it posed
to local control.

...The defeat of the proposition to cut down sharply on property

taxes to finance schools and substitute higher income taxes,

“smacked of local control. Anything that stood in the way of

antibusing Tuesday got run over.”
The issue of busing and school integration was also very real in Lansing,
Saginaw, Grand Rapids, and Jackson. All four areas either had already
confronted the issue (Saginaw), or had desegration plans under con-

sideration or in operation because of pending lawsuits forcing inte-

gration of the schools.

109

To many voters in Michigan, any issue related to education was
directly related to the issue of busing. Although the issue was never
publicly brought into the debate over Proposal C, both the Governor's

01

office and the MEA felt busing was ”behind the scenes.” One week
prior to the election, the Governor held a staff meeting to discuss
whether or not he should try to separate the issue of busing from the
issue of school finance reform. The Governor and his staff decided
not to raise the issue in the public's mind. The decision was based
on a public opinion poll showing that Proposal C would pass and that
the issue of busing was not related to the public's attitude toward
school finance reform. In retrospect, this has been seen as a major
error.

In summary, proponents of Proposal C argued that its passage
would mean property tax relief, equitable funding of public schools,
and a method of easing the financial crises facing public education.
Opponents argued that the economic effects of Proposal C were not known.
Its passage would probably mean a substantial increase in the income
tax. They also suggested that adoption of the amendment would reduce
local control. The issue of busing was not used by either side, al-
though it was felt to be a major factor in the election outcome. It
is difficult to assess the amount and effectiveness of information
available to the voter. The campaigns in favor of Proposal C attempted
both to stress economic aspects and to reduce fear over loss of local
control. Basically, the debate over Proposal C and the conjecture

about the vote reduced to the question of whether or not property tax

llO

relief would sell the proposal. Would property tax relief overcome the
desire to maintain the status quo and the present degree of local con-

trol?

Election
The final vote on Proposal C came as a surprise to many state
policy makers. The Proposal was defeated by a margin of 1.3 million
to 1.8 million votes. Their surprise was undoubtedly a result of polls

conducted by Market Opinion Research of Detroit for the Detroit News

 

showing Proposal C passing by a 60 to 00 percent margin.
The polls conducted prior to the election had shown a steady
trend toward increasing support of Proposal C. Table 5.1 shows this

trend.

TABLE 5.l.--Public Support for PrOposal C in the Detroit Metropolitan

 

 

Area.
Poll 1 Poll 2 Poll 3
(9/1/72) (9/21/72) (lO/ll/72)
% Yes 58 60 65
% No 33 30 29
% Undecided 9 IO 6
Total 100 100 100

 

Source: The Detroit News, September 12, 1972, October 3, 1972, and
October 13, 1972.

 

N = 050.

111

In the October poll the research firm tried to determine if there was
a relationship between support of Proposal C and attitudes toward bus-

ing. The results of the poll are presented in Table 5.2.

TABLE 5.2.--Support of Proposal C and Attitudes Toward Busing.

 

Position on Proposal C

 

 

Yes No Undecided Total
% Opposing Busing 59 32 9 100
% Supporting Busing 67 20 13 . 100
Total Sample ' 6O 3O 10 100

 

Source: Detroit News, October 3, 1972, p. 10A.

 

As can be seen, Market Opinion Research found that the effect of busing
attitudes on support of Proposal C was rather small.

The Michigan Education AssociatiOn also contracted with Market
Opinion Research to conduct a public opinion poll on attitudes toward
Proposal C. The MEA-sponsored poll reached the same basic conclusion
as did the Detroit News poll: Proposal C would pass. One of the more
interesting questions asked by the MEA poll was why voters who were
against Proposal C objected to it. Table 5.3 lists a breakdown of the

major reasons people gave for opposing Proposal C.

112

TABLE 5.3.--Reasons Given for Opposing Proposal C.

 

 

Reason Percent of Those Against
Will increase income taxes 10
Taxes are already too high 13
Lose local control 15
State control/no limitation 6
Will still increase property tax 17
Poor will get hurt 7
People with no children still pay tax 0
Other 20
Total 100

 

Source: Report of Market Opinion Research to Michigan Education Asso-
ciation September, 1972.

Table 5.3 indicates that: (l) 27 percent were concerned with higher
taxes because of the limitation of the property tax, (2) 21 percent
were concerned with the loss of local control or future state domina-
tion, and (3) 17 percent were confused about whether it would really
limit property taxes.

The polls taken for the Detroit News and the MEA gave the

 

Governor and other political leaders a false sense of security about
the passage of Proposal C. The questions about the impact of busing
and the loss of local control were not seen to be major obstacles to
the proposal's success.

It is difficult to determine why the polls were so incorrect
about the passage of Proposal C. Two explanations seem plausible.

First, the last poll was taken approximately a month before the election.

113

The campaign against the proposal, conducted by the Michigan AFL-CIO,
Tad not yet taken place. There may have been a major shift of opinion
during that month. The second explanation is that the sample used by
Market Opinion Research was not representative of the voting population.
In either case, the polls misled policy makers into thinking Proposal
C had an excellent chance of passing and that the issue of busing or
local control did not need to be separated from the question of school
finance.

Proposal C was favored in 26 of Michigan's 83 counties. Appen-
dix B provides a county-by-county breakdown of the vote. An analysis
of the vote by county shows that Proposal C was viewed more favorably

in rural than urban counties. Table 5.0 shows this distinction.

TABLE 5.0.--Vote on Proposal C by Percentage of Rural Population in
Michigan's 83 Counties.

 

l 2' .
«M
flag» '.

 

 

0 - 25 2 Rural 26 - 50 % Rural 51 - 75 % Rural 76 - 100 % Rural

IT

Yes 790,808 187,630 218,815 133,011
No 1,219,606 183,071 169,265 102,505
8 Yes 39 z 51 z 05 % 08 8

N = 8 I3 28 33

I;

Source: Federal Census for 1970, State of Michigan and Michigan Secre-
tary of State.

Proposal C was defeated primarily because of the vote in Detroit and
Grand Rapids. The difference between ”no” vOtes and ”yes” votes in

those two areas accounted for 050,000 of the 090,000 votes separating

 

 

110

victory from defeat. The defeat of Proposal C was primarily an urban
phenomenon. Consequently, it is analyzed by comparing the two models

of public choice and public polity in the urban areas of Michigan.

Summary

The history, campaign, and voting behavior associated with Pro-

posal C were traced in this chapter. The review has shown the competi-

--p .E I
’.
‘

'1'

tion between the economic and social distance models and how the two
perspectives influenced the discussion of reforming educational fi-

nance. The next chapter explores which model best explains the vote

 

TELL-(3 anfal' 1:‘-.'AA._H
1_ v V '
‘1- al.. ..

on Proposal C.

 

CHAPTER V--FO0TNOTES

lLaurence lannaccone, Politics in Education (New York: The
Center for Applied Research, 1967). Stephen Bailey, et. al.,
Schoolmen and Politics: A Study of State Aid to Education in the
Northeast (Syracuse: Syracuse University Press, 1962). Nicholas A.
Masters, et. al., State Politics and the Public Schools: An Explana-
tory Analysis (New York: Alfred A. Knopf, 1960). Michael Usdan,
The Political Power of Education in New York State (New York: The
Institute of Administrative Research - Teachers College, Columbia
University, 1963); Michael Kirst (ed.), The Politics of Education
at the State and Federal Level (Berkeley: McCutchan Publishing
Corporation, 1972); and Dean B. Bowles, ”The Power Structure in
State Education Politics,“ Phi Delta Kappan, February, 1968, pp.
337'300-

 

 

 

 

 

 

 

 

2 . .
Bailey, Op. Cit.

3See Masters, et. al., op. cit., especially chapter six on
Michigan.

“Masters, op. cit., p. 267.
5Frank Pinner, et. al., Education and the State: Decision Mak-
ing on the Reform of Educational Finances in Michigan, a Report to the

Urban Coalition, Washington D. C., 1971. Chapter 5.

6J. Alan Thomas, School Finance and Educational Opportunity in
Michigan (Lansing, Michigan: Michigan Department of Education, 1968).

7George Montgomery, William Ryan, and William Copeland, Gongwer
News, April 5, I969.

 

 

 

 

8William G. Milliken, Special Message to the Legislature on
Excellence in Education, April 12, 1971, p. 2.

 

 

9Ibid., p. 2.

 

ll

 

'2Constitional amendments to allow a graduated income tax had
been defeated in 1966 and 1968 by about a three to one margin.

115

Speaker of the House William Ryan, Gongwer News, April 15, 1971.

.o

 

 

116

13Senate Republican Leader Robert Vanderlaan, Gongwer News,

September 1, 1971.
Representative Roy Spencer, Journal of the Michigan House of

Reparesentatives, No. 108, July 29, 1971, p. 2360.
5House Speaker William Ryan, Journal of the Michigan House
of Representatives, No. 108, July 29, 1971, p. 2363.
Representative Louis Crampton, Journal of the Michigan House
c>f' Ftepresentatives, No. 108, July 29, 1971, p. 2365.
17Representative Bryant, Journal of the Michigan House of Repre-
sen tatives, No. 108, July 29, 1961, p. 2369.

'8Robert E. Butcher and Gladys K. Butcher v The Township of
Grosse Ile, State of Michigan Supreme Court, No. 52935, June, 1971,

P- 13.

 

'9Michigan Education Association, ”Tax Proposal Answers,”
Teachers Voice, Vol. 50, No. 1, September 0, 1972, p. 2.
0Michigan Chamber of Commerce, Special Report, April, 1972,

P- ls-

2'Ed Keller, ”Property Tax Relief and Local Control: Separable

ues," The Michigan Elementary Principal, Vol. 07, No. 0, Summer,

 

'SS
‘972,m0.

22lbid.

23Governor William G. Milliken, Press Release, March 18, 1972.

t - Terry Herndon (Executive Secretary Michigan Education Associa-
Ion) Gongwer News, May 30, 1972.
25Michigan Education Association, Op. cit., p. 2.

Interview with Kai Erickson, Director of Public Relations,

M‘
. (:lf‘ igan Education Association, February 9, 1973.

27Interview with Dick Pastula, Special Assistant to the Governor,

F‘
Qt) I‘uary 9, 1973.

28Michigan State AFL-CIO, Press Release, October 17, 1972.

29Michael Deeb, Gongwer News, October 11, 1972.

30Detroit Free Press, ”C and 0 Would bring Equity in School

1:.
I
I ‘teance,” Detroit Free Press, November 5, 1972.

 

 

 

117

3'Editors, “No on Proposal C,” Detroit News, October 0, 1972.

 

32Editors, “Problems Remain, but Vote Yes on Proposal C,“
fihJskegon Chronicle, October 26, 1972.

 

33Editors, ”Proposal C Presents Tax Reform Opportunity,”

Saginaw News, November 2, 1972.

3“Editors, ”Yes on C but a No for D,” Lansing State Journal,
October 31, I972-

 

35Editors, ”Legislature Key to Proposal C,“ Jackson Citizen
Pat r'iot, October 25, 1972.

36Editors, “Clever Selling Job May Backfire on Proposal C at
Ba I I ot Box," Grand Rapids Interpreter, October 19, I972.

 

 

37tditors, ”Busing Cuts Big Swath in Statewide Balloting,”
flacohwb Daily (Mt. Clemens, Michigan), November 18, 1972, p. 6-A.

38william G. Milliken, Gongwer News, February 10, 1973.

 

 

39"Busing Cuts Big Swath in Statewide Balloting,“ Macomb Daily,
NOVenlber 13, 1972, p. 6-A. ‘

Ibid.

 

h'lnterviews with Dick Pastula and Kai Erickson, op. cit.

 

 

CHAPTER VI

ANALYSIS OF DATA

Analysis of the vote on Proposal C shows that neither the econ-

 

om i c: nor the social distance explanation, alone, was capable of describ-

ing the voting behavior of the Michigan electorate. However, when both ‘3‘

exp I anations are used and the vote is analyzed according to metropolitan

 

reg ions, a fairly comprehensive picture can be drawn of what factors .;
in F 1 uenced the vote on Proposal C.

This chapter analyzes the vote on Proposal C in four ways.
Fl r“ s t, zero-order and partial correlation matrices are presented show-
'n9 a variety of demographic variables and how they correlate with the
Vote on Proposal C. From these matrices, it is clear that two explana-

t 3 0715 of public choice and public policy are relevant to an analysis

OF the vote. Secondly, multiple regression analysis and multiple

[)3 "tial correlations are used to compare the two explanations of public
Cho 3 Ge and public policy within each of the seven metropolitan areas.
Th 3 "(1, different hypotheses are discussed and analyzed in an effort to
exp 1 ain why the metropolitan regions differed in the "explanatory"
Dow‘ar of the two sets of variables. The fourth section analyzes
WHQ ther or not the concept of social distance is as relevant for social
Cl 3 f: Fer-ences within school districts as among SChOO' districts. In

th
Q last section, some of the major findings of the research are re-

\

Q\rved.
118

 

119

The Over-All Picture

 

The correlations between demographic characteristics in the
183 school districts and the vote on Proposal C are shown in Table 6.1..
Three conclusions may be drawn from the correlation matrix. First,
support of Proposal C was negatively related to the wealth of the com-

munity and the school district. Since school district expenditures

are basically a function of community wealth, it is not surprising to F3 1

find school district variables and general wealth variables similarly
correlated with the vote on Proposal C. Support for Proposal C was

correlated with family income (-.36), value of the home (-.03), and

 

"5"
"

average rent (-.28), to about the same extent as with instructional
expenditures (-.28), average teacher salaries (-.00), and state equali-
zed valuation (-.29). The precentage of the total budget received

from state aid was positively correlated with the vote on Proposal C
(.31) because the percentage of funds received from the State is
determined by the wealth of the district.

The second conclusion to be drawn from Table 6.1 is that home
ownership and property tax rate were negatively associated with support
of Proposal C (-.26 and -.28, respectively). Although proponents of
Proposal C argued that its passage would mean property tax relief for
the homeowner, it is apparent that home ownership and high millage
levels were associated with the desire to maintain the current system.

The third and final conclusion is that districts with a large
percentage of people age 65 or over tended to support Proposal C more

than communities with a smaller percentage (+.27). Possibly, tax

In”, . -
-.-n—m.

 

120

 

fill:

trillFltbn’

 

 

 

xx Nm. om. m_. on. cm. m..- -.l ca. mm. mm. mm. _:.1 om.i oeooc_ ommuo><
xx mm. ~_. oo. mm. mm.1 mm.1 :m. :N. _:. mm. _:.i m:.1 acom ommuo><
xx o_. 0.. ow. mm. NN.1 ma. mm. :m. mm. m:.1 m:.1 o:_m> OEOI cm_poz
xx m_.1 _N. on.: mm. mo.l mm.1 m_.1 mm.1 mm. w~.l mcoczooeo: uc00coa
xx xm. N_. mo.- Nm. mm. _m. .m. ma.- o_.- o_omx eo_umeauuo
xx mm.1 mm.1 Na. mo.l mo. mo. N_.1 m..- o_umm co_umo:ow
xx Om. __.l mm. __. mm. m_.1 mm. mm Lo>o HCOOLOQ
xx mo. om. mm. Rm. m_.1 w_. xom_m ugOOuom
xx mo. mm. am. mm.- oN.- oomx oom___z
xx am. mm. .m.1 m~.1 co_um:_m>
cON__m:om Oumum.
xx mo. xm.- aa.- xcm_mm
cucumOH ommco><
xx mm.1 m~.1 OmCOQXw .mco_uo:uumc_
xx _m. u_< Oumum uCOOLOd
xx 0 mc_co>mu ommucooLom
IV ro H 9o 0 3 o/o .70 N 5 31V 31 o/o 06
u A a A o O p. 1. a. e a A x u
3 a u a w u” 3 n no as I. A” 1.9 a .o s Q. .A
OJ IJ 9 O n D A I. l EDJ 81 .4 a
we 9 m d E 3 9 D. Jule UJ 9 S
95 5 A a 9 1. J D 5 A85 Sh ..4
a a e o J 1. x. a J a a O a o
l. M I. O 9 1+ U
n U 0 U S I. V
a a U 0 I... 3
w u p.

 

.muu_cum_o .OOLOm mm. c_ u .mmOQOLm mo
co_uaoo< m:_co>mu ommuc00coa ocu new mo_nm_cm> o_;amcmoEoo m. coozuom mco_um_0cc0u cOpcOuOuONil._.m mom<p

lZl

relief was an important factor in influencing Opinion in that segment
of the community. It may also be spurious because of the relationship
between the more highly urbanized districts, the high percentage of
people age 65 or over, and support of Proposal C.

In addition to school districts in the seven metropolitan areas,
correlations were computed for the census tracts within the central
cities of six of the areas, and for subcommunities in Detroit. The
correlation matrices are presented in Tables 6.2 and 6.3.

Table 6.2 gives the correlations between l3 demographic varia-
bles in I82 census tracts in six metropolitan central cities (excluding
Detroit). Only two variables were significantly related to the vote
on Proposal C. The percent black was positively correlated with
support of the amendment (+.55). The percentage of people buying or
owning their own home was negatively correlated with such support
(-.h3). The difference in the two correlations is interesting, in
that it suggests social characteristics rather than economics may ex-
plain the vote within school districts as well as among school dis-
tricts. If property tax relief had been important, the percentage of
people owning or buying their homes should have been positively corre-
lated with the vote on the proposal. However, this was not the case.

The same conclusion can be drawn from Tables 6.3 and 6.4.
Table 6.3 shows the correlations among lh demographic variables in the
#9 subcommunities of Detroit. The percentage nonwhite was highly
correlated with the vote on Proposal C (+.87). The relationship

between home ownership and support of the amendment was similar to the

l22

 

 

 

 

ELF. .r .I. uh: . Pi... ..FII...I..I.VIIWMI:
._o>m_ .0. am ucmu_m_cm_m m_ cm. ”ouoz
xx wo.u :0. mo. mo.u m_. 00. o_. N..- NN. mumm xmk
xx om. xx. mm. ma. mm. m..- mm.- m..- 6500:. mmmtm><
xx om. .m. o:. mm. xm.- x~.- mo.- acme mmmtm><
xx o_. :m. cm. ~m.n mN.I No.u o:_m> oEo: cm_omz
xx o_. om. NN.I .m.u mq.u mcoczomeo: ucmocmm
xx 00. _N. mm.- No. o_um¢ co_uma:uuo
xx _~.. 0... w..- o_umm co_umo:om
xx Nm.n m..- mo Lo>o ucmuLmd
xx mm. xum_m achLom
xx u mc_co>mu mmmucoocom
me... W. % nuwu n70 «nJU m xw/o m/o n70
X 3 U m 0 3 n 0 9 I...
O 1 a M n 3 A .l O
uo m U .0 P. 8 9 J
9 a .A I. e 1 J 3
1+ 9 U 1+ I . v4 3
a I. .b I. 0 Q;
n O u S
a H” u
o
w
a

 

.mo_u_u .mcucou x_m c_ muomch mnmcou mm. c_ u .mmoa0cm mo co_uaoo< mc_co>mm
ommucouLom ozu ucm mo_nm_cm> u_comcmosoo mc_z coozuom mco_um_occ0u concoIOLoNII.~.e mom<h

 

 

 

 

 

123

xw. _N.- m~.- o..- No. No. Rm. mm. xN. :o. 00.- m..- xm.- mucmo_mmx co Lumcmu
x mm.. .m.- m:.. 0:. __. Na. ma. 0:. as. mm. :x. No.- _oo£um _m_£uotma c_ cuta__cu x
x mo. mm.- 0... mo. mm. mm. mN. mo. ax.- mm. m_. m_ cans: mu_x zu_z mm___Emu x

x mm.. mq.- m..- om.- :m.- mq.- .m. xN.- me. me. amxo_aemc: x

x .m. Rm. N. mm. ma. N:.- am. .¢.- 3”,- tm__0o mu_£3 x

x we. 94. x4. mo. am.- mm. mq.- mm.- nsumom u_a0couu-omuom cm_I

x mm. oz. mm. w... No. 00.- oo. co_umusom

mmEOI >__Emd v_mc_m w

l

x mm. _w. m..- mo. mq.- mo.

 

 

 

x mm. .N.1 mo. N:.1 no.1 meo: mc_>:m co mc_czo N
x NN.- :0. mm.- m:.- 9:8:_ cm_amz
x 051 m:. .3. m_ Lone: N
x :m.- 44.- mo Lo>o N
x 2.. 82382 .x
x u :0 moo> mo> w
.I. S d 7. VA ...zm mm 7. S 3 H .11 H 1 OI... Z. w . ./...
a 3 p. I. 1 3 I. P o. n a
u 11.1 I. 3 D. 4.. n M e 0 6 n .1. S .A O D. H O N .A
b 0 o u- S e u u. l u u. 3 a I. 1. M 1. u A o a
J O 3 I. m a 1. n O p. S U u u 9 D. a u S
q I. U. II. n l. m \4 Q. W n5 1 5 5 l. U 3. J M.
1. D. u I. d a 1. n. 1. l J J u. A
o a J 0.1. 1. 3 3 o 9 Hrs I Q, 1. o mo_nm_cm)
1. l a a a 0 3 1. u o u l S l 1 . .
u J s 1... o 0 d m C 3 00 a 9
Nu 3 .l . P .0 J O
a 1. I. M P .l m m 3
S U 8|. 9 l. a
. ‘0 J l
u. .A

.u_ocumo mo mo_u_c:650on:m m: sot» mo_nm_cm> oouoo_om to» mco_um_mcc0u cmoLOIOLoN11.m.m mom<p

12h

 

 

 

 

+ .
F .. . III 5 it. 1“!.I,I{IC.._-.... L“
:w. mm.1 :m.1 mm.1 _o. _0. mm. mm. mm. m_. _~.1 x mm.1 oucoo_mom we Lumcoo
x .0. mo.1 NN. NN. mo. m_. :m. :N. m_.1 m_. x o_.1 m_ooLom _m_;o0cma c_ coco__£u w
x NN.1 o_.- :0. N_. mm. m:. ma. mm. mm.1 x _N.1 m_ Lona: mo_x cu_3 mm___Emu N
x mo. :m.- o~.- mm.- mm.- mm. m..- mm. x .m. amxo_aEmc: x
x mm. mm 00. mo. mm. m.1 mm. x mm. Lm__0u oo_£3 N
x _m. "m. mm. Rm. _x.1 mo 1 x q_. magnum u_eo:0om-o_o0m cm_I
x mm. _a. mm. m..1 No.1 x __. co_umo:om
x mm. xx. mo. xm.- x _Q.- maze: x__e~u m_ac_m x
x mm. _o.1 :m.1 x om.1 oeox c_ocu mc_>3m co mc_czo w
x o_.1 mm 1 x mm.1 mE0oc_ cm_ooz
x mm.1 x 00. w. cove: N
x x mm. mm to>o N
x x m._53coz x
x o co muo> mm> N
..l C., 10 .7. fin ../ ./ x. S 1.. H 13 H .7. 8 t N a/ n/ / /.
.0 \J 9 l. 19 3 I. D..\I1 n a
...t Mum 014. n m 00b n wt. A0 or n O N .A
.b K H l C m U l 14 U U. «J O .I. l. M l. U fl. 0 3..
1.. I l. .11. O l. n! O P. S U U U 9 nit 3.. U r3
u l: U1 H n. l. W 1.1 C. WI... aka 10 6 6 l. U J .J M
o -. m .1 m .1. m mm m. a Hm 1 . 9 w m $225..
I.) 11. ...n 9 «v 0 fiJ I. U 0 U I. C... 14 1|
11 J 5 ,A O O .3 w 0 D 00 a. a
H. a l. . P. a 1.. O
a H l M P l w w 3
C) L 8 .l. mu Ill” 9
.M . J. 111.11., 1111

I'
ll
.1
‘1
l
‘I

 

 

.coe oo__ocucou ou_c3coz

ommucmucom zu_; u_ocooo mo mo_u_c:EEouo3m a: 60c» mm_nm_cm> oouuo_om to» mco_om_occ0u _m_ucmm11.:.m mom<h

 

125

other six areas, with correlations of -.6h for the percentage owning

or buying their own home and -.68 for the percentage of single-family
dwellings. Because of the high correlations between support of Pro-
posal C and percentage nonwhite, partial correlations were computed for
12 variables while controlling for percentage nonwhite. Controlling
for the percentage nonwhite (Table 6.“), did not significantly reduce
the correlation among the percentage owning or buying their own home,
the percentage of single-family dwellings, and the vote on Proposal C.
However, other relationships that were significant before were no
longer significant at the .0] level. Only the percentage of white‘
collar workers in a subcommunity was significantly correlated (+.38)
with the vote on Proposal C. This may substantiate the theories of
some urbanists that “cosmopolitans” support consolidation or integration
more than ”locals” do. However, this was the only instance of a
positive relationship between the percentage of white-collar or pro-
fessional employees and the percentage voting for PrOposal C.

All four matrices show the vote on Proposal C was more strongly
related to the social characteristics of the school district, census
tract, or subcommunity than to home ownership or tax rate. Thus, the
mate on Proposal C may best be explained by social differences rather

than property tax relief for homeowners.

Comparisons of Two Explanations
A second purpose of this chapter is to compare two such explana-
tions of public choice and public policy. Chapter II gave the theoreti-

cal rationale for each explanation. The social distance explanation

126

suggests that as social distance between school districts or census
tracts increases, support for Proposal C will decline. Social distance
was measured on three dimensions: social rank, segregation, and life
style. The economic explanation hypothesizes that the vote on Proposal
C should be explained best by the economics of property tax relief.
Three variables are used to measure the current burden of property
taxes in the school district: tax effort, tax burden on homeowners,
and tax burden on income.

The first step in comparing the two explanations was to compute

the multiple regression equation for each set of variables for all 183

 

school districts. The variables were standardized by using the mean
and standard deviation for all 183 districts. Table 6.5 gives the
multiple regression coefficients, zero-order correlation coefficients,
and Beta scores for each model.

As can be seen, the regression equations, together, account
for only 7 percent of the variance (R2). Two explanations come to mind.
Either the independent variables selected have scarcely any effect on
the vote, or else such relationships as exist vary in strength and
direction over metropolitan areas, so that they become obscured when
the data for all seven areas are combined.

To test these possiblities, multiple regression equations and
multiple partial coefficients were computed for each metropolitan area.
The social distance variables were standardized by using the mean and
the standard deviation of each metropolitan area, to insure compara-

bility among metropolitan areas. This is a different approach than

127

TABLE 6.S.--Multip1e Coefficients of Determination, Simple Correlation
Coefficients, and Beta Scores for 183 School Districts on
Two Explanations of Public Choice and Public Policy.

 

 

 

 

 

R r B
. a
Econom1c Model
Tax Effort .01 -.11 -.12
Tax Burden on Homeowners .00 -.01 -.02
Tax Burden on Income .00 .00 -.02
Total .01
Social Distance Modelb
Segregation .03 .18 .13
Social Rank .01 .15 .16
Life Style .02 -.13 -.15
Total .06
8F = .811
bF = h.25
N = 183

the one used by Shevky and Bell in their original analysis.l The re-
searchers standardized their social distance variables by using the
range of the variable over all metropolitan areas. Standardizing by
metropolitan area, rather than by all seven areas, is based on the
assumption that social distance is relative and that individuals com-
pare their social status with surrounding communities rather than by
some absolute or statewide standard.

Before computing the regression equations, correlations between
independent variables in each explanation for all seven metropolitan
regions were calculated. This was done to test for multicollinearity

between independent variables. The intercorrelations are given in

128

Table 6.6. A number of correlations present a problem of collinearity.
In Lansing and Jackson the correlation between segregation and tax
effort is above .802 (.92 in Lansing and .94 in Jackson). However,
since they are in different models the multiple partial correlations
will control for their interaction.

In four cases there were serious problems of multicollinearity.
The correlation between tax burden on homeowners and tax burden on in-
come was -.79 in Grand Rapids, -.88 in Jackson, -.79 in Flint, and
-.8A in Muskegon. To compensate for the high correlations, tax effort
and the more powerful of the other two economic variables (tax burden
on homeowners and tax burden on income ) were used to calculate the
multiple regression equations for the economic explanation in the
four metropolitan areas. The following formula was used to adjust
for the unequal number of independent variables and degrees of free-
dom.

-2 N - l

2
N-K-l ("R )

 

N is the number (fl: school districts in the metropolitan area. K is
the number of independent variables used in the regression equation.
R2 is the squared multiple coefficient of determination. Calculating
the multiple partial correlations between the two explanations after
using this formula to adjust for the unequal number of independent
variables and degrees of freedom, allowed a comparison of the two ex-
planations in the seven areas.

Table 6.7 presents the multiple coefficients of determination

(R), the squared multiple coefficients of determination (R2), the

1229

oeouc.
m -. m CZOQEOI

co coouam xoh u
:0 coocam xmh u Iomxh

_omx»

“Lexem xmh u F¢Ouum

o_>um ec_4 u xxmco

xcmm

_m_uow a x24:

co_0moucwom n umm

 

 

 

 

 

 

 

 

 

 

 

mm_ . 2 m0_ 1 z 0N . z 0N 1 z
.nmx»
NJ... Jm... mN.I mm.l IOOXH
00.- N0. *4. mm.. 0..- .0. 04.- MN. #00000
00. m_.- 0_.- NE. 00. 00.- xx. 4..- 0m.- 0x. 04 - 04.- xcmca
_o. oo.1 mo.1 NN. 40. on.- _o.- m.. No. ~m.- __.- mm. mm 04.- mo. 0m. xz«m
co. 00. mm. m_.- ON. ~m. m. - m”. mo.- 4». m_.- mo. Pm. mo. mm mm. mm. 4m. 44.- mo.1 umm
cu>om __< comoxpax uc__u Cwaumw
m. n 2 4m u 2 4m u z m_ u z
_omx»
05.- mm.- _m.- 00.- . romxh
40. .0. mo. mo. om m.. mm. o~.- hmouum
0.. x,.- x4.- 00. m~.- x0. 0N. 04.- am 0m. 00.- 0m.- xcmua
m_. m4.- mm.1 0m. m_. wm.- __ - _~. mm mm - m~.- om. mm - m.. mm.- m". xz<¢
mo.1 ~_.1 x4.1 4m.- mo. um. mu.- mm. m..- mm o..- 4. Nm. oo .m.- 44. 4m 1 \m mm.- ~_.- mum
monomm ocoLm
la IA 11, 11 H S I. I; .J 1. u. 1 r) I... l 3 11 ud Q. l l 3 1. 3 S
X x .1. .1 V. 2.. v; v. 4.. .1. 7 3 y.» Y 4.. .1. V 3 X VA .1 -J V 3..
a e .1 S N 9 p. 8 A]. S N “J g a 11. S NH O 8 8 .3 S N nJ
0 0 0 I. N 0 0 0 L X 0 G 0 I. X 0 n. 0 l M mega-Jeni
l H H. A l H B A 1 H E A l H U A
l l 1 1 co_umcm.QXw
u~E0coow mucoum_o .uom u_:0couw uucmunmo .uOm u_Eocouw vocmom_o .qu u_EOcCum oucmum_o .UOm mead

 

IE. .. II.

omcv>_c3 oc_ucm oz“

new mco_oo¢ cu>wm c_

mo_nm_cm> “ocucoomoc_

coozuom mco_uo_occ0u Lance-oLo-1.e.o w4m<k

 

laDUBISIQ ‘3031 [:onuoa3 ]

 

31m0u033 l

 

 

[31091910 “DUSII

130

corrected R2 (fiz)

, and multiple partial coefficients of determination.
The table shows the striking differences in the explanatory power of
the two sets of variables in each metropolitan area. In three of the
metropolitan regions the social distance variables clearly explained
more of the variance (R2 and fiz) than did the economic variables. In
Jackson and Grand Rapids the social distance variables explained more
of the nonadjusted variance (R2) than did the economic variables. How-
ever, when the amount of variance explained was adjusted to compensate
for the number of independent variables and degrees of freedom, the
relationship became nonexistent. In Muskegon and Flint the economic
variables explainedrmnwaof the variance than did the social distance
variables. It would appear that the lack of any significant finding
in Table 6.6 was due to compromising the power of each explanation by
combining all metropolitan regions.

The independent variables differ in the amount of variance ex-
plained. Table 6.8 subdivides the two explanations or models into the
component independent variables. It lists the amount of variance ex-
plained by each variable (R2), the zero-order correlation between the
variable and the dependent variable, and the Beta weight for each vari-
ble. It also gives the same data for two economic variables in the
four metropolitan regions where tax burden on home ownership and tax
burden on income were highly correlated.

In Detroit and Flint, segregation contributed the greatest
variance explained by the social distance variables. Social rank was

most important in Lansing. In the other four areas life style differ-

ences contributed greatest to the amount of variance explained by the

13l

.c0.umcm_axm comm c. mo_nm_cm> .0 c0333: .macoc: on“ .0. umaaom 00 com: m_ awn. c0_umc_Ecoumc .0 0co_o_wwmoo

c0.mno.omc v.a_u_:s umuom.cou och .mo.nm_cm> 00cc“ __m mm __o3 mm co_.ocm_qu u_50coom 0:0 :_ mv_nm_cm> 030 mcmma Dogan

-EOu m.m2 mco.um_mccoU .m puma o_a_0_3E on“ ere co_uoc.Ecwumo co mu:o_o_c.ooo c0_mmmcmoc o_a_0_:E ecu .uc__u uco .cOmoxmsz
n

a

 

 

 

 

 

 

 

 

.Cimxum .nu_amm croco c_ 0.00:_ :0 cmocnn xm. 0cm mLQCROmEOL co cwocnn x0. cmmzuvn co_um_mc.00 £m_L or. .0 emsmuvmfi

mm. m: 00. “J 00 m. on. me. .mmv oucm0m_o _m_uom

mm a :0. ma. co m_. co. mo. Q.. 0.. mo. mo. :n cu .n. .0.
nu..o..c0u 0.:0coow

no. aw go 0. an am so. a. .N. u.:o:oum

wm mm 09. m. cm. _m on. N“. .m. .m. mm. on. so. on. Am. co. co__o.u:oU
wocm0n_o _m_00m
.Nm 0cm am cow v0.3aaou mm co_umc_Ecmumo .o m.cm_u_ewo0u _m_ucma m_a.0_:z

II A.

:4. mm. mm. _.. om. oo. 00. mo. mm. 00. mo. Om. mm_nm_.m> 03F

ma. mm. mm. m.. .4. no. go. mo. 0 . oo. ._. am. No. No. o.. co. m_. om. ... .m. cm. mm_nm_cm> ooczp
0.50couw

mm.um_cn> vocab
m_. mm. oo. no. mo. mm. co. mm mm 0s :m. mm mm mm mm. mm. an. mm. mm. mm. mm. monoumnn _o_o0m

0m. Nu. .m. Om. on. mm. up mm ma. Ob. ow. .A. ¢¢c.om_cm> w>_u
ma. mo. mm. :0. mm. a.. so am. mm. co. om. .m. am. an. mm. Nu. m:. we. mg. :m. mm. mm.om..c> x.m ..q

. .. z . z: . . m
mm mm a mm mm a mm mm mm -amm x Nm N. x nm N. «. mm Na
comoxmaz conxomw mn_ucm ocvco 0.0cuoo oc_ncm4 zmc_mmm

.mo_o;u u__nzq ncm >o__0a u__n3a .0 mc0_.mcm_axm 03p .0. co_0mcmecwgmo .0 mco_.m_occ0u _m_ucma u_a_u_3z
ocm .A m. co_.mc_eccaoo .o mocumu_..w0u :o_mnocmmm m_a.u_nz nouooLLOU .co_umc_Eco.uo mo muco_u_0000u o_a_0_:z-u.n.o ubm<w

l32

model. Tax effort contributed the most variance explained by the econ-
omic variables in Grand Rapids. Tax burden on homeowners was the most
important economic variable in Saginaw, Lansing, Detroit, Jackson, and
Flint. Tax burden on income was the most powerful economic variable

in Muskegon.

Although the ranking of the variables differed, the average
contribution made by each variable to the power of each explanation
shows a relatively equal contribution. The average percentage of vari-
ance explained by life style, social rank, and segregation was l6,
l0.9, and 9.7, respectively. For the economic variables, tax effort,
tax burden on homeowners, and tax burden on income had averages of
5.4, II, and 6.7 percent, respectively.

The multiple partial coefficients of determination show that
the amount of variance explained by each model varied from one metro-
politan region to another. In five of the seven regions, the social
distance variables were better at explaining the variance in the vote
on Proposal C than were the economic variables. The economic variables
were more successful than were the social distance variables in Muske-
gon and Flint. Table 6.9 highlights these differences. The seven
metropolitan regions are ranked according to the success of the social
distance variables in explaining the voting on Proposal C. The ranking
is based on the difference between the multiple partial coefficients
of determination for the noncorrected variance explained. fie was not
used because of the very unstable nature of the statistic for regions

with few independent variables and small number of school districts.

 

-7

r.‘
(_3

1
cv\

UN
‘l'\

La‘

'33

\n

t(\
’\
i‘i
M
g.“

4

-f\ O
C) C'

U\D
a

(win-3

\L)-3 \D

J -
("I
-rh-
a

wm. mo. mm. noc_m_axw mucm_cn> _QHOF
N Om. mm. QEOUC_
:0 coccjm xm»
”3.- No.- me. _o.. m..- Mi. n.0ckoycoz
:0 cmoczm xm»
mm. mm. .3. u. mo. .m. mm. cg, “LC..w xm»
wu_nmm;a>
..m.-,.m...0.:m 0...?
c. .6 7:00: ”3... ...Juw . . __
3.. ma. ._ 4.. m_. NM. a;qu.g.. mucm..m> .m...
..w am. mm m.m- no. .0. ...- 4... mo. go. no. on. ...- a..- .o. “.m- m... mo. 2500:.

:0 conczm x0»
ncuczovCCI
:0 concam xmp
. 1o. 5.. mo. __. mm. .0. ms. __. .o. no. -.. No. NO. ON. go. ..o... xx.
U_E0c00m .._

.T
('5
N
C)
(J

T)

l
"a
f.)

I
U'\
C.
(3

O

I
rn

(7‘
’1)

”a
Q
rx
N

m
0
O
O
U

1‘-

C.

\d

\c
m
|\
CW
n

 

 

 

0:
LC

 

 

 

 

..—

 

mo. mm. :m. RN. :4. mm. osc_m.qxu mucm_cm> _muoh
mo. on. um - wq - on m“. 4.- m.. a. - a..- o. m..- mm.- m. 3m. - _x.- :4. my.34. a...
m., mo. u: - on - mu. m. _. .o. c.. m. .0. mm,- a: m.. 3N. - mm. mm. «cam _o_u3m
m_. mg. .m.- we co. \. :w as. mm ma. m.. cm. ow. mo. 33. mm kc. :o..ommcnmm
pucm.m_o .m.u;m ._
xo_nm_cm>
-_ ix. . .. .n. . -. i..u . s .r -. «y. . L ..x
. mil: .IilllIi-x-Iilil,vl .ll-i 1. . -simi-. --o -iiii-i-illimii- - - l i .I- mum-l. pg... _anch,uz
...r«.-.r..£....l...l ..u..r(..)fi-_. {wamNzM—m . .

.mqufm UL1.; 0.C.w;: UC./LT.

Lm\’ xvremquQ —F..U.fg\w pu.~.r.._ knu...‘ ..I.u;&mw_nd3 ”smudmu. w...fi.. nfr...mufiv-I~I.Luv.v LfllfiJLDCr _qvm .r...u_‘|!..m.._L...-wno .*A.d f~r_1J-UNL|bAUfi—U m—Qmwfilr-’.mc® Ul—M4NP

Ln . _WZui_._JI v.uL> nfimcf inh._tcch;r 1;. 1. JC.:0~.¢> 0.2»6Lcm .3..m?3& .502 32% OK. 1:n ¢D_nfl_um> UmecCum

13h

 

 

 

 

 

00. NN.- 00. ms. 00. 0.. mm.0m.tm> 0.200000 02F

N 00.- 0m.- :0. 00. 00. 0.. mm.0m.cm> 0.500000 mmczh
commxmaz

0m.- 00.- NM. 00. 00. 0N. mm.0m.tm> 0.500000 oz»

0 mm.- 0... 0m. 0:. 00. 0N. mm.am_tm> 0.000000 moLLF
as...

00. m.. 00. m.. 00. 00. mm.0m_cm> 0.500000 03H

m 00. 0.. 00. 0.. 00. .m. mo.0m.cm> 0.50:000 match
comxumw

00. .. 00. m0. 00. 0.. .832.5.. 0.500000 oz»

0 00. 0.. 00. m0. 00. 0.. mm.0m..m> u.eocoum 00.00
mn_amm vcmco
m .N. .N. 0.. 0.. .m. .m. 0.0.0m0
N .N. mm. 00. 00. RN. 0:. 00.mcm.
. 00. Nm. :N. 0N. :0. 00. zmc_0mm

mm 00 mm N0 an. N0
mucu_o_wmoOu oucmum_o _mWUOm o_Eocoom

mc_xcmm _m_ucm¢ m_a_u_:z co» uo__0cucou

coozuom oucocomm_o

u_60coum

Lo» no__0cucou
oocmum_o _m_00m

moc< cmu__000cumz

 

.comumcm_axm comm mo
m:_xcmm o>_um_mm 050 can mmoc< cmu__oa0cumz co>om c_ ou_ozu u__n:¢ vcm >u__om
u__n:a wo mcomumcm_axm oz» co» co_umc_Ecouoo mo mucu_u_mmoOu _m_ucm¢ o_a_u_:zun.m.m mum<h

I35

Metropolitan Differences

 

Why should metropolitan areas differ as to the model that best
explains voting behavior? Two general explanations may account for the '
differences. One explanation lies in the characteristics of the metro-
politan region. The social and demographic characteristics of Flint
and Muskegon may be sufficiently different from the other five areas
to account for the differences in the two sets of variables. Three
hypothesis are related to this explanation.

The first hypothesis about social and demographic characteris-
tics is that the differences in the strength of the two models may be
a result of the nature of the independent variables. Conceivably,
values for the social distance model may be greatly different for
Muskegon and Flint than the other urban areas. Table 6.]0 presents
the mean, ranking of the mean, standard deviation, and ranking of
the standard deviation for each metrOpolitan area. A Spearman rank-
order correlation between the mean of each variable and the ranking
of the power of the social distance model is also presented. The
rank-order correlations indicate that there is little relationship
between the mean values and social distance. Only the relationship
between the mean value on tax burden for homeowners and strength of
the social distance model is significant. The high correlation is
probably a function of the importance of the school in maintaining
the suburban life style. In areas where homeowners are willing to
accept a greater tax burden to support schools, there may be a greater

tendency to be concerned about maintaining the role played by the

I36

.00000 :00; :0 0000 cam .0000 mocm00_o _m_uom :0 xcmc cmmzumn c0_0m.00000 c0000 xcmc cmEcmuamm

 

 

0.0 0.0 0.0 0.0 0.0 0.0 0.0 .0.0
00 xcmm ommco><

.0..00.. 00.0000. .m.~0. .0..00.. .0000. 00000. .0.0000. 0000 000 .0.0

00. 0.0mm. Am.~v_m Am.~0_m Am.mvm4. Am.m0mq. 04004. 00040. 0:00 0cm cmmz
0000c. co .um x00

.00 000 .0. 00. ... .0. .0. 00. .0. 0.0 .0. .00 .00 000 0000 000 .0.0

00. .0. 000 .0. 0.0 .00 000 .00 000 .0. 000 .0. 000 ... 000 0000 000 000:
QKOI CnvLDm XMP

._. 00 .0. 0.. .00 0.00 .00 00 .00 000 .0. 00 .00 00 0000 000 .0.0

00. .0. 00 .0. 00 .00 0.00 .00 00 ... 00. .0. 0.00 .00 0.00 0000 000 0mm:
:8:

.00 0.0 0.0 0.0 .00 0.0 .0. 0 0 000 ..0 .00 0.0 .00 0.0. 0000 000 .0.0

mm. 0.0 0... Amy m.: 0:0 m.0 0N0 _ m A00 0.: Amy m.m i Amy m.m xcmm Ucm cmw:
0.000 00.0

..0 0.0 .0.0. ..0 .0. ..0 .0. 0 0 .0.00 ..0 .0. 0.0 .00 0.0 0000 000 .0.0

0.. .0. 0.0 .00 0.. .00 0.0 .0. 0 m ... 0.0 .0. 0.0 .00 0.0 0000 000 0000
0000 .m.000

.0. 0.0. .00 0.0 .0. 0.0 .0. 0.0 .0. 0.0. ... 0.. .00 0.0 0000 000 .0.0

mm. 0M0 m.¢~ 000 0.0 - 0:0 0.0. a Am. m.o. n on 0.0: 000 J.m - “my o._m n Jena v70 cam:
co_uom00m¢m
mu :Omognax 0r__0 rcmxumfi nzmcu “_C-000 oc_0c00 30:_umm 0_nc..m>

nv_umm

.mm_cm_cm> 0000000000.

x_m c0 oc_xcmm com: 020 000 mc0_0m_0000u canOIxcmm
cmEmeam use .:o_0m_>mo vconcmum 0:0 00 mc_xcmm .co_0m_>oo ngmccm0m .cmoz 0:0 00 m:_xcmm .cmcziu.o_.m m40<0

137

local district in financing education. In addition, the procedure of
rank ordering metropolitan areas and tax burden on homeowners most
likely distorts to some extent the relationship between the two varia-
bles. Although the relationship is significant, neither Flint nor
Muskegon differed greatly from the other areas on the mean value of
the independent variables.

A second hypothesis is that the social distance model may be
strongest where there is a high degree of homogeneity or similarity
between districts, both economically and socially. If Proposal C
had been interpreted as an integrative mechanism, a metropolitan area
that was hetereogeneousinight have been more worried about the integra-
tive implications of Proposal C than a community that was already
homogeneous. This hypothesis is tested by comparing the average
ranking of the standard deviations for the six variables with the
strength of the social distance model. A low standard deviation would
indicate a high degree of homogeneity, whereas a high standard devia-
tion would indicate a low degree of homogeneity. The correlation is
-.l9, which would say there is very little relationship between homo-
geneity and strength of the social distance model.

The third hypothesis that may account for the difference in the
strength of the social distance model is that the model may be influ-
enced by the economic conditions of the areas. It is possible that
in poor areas the interest in the economic aspects of Proposal C may
have been greater than in relatively wealthy areas. Table 6.ll lists

a number of economic measures for the seven areas and the Spearmen

138

co xcmm ucm co_umcm_QXm mocmum_o _m_uom co 0cm; cmozumn mco_um_mccoo conco xcmm cmEcmoam:

.mo. 00 _o>m_ mucmu_0_cm_m co» .0. m.

m 00 2 cm .00 _m>o_ mucmo_w_cm_m.

.0.00.00> 00.002

 

 

00. .0. .0. .0. .0. ... .0. .0.
000.0.0 000.0.0 000.0.0 .00.0.0 000..00 000.0.0 000.0.0 00.00:.0)
oo~__macm mumum
00. .0. .0. .0. .0. ... .0. .0.
000.0.0 0.0.0.0 00....0 000.0.0 0.0.0.0 000.0.0 000...0 00000. 00000><
00. .0. .0. .0. .0. ... .0. .0.
00 0 0.. 0 00 0 00 0 00. 0 00. 0 00 0 0000 00000>0
00. .0. .0. .0. .0. ... .0. .0.
mom.m_m on~.m_m mmm.m_m __w.0_m 0~o.m~0 00_.0_m ~00.0_0 oa_m> meox mmmcm><
00. .0. .0.0. .0. ... .0. .0. .0.0.
0.00 m._m ...w N.Nm ..wm w.~0 m_.w mcmczooeo: “coucmm
00.- .0. .0. .0. .0. ... .0. .0.
000.0.0 000.0.0 000.0.0 000.0.0 .0..0.0 00_.0.0 000.0.0 000.00
cocommh mmmcm><
0o. .0. .0. .0. .0. ... .0. .0. 0000
000 0 000 0 000 0 000 0 000 0 000 0 000 0 00000X0 .00o.0000000_
00. .0. .0. .0. .0. ... .0. .0. 0000
_.:m m.m: o.m¢ 0.0: 0.0m ..Nm 0.m: v_< mumum achme
ma comoxmsz uc__0 comxomfi 00.0mm ucmeu u_Ocumo mc_mcm4 3mc_mmm m_nm_cm> 0.60:000

.mmoc< cmu__oQoLumz cm>mm c_ mm_nm_cm> :u_mm3 u£m_m mo mmc_xcmm ocm mm:_m> :mmZIu.__.m m4m<H

139

rank-order correlations between the economic characteristic and the
strength of the soical distance model. A review of the data in Table
6.ll shows that neither Flint nor Muskegon was much different from the
other metrOpolitan areas. Muskegon was the poorest metropolitan area
(average income of l0,500, SEV of l2,h32, and average home value of
l3,308). This may account for some of the voters' interest in property
tax relief in the Muskegon area. However, the same is not true for
Flint. Flint ranked second on average income, fourth on SEV, and
second on home value. The Spearman rank-order correlations show no
clear relationship between the economic characteristics of the metro-
politan areas and strength of the social distance model.

A second explanation for the differences in the two models is
the variety of local issues affecting the vote. It is hypothesized that
social distance variables may be more important in areas threatened by
cross-district busing or some form of forced school desegregation.
Indeed, social distance maintenance may affect the vote most when
social boundaries or social distance are preceived to be threatened.

To test this hypothesis, the major newSpapers in all seven
areas were analyzed for the three months prior to the election. The
number of column inches printed on the topic of integrating the public
schools was computed for both the state and local level. The column
inches were further subdivided into news items, editorials, letters
to the editor, and special features. The number of column inches was
then divided by the average length of the newspaper. This gave a ratio

of the relative importance of school integration in each urban area.

lhO

 

 

 

 

 

00.0. 00.0 0... 0 0 0... 00.. ...0 0 0 00. 0000000:
0..0. 00.0. 00.. o 0 00. 00. 00.0 00. 00. 0.. 00..0
00.0. 00.0. 00.0 0 0 .0.. 00. 00.0. 00.0 0 0.0 0000000
.0.0. .0.0 00.0 0 0 00. 00. 0..0 00.0 0 00. 00.000 00000
00.00 00.0 00.00 0 00.0 0 .0.0 00.0 .0 00 0 0..0 0.00000 .0000
00.00 00.. 0.00 0 00.0 0 00.0 00.. 0.00 0 00.0 .0202.
00.00 0.0 0. .0 0 00.0 0 00.0 00.0 0.0. 0 00.. .00000 0000.
0.0000o
00.00 00 0. .0.00 0 00.0 00. 00.. 00.0. 0..00 0 00.0. 00.0000
00.0. 00.0. 00.0 00. 00.0 0 00. .0.0 00.. 0 0 200.000
mumuw _moo. manum .mo00 oumum _moo. oumum _moo. oumum _m004
__< _mHOF mocaumou m_m_000_vm mzmz 000_Um 00 mcmuum.

 

.mmoc< cm0__000cuoz cm>om c. co_0mcmou:_

_oo;om 00 oucmucoQE_ 0:0 00 00.0mmun.~_.m m40<h

lhl

The ratios of column inches devoted to stories or comments about racial
integration in the schools indicate that there existed a variety of
interest and issues in the different metropolitan areas. Before further
analyzing the content of Table 6.ll, it may be worthwhile to look at
the status of school desegregation in the seven areas.

This brief analysis reveals the startling fact that in Muskegon
and Flint, the two areas where the economic model best explained the
vote, there was no pending lawsuit that would have forced the integra-
tion of the schools. In all other areas there has been, or was, a
lawsuit pending which threatened school desegregation. The threat was
exacerbated by the Detroit cross-district busing controversy.

In Lansing, five members of the local school board were re-
called in the November election because of a busing plan integreting

the city schools. The Lansing State Journal gave considerable cover-

 

age to the recall and the relationship between the recall vote and
integregation in the schools. The recall vote was highly correlated
with the vote on Proposal C (-.79 yes on recall and no on C, for 69
precincts in Lansing). In Detroit, Judge Stephen Roth had ordered a
massive cross-district busing plan that would have involved 53 school
districts, thus posing a threat to the viability of the local district.
The Roth decision, and the appeal of that decision, remained a major
factor in the November election, particularly in Detroit. In Grand
Rapids, the local school board was faced with a lawsuit aimed at
forcing desegregation of the schools. This plus a lawsuit in nearby

Kalamazoo drew considerable attention to the issue. In Jackson, the

th

local school superintendent was fired because of his support of a de-
segregation plan. A lawsuit followed. In Saginaw, the question of
integrating the public schools had already reached a climax and had
subsided somewhat. ln l97l the Saginaw public schools were closed for
over one month because of racial violence. The school board decided
to promote a county-wide desegregation plan that would have included
all the school districts in the Saginaw region. The plan was hotly
debated throughout the Saginaw area. The suburban districts were
successful in their opposition to the plan. Consequently, the Saginaw
Board of Education passed a desegregation plan involving only the city
of Saginaw Public Schools.

Comparing the ranking of the urban areas on the importance
of the social distance variables in explaining the vote and the rank-
ing of the areas on the amount of space given to the issue of racial
integration in the public schools by the newspapers, as measured by
column inches of new5print divided by average length of newspaper,
shows a close relationship between the two. Table 6.l3 presents the
rankings of the metropolitan regions on the two measures. The amount
of space given to the issue of school integration by the newspaper
is named the School Integration Ratio. Only Saginaw was dramatically
different on the two rankings. In the other six regions, the order
of importance of the social distance explanation was almost identical
to the order of importance given to local concerns over school inte-
gration. The reason why Saginaw did not score highly on the School

Integration Ratio may be that the public debate over school

lh3

TABLE 6.l3.--Comparative Rankings of the Social Distance Explanation
and the School Integration Ratio.

 

Rank of the

Social Distance Rank School Integration Ratio

 

Saginaw l 5
Lansing 2 l
Detroit 3 2
Grand Rapids 4 A
Jackson 5 3
Flint 6 6
Muskegon 7 7

 

Spearman Rho = .6] with Saginaw, .83 without Saginaw.

desegration was held during l97l ratherthan I972. The issue may have
been less relevant to the newspaper in Saginaw than to the citizen who
remembered the previous threat to his school district boundaries.

It appears that local issues were most important in explaining
why some metropolitan areas voted according to the social distance
model while other areas voted according to the economic model. This
initial analysis indicates that where social boundaries were felt to
be threatened, the voters responded by calling for maintenance of
present boundaries. However, when social boundaries were not threatened,

voters responded according to economic interests.

Social Differences Within School Districts

 

A fourth objective of this analysis is to determine if the con-

cept of social distance can be applied to social differences within

lhh

school districts as it was to social differencea among school dis-
tricts. This objective is important because of the large size of the
central city school districts. For example, the Detroit Public Schools
enroll over 280,000 pupils every year. This is approximately 13 per-
cent of the total student population of Michigan.

The seven central school districts were used to compare the
relationship between social distance variables within a district to
the social distance variables among school districts. Census tracts
were used as the unit of analysis in all cities except Detroit. In
Detroit, the large number of census tracts (800) prohibited their use.
However, #9 subcommunities, determined by the Detroit's Mayors Committee
for Community Renewal, were used. The subcommunities were composites
of similar census tracts, both socio-economically and geographically.
Measures of percentage of nonwhite, life style, and social rank were
computed for all census tracts and subcommunities. The variables were
then standardized. The census tract or subcommunity with the most
ghettolike characteristics were used as the reference point. The stand-
ardized values of the three variables for the census tract were sub-
tracted from all of the other census tracts or subcommunities thereby
yielding a social distance measure. The difference was assigned to the
census tract that was not used as the reference point. The referent
census tract was given the value of 0.0 on all three measures. The
three social distance variables were then regressed on the percentage
favoring adoption of Proposal C.

Table 6.lh gives the multiple regression coefficients of deter-

mination and Beta weights for the three social distance variables in

lbs

 

 

 

0. 00 00 00 0. .0 00 u z

m_. mm. 00. mm. mm. om. _m. N0 .M000

0... _o. :m.4 __. m..- mo. 0m.u 0.. om.u mo. 0.. _o. m..- 00. m_>0m o0_0

m:.i m.. mm.n 0.. .0.- mo. 00.- NC. om.u ow. mm. :o. 0.. No. xcmx _m_u0m

00. 0o. 00. 00. 00. 00. 00. 00. om. 00. 00.. 00. .m. 00. 00.00000000
0 00 0 00 0 00 0 00 0 00 0 00 0 00
comoxmzz 0:..0 00.0mm m:_mcm0 comxomw ch_mmm 0.00000

ucmcu

 

uOcuoz co>om c. mm.u_cneeoun:m co muomch mamcou c.

.mo_u_u cmu__oa

mo_nm_cm> mocmum_o _m_o0m

mmcLh co» mucm_o3 muom ocm co_umc_Ecmuoo mo mucm_o_mwo0u co_mmmcmox o_a_u_:ziu.:_.w m.m<k

lh6

seven central city school districts. Segregation, life style, and
social rank accounted for anywhere between 8l and IS percent of the
variance. Unfortunately, the explanatory power of the three variables
within the central city school district was not consistent with the
entire metropolitan area. Table 6.15 shows the amount of variance
explained in each metropolitan area and central city, the relative
ranking according to the percentage of variance explained, and the
ranking of the seven metropolitan areas on the School Integration

Ratio.

TABLE 6.lS.--Percentage of Variance Explained by the Social Distance
Variables in Metropolitan Areas and Central Cities.

 

Rank on School

2 2 .
R Metro Rank R City Rank Integration Ratio

 

Saginaw .79 (l) .80 (2) (5)
Detroit .27 (3) .81 (l) (2)
Lansing .hh (2) .39 (6) (1)
Grand Rapids .24 (6) .67 (h) (R)
Jackson .28 (5) .69 (3) (3)
Flint .36 (3) .53 (5) (6)
Muskegon .05 (7) .l8 (7) (7)

 

The Spearman rank-order correlations between percentage variance ex-
plained by the social distance variables in the metropolitan area and
the central city was +.h9. The correlation between the ranking of the
importance of school desegregation and percentage variance explained
in the central city was +.36. Neither correlation was significant,

becaue of the small n.

147

The explanation for the difference between central city and
metropolitan area on the variance explained by the social distance
model is not clear. It may have resulted from the nature of the data.
The use of a considerably smaller unit of analysis, i.e., census
tract rather than school district, may have accentuated the importance
of the percentage nonwhite. Segregation explained 75, 75, 35, 23, 59,
26, and h percent of the variance in the different central cities.
This was considerably more than segregation explained for the metro-
politan areas (7, l8, 8, 6, 2, 26, and 3 percent).

A second possiblity is that metropolitan voters saw Proposal
C as being closely associated with the Detroit cross-district busing
plan. Metropolitan voters may have been voting on the basis of main-
taining the school district. The central city voter may have been
more concerned with maintaining the character of the neighborhood.
Desegregation plans had already been put into effect in most cities,
thereby lessening the concern over the issue of the loss of local con-
trol for the central city resident but not for the metropolitan area
resident. This would account for the differences in the correlation
between the importance of the school desegregation ranking and the

variance explained by the social distance model.

Conclusions

 

Four basic conclusions can be drawn from this analysis of the
vote on Proposal C. First, the vote on Proposal C cannot sufficiently

be explained by only one model of public choice and public policy.

lh8

When multiple regression equations were computed for all l83 dis-
tricts the two models, combined, explained only 7 percent of the
variance.

Secondly, the seven metropolitan areas differed on which model
best explained the vote in that area. in five of the seven areas, the
social distance model explained the vote better than the economic model
did. In Flint and Muskegon the economic model was most successful in
explaining the vote.

Third, the social distance model was useful in explaining the
vote on Proposal C within the central city school district as well as
between school districts. The amount of variance explained varied from
l8 percent in Muskegon to 81 percent in Detroit. In the central city
districts, segregation played an important role in affecting the vote.
Segregation was much less important in the suburban districts.

Finally, the differences among metropolitan areas on the strength
of the social distance model is most closely associated with the amount
of newspaper coverage given to integration in the public schools. The
busing controversy had a big impact on all elections in Michigan in
l972. The issue was reflected in the strength of the soical distance
model.

The relationship between the importance of school desegregation
and the vote on Proposal C points to an interesting hypothesis that
deserves further inquiry. If Proposal C was interpreted as a threat
to social and political boundaries, what role does “boundary mainte-

nance” play in other political issues? Anthropologists, sociologists,

1A9

andiothershave described man as a ”territorial animal” who will fight
to maintain his place in the world. For the political scientist,
boundary maintenance behavior may have important ramifications for
political preferences and the interpretation of referenda voting. This
is particularly important in discussing either metropolitan reorganiza-
tion or educational politics.

The two models of public choice and public policy were quite
successful in explaining the vote on PrOposal C. The analysis indi-
cates that the vote, like the debate over Proposal C, was divided on
whether the economic aSpects of property tax relief and tax redistribu-
tion would ”sell” the proposal or whether the public would vote to
protect their social and political boundaries. In the final analysis,
the social distance model was most successful in explaining the vote

on Proposal C.

CHAPTER Vl--FOOTNOTES

lEshref Shevky and Wendell Bell, Social Areas Analysis:
Theory, Illustrative Application and Computational Procedures
(Westport, Conn.: Greenwood Press, l972), p. SEC

 

 

2Traditionally, econometricians have established a criterion
of .80 and above as indicating ”harmful multicollinearity,” see
L. R. Klein, An introduction to Econometrics (Englewood Cliffs,
New Jersey: Prentice-Hall, I962).

 

3See Herbert M. Blalock, Jr., Social Statistics (New York:
McGraw Hill, I970), pp. #29-470.

 

lSO

CHAPTER VII

CONCLUSIONS

The final vote on Proposal C caught a number of state policy
makers off guard. Prior to the election, there had been a general con-
sensus that Proposal C would pass because of the “property tax relief”
it would have given to homeowners throughout Michigan. Lulled by the
results of public opinion polls taken by Market Opinion Research of
Detroit, the Governor, the Michigan Education Association, the legisla-
ture, and other policy makers were confident of the passage of Proposal
C. However, when the vote count was taken, PrOposal C was soundly de-
feated. Therefore, the question is raised, why was Proposal C defeated?
This is the basic question motivating the research presented in this
dissertation.

Proposal C contained two major provisions. The first was a
constitutional limitation on the property tax. It was this provision
that many leaders assumed would “sell” the proposal to the public. The
second part of the proposal would have shifted the primary responsibil-
ity for funding public education from the local school district to the
state. It is this provision that may explain why Proposal C was de-
feated. Many of the opponents of Proposal C argued that its passage
would mean the loss of local control. This, coupled with the contro-

versy over busing, drew attention to the second aSpect of the proposal.

l5l

I52

The two provisions in the amendment led to two explanations
about why Proposal C was defeated. The first explanation is that the
vote on Proposal C should be positively related to the rate of taxation
and the tax burden imposed on local school districts. Three variables
were hypothesized to be correlated with the vote on Proposal C. They
were tax burden on homeowners, tax burden on income, and total tax ef-
fort. It seemed logical that people who stood to gain the most ”prop-
erty tax relief” would support Proposal C more than people gaining
less property tax relief. This explanation was termed ”economic“ be-
cause it stressed the economic aspects of Proposal C.

The second explanation of the vote on Pr0posal C is that the
public did not want to shift the primary reSponsibility of funding
public education from the local school district to the state. Implicit
in this hypothesis was that the public wanted to maintain present
”boundaries” between themselves and other school districts. To opera-
tionalize this hypothesis, the concept of social distance was used.
Social distance was computed by taking the difference between the
central city school district and the suburban school district on
measures of social rank, life style, and segregation. If the public
interpreted Proposal C as an integrative mechanism, opposition to
Proposal C might best be explained by studying the social differences
between the suburban districts and the central city school district.

These two explanations were tested by using multiple regression
analysis and multiple partial correlations. It was discovered that in
five metropolitan areas the social distance variables explained more

variance in the vote on Proposal C than did the economic explanation.

l53

In two areas, the economic variables explained more of the variance
than did the social distance variables.

A brief analysis of the seven metropolitan regions indicated
that local issues, particularly school integration, were most important
in determining which metropolitan regions voted according to the econ-
omic explanation and which voted according to the social distance ex-
planation.

From the research conducted on the vote on Proposal C, a number
of findings stand out in their importance for future policy decisions
and research. First, the two theoretical models used in this research
provided a meaningful conceptualization for stuyding educational finance
reform in Michigan. The history, the political campaign, and the vote
were all discussed in terms of the two explanations of public choice
and public policy. The same type of conceptualization may prove use-
ful in studying voting behavior on other referenda such as millage
elections for school districts, metropolitan consolidation, or any
referenda that might be considered ”integrative” in nature.

The success of the social distance variables in explaining the
vote on Proposal C points to the conclusion that not all political be-
havior can be explained according to economic theory. The vote shows
that few voters calculated the economic benefits that ”property tax
relief” might bring them. Thus, political sociology is also an impor-
tant aspect of the study of political behavior.

Probably the most significant finding of this research is the

relevancy of the concept of “social distance” for understanding public

ISA

choice and public policy. Chapter VI suggested that a possible explana-
tion for the relationship between the amount of news space given to
school integration and the power of the social distance variables is
that when the social boundaries of voters are threatened, they react

in defense of those boundaries even if it means increased financial
costs.

Anthropologists have theorized about this territorial nature
of man. However, there has been very little discussion or study of
the importance of social boundaries in affecting political behavior.
This research suggests that this type of analysis might prove bene-
ficial in understanding politics. For example, the concept of rep-
resenting geographical areas, political struggles over reapportionment,
and metropolitan consolidation may all be discussed by using these
concepts.

The concept of social distance was found to be as applicable
to social differences within school districts as it was to social dif-
ferences among school districts. It may be that politics within large
cities can also be analyzed according to the social distance model.

It is apparent that Proposal C was not interpreted as a viable
means of attaining property tax relief. Although the public, at large,
was supportive of the need to limit or reduce prOperty taxes, support-
ers did not succeed in convincing the electorate that Proposal C was
the vehicle to accomplish that goal. For the policy maker, this might
indicate that a legislative solution to the problem might be more

feasable than a constitutional amendment.

155

Analysis of the vote on Proposal C showed that the issue of
school integration or busing was crucial in determining which explana-
tion would best fit the vote on Proposal C. It is possible that when
the issue subsides in public interest, the maintenance of social
boundaries may be less important. If Proposal C were to be presented
to the voters in Michigan again in four years, it is conceivable that
the economic explanation may best explain the vote.

The research has attempted to reflect the dynamics of public
school finance reform in Michigan from I969 to I973. The climax of
this dynamic came in the constitutional amendment to limit the property
tax as the primary source of educational revenue. Two explanations
were developed to explain the debate and vote on Proposal C. Hopefully,
the analysis of this important issue will assist the policy maker as
well as the political scientist in understanding the reform of educa-

tional finances in Michigan.

APPENDICES

1.56

APPENDIX A

School District

 

What percentage of local property tax revenue is derived from
business property (non-homeowners) in your district?

Exact percentage (if possible) %

 

Estimate (If exact percentage is not available)

 

 

 

 

 

 

 

 

 

(Check one) 0 - l0%
ll ' 20
2| ' 30
BI - AD
“I ' 50
SI - 60
6| - 70
7| - 80
8| - 90
Si ‘lOO

 

Has your district had to reduce the educational program offered to your
students in the past two years due to financial constraints?

Yes No

l57

County

Alcona
Alger
Allegan
Alpena
Antrim
Aranac
Baraga
Barry

Bay
Benzie
Berrien
Branch
Calhoun
Cass
Charlevoix
Cheboygan
Chippewa
Clare
Clinton
Crawford
Delta
Dickinson
Eaton
Emmet
Genessee
Gogebic
G.D. Traverse
Gratiot
Hillsdale
Houghton
Huron
Ingham
Ionia
Iosco
Iron
Isabella
Jackson
Kalamazoo
Kalkaska
Kent
Keweenaw
Gladwin

APPENDIX B

Vote Returns on Proposal C for 83 Counties

 

122
1796
1656

l04l3
6687
30l0
2125
1597
7232

21896
1887

21047
5632

27227
7302
2910
3415
6157
29l0
8685
I702
6231
3886

l272l
3095

7l422
6359
8761
7806
5231
7530
6355

6l228
7286
6575
29l0
89l0

26l8l

39015
1386

66636

696
2666

53
I757
1923

16703
6648
2975
2176
1601
8094

30894
2026

3277i
706I

23067
645i
6311
3662
4756
3628

10231
II83
7895
6586

l5078
4692

80104
5143
7769
5979
7452
6168
6984

52710
9055
3967
3855
7763

24475

60690
1601

101602

566
2707

158

County 133
Lake l557
Lapeer 8848

' Leelanau 2855
Lenawee l409l
Livingston lll6I
Luce ll93
Mackinac 24I0
Macomb 60420
Manistee 3668
Marquette Il938
Mason 4408
Macosta 5294
Menominee 3672
Midland l068l
Missaukee I647
Monroe l9085
Montcalm 6758
Montmarency l274
Muskegon 27043
Mewaygo 5673
Oakland l20664
Oceana 3577
Ogenaw 3278
Ontonagon 2227
Osceola .2I03
Oscoda I070
Otsego I998
Ottawa 20628
Presque Isle 2550
Roscommon 3340
Saginaw 3l022
St. Clair 9936
St. Joseph 6277
Sanilac 7274
Schoolcraft I830
Shiawassee l076l
Toscola 8489
Van Buren 8043
Washtenaw 62768
Wayne 3I0655
Wexford 3l38
TOTALS

2322
I55433
4650
l0343
5932
489i
6367
I5l68
l853
I8606
7069
I349
27263
6I66
23I833
3303
I904
2497
3895
968
2544
34874
3386
2699
3682l
l2334
l02l4
7548
I978
I2846
8I96
l0859
35507
5I6258
5l22

l,324,702 I,8l5,l26

BIBLIOGRAPHY

159

BIBLIOGRAPHY

Books and Articles

 

Advisory Commission on Intergovernmental Relations. “Financing Schools
and Property Tax Relief - A State Responsibility.” Washington,
D. C.: Advisory Commission on Intergovernmental Relations,
I973. Report A-40.

Alker, Hayward R. ”A Typology of Ecological Fallacies.” In Dogan,
Mattei and Rokkan. Stein. Quantitative Ecological Analysis
in the Social Sciences. Cambridge, Mass.: M.I.T. Press, I969.

Bailey, Stephen. Schoolman and Politics: A Study of State Aid to Edu-
cation in the Northeast. Syracuse, N. Y.: Syracuse University
Press, I962.

Bauer, Raymond and Gergen, Kenneth (eds.). The Study of Policy Forma-
tion. New York, N. Y.: The Free Press, I968.

 

Bell, Wendell. ”Economic, Family and Ethnic Status: A Empirical Test.”
American Sociological Review. Vol. 20, I955. 45-52.

”Social Areas: Typology of Urban Neighborhoods.“ In Shussman,
Marvin (ed.). Community Structure and Analysis. New York, N. Y.:
Thomas Crowell, I959.

Berke, Joel. “The Current Crises in School Finance: Inadequacy and
\Inequity.” Phi Delta Kappan. September, I97I.

 

Berke, Joel and Callahan, John. ”Serranb vs Priest: Mileston or Mill-
stone for School Finance.” Journal of Public Law, Vol. 2l, No. I,

l972.

 

Blanck, Herbert M., Jr. Social Statistics. New York, N. Y.: McGraw
Hill, I970.

 

Booms, Bernard H. and Haldorson, James R. ”The Politics of Redistribu-
tion.” American Political Science Review, September, I973.

 

Bowles, Dean B. “The Power Structure in State Education Politics.”
Phi Delta Kappan. February, I968.

 

I60

l6l

Buchanan, James M. and Tullock, Gordan. The Calculus of Consent: Logi-
cal Foundations of Constitutional Democracy. Ann Arbor, Mich.:
University of Michigan Press, I962.

 

Burkehead, Jesse. Public School Finance: Economics and Politics.
Syracuse, N. Y.: Syracuse University Press, I963.

 

Dawson, Richard and Robinson, James A. ”Inter-Party Competition, Econ-
omic and Welfare Policies in the American States.” The Journal
of Politics, Vol. XXV, I963.

 

 

Deutsch, Karl; et. al. Political Community in the North Atlantic Area.
Princeton, N. J.: Princeton University Press, I957.

 

DeW. Lewis, Karen. ”Educational Report/Federal Policy Toward School
Finance Faces Review by Administration, Courts.” National
Journal, December, l972.

Dogan, Mattei and Rokkan, Stein. Quantitative Ecological Analysis in
the Social Sciences. Cambridge, Mass.: M.I.T. Press, I969.

 

 

Downs, Anthony. An Economic Theory of Democracy. New York, N. Y.:
Harper and Row, I957.

 

Dror, Yehezkol. Public Policymaking Reexamined. San Francisco, Calif.:
Chandler Publishing, I968.

 

Duncan, Ottis 0. ”Review of Social Area Analyais by Shevky and Bell.”
American Journal of Sociology. (Vol. 6I, I955.

 

Dye, Thomas R. “A Model for the Analysis of Policy Outcomes.” In
Sharkansky, Ira (Ed.). Policy Analysis in Political Science.
Chicago, Ill.: Markaham Publishing, I970.

 

Dye, Thomas R. “City Suburban Social Distance and Public Policy.” In
Dye, Thomas R. and Hawkins, Brett (Eds.). Politics in the
Metropolis. Columbus, Ohio: Charles Merril Books Inc., I967.

 

 

Dye, Thomas R. et. al. ”Differentiation and Cooperation in a Metropoli-
tan Area.” Midwest Journal of Political Science, Vol. 7, No. 2,
May, I963.

 

Dye, Thomas R. Politics, Economics, and the Public: Policy Outcomes
in the States. Chicago, Ill.: Rand McNally, I966.

 

 

Dye, Thomas R. Understanding Public Policy. Englewood Cliffs, N. J.:
Prentice Hall, I972.

 

Fabricant, Solomon. The Trend of Government Activity in the United
States Since I900. New York, N. Y.: National Bureau of
Economic Research, I952.

 

 

I62

Fisher, Glenn. “Determinants of State and Local Government Expendi-
tures: A Preliminary Analysis.” National Tax Journal, Vol.
XIV, December, l96l.

 

Froman, Lewis A., Jr. ”An Analysis of Public Policies in Cities.”
Journal of Politics, Vol. 29, I967.

 

Fry, Brian R. and Winters, Richard F. ”The Politics of Redistribution.“
American Political Science Review, Vol. 64, June, I970.

 

Guthrie, James; Kleindorfer, George; Levin, Henry; and Stout, Robert.
Schools and Inequality. Washington, D. C.: The Urban Coali-
tion, I969.

 

Hawkins, Brett W. ”Life Style, Demographic Distance, and Voter Support
of City-County Consolidation.” Southwestern Social Science
Quarterly, Vol. 58, September, l972.

 

Hirsch, Werner 2. Analysis of the Rising Cost of Public Education.
Washington, D. C.: Joint Economic Committee, I959.

Hofferbert, Richard I. ”The Relation Between Public Policy and Some
Structural Variables in the American States.“ American Politi-
cal Science Review, Vol. LX, MarCh, I966.

 

 

lannaccone, Laurence. Politics in Education. New York, N. Y.: The
Center for Applied Research, I967.

 

Jacob, Herbert and Lipsky, Michael. ”Outputs, Structure, and Power:
An Assessment of Changes in the Study of State and Local
Politics.” In Hofferbert, Richard and Sharkansky, Ira (eds.).
State and Urban Politics. Boston, Mass.: Little, Brown, and
Company, l97l.

JODDS. Roe and Morphet, E. L. Problems and Issues in Public School
finance; New York, N. Y.: Teachers College, Columbia Uni-
veristy, I952,

 

Kaufman, Walter and Greer, Scott. ”Voting in a Metropolitan Community:.
An Application of Social Area Analysis.” Social Forces, Vol. 38,
March, I960.

 

Keller, Ed. ”Property Tax Relief and Local Control: Separable Issues.”
The Michigan Elementary Principle, Vol. 47, Summer, l972.

 

Kirst, Michael. The Politics of Education at the State and Federal
Level. Berkeley, Calif.: McCutchan Publishing, I972.

 

Klein, L. R. An Introduction to Econometrics. Englewood Cliffs, N. J.:
Prentice Hall, I962.

 

I63

Klingele, William E. ”Curtailing Expenses in Michigan Public Schools.‘I
Michigan School Board Journal, Vol. 20, April, I973.

 

Master, Nicholas A.; et. al. State Politics and the Public Schools:

An Exploratory Analysis. New York, N. Y.: Alfred A. Knopf,
1964.

 

 

 

Menzel, Herbert. ”Comment on Robinson's Ecological Correlations and
the Behavior of Individuals.” American Sociological Review,
Vol. l5, October, I950.

 

Michigan Education Association. ”Tax Pr0posal Answers.” Teachers
Voice, Vol. 50, September, I972.

Mitchell, Joyce and Mitchell, William C. Political Analysis and Public
Policy. Chicago, Ill.: Rand McNally, I969.

Mitchell, William C. ”The Shape of Political Theory to Come: From
Political Sociology to Political Economy.” The American Be-
havioral Scientist, Vol. II, November-December, I967.

 

 

Olson, Mancur. The Logic of Collective Action. Cambridge, N. J.:
Harvard University Press, I965.

 

Pinner, Frank, et. al. Education and the State: Decision Making on
the Reform of Education Finances in Michigan. Report to the
Urban Coalition, Washington, D. C., l97l.

 

 

 

Ranney, Austin (ed. ). Political Science and Public Policy. Chicago,
Ill.: Markham Publishing, I968.

 

Ranney, Austin. ”The Utility and Limitations of Aggregate Data in the
Study of Electoral Behavior.” In Ranney, Austin (ed.). Essays
On the Behavioral Study of Politics. Urbana, Ill.: University
of Illinois Press, I962.

 

Rao, Tohluri and Miller, Roger LeRoy. Applied Econometrics. Belmont,
Calif.: Wadsworth Publishing, I97l.

 

Rees, Phillip. “Problems of Classifying Sub-Areas Within Cities.” In
Berry, Brian (ed.). City_Classification Handbook of Methods
and Applications. New York, N. Y.: Wiley, Inc., I972.

 

 

Riker, William. The Theory of Political Coalitions. New Haven, Conn.:
Yale University Press, I962.

 

Robinson, W. S. ”Ecological Correlations and the Behavior of Individu-
als.” American Sociological Review, Vol. I5, June, I950.

 

Ross, Robert. ”Public Choice in American Politics: An Introduction.“
In Ross Robert (ed.). Introductory Readings in American Govern-
ment: A Public Choice Perspective. Chicago, Ill.: Markham
Publishing, l97l.

 

 

I64

Rossmiller, Richard. Fiscal Capacity and Educational Finance. Gains-
ville: National Educational Finance Project, I97].

 

Scheuch, Erwin K. ”Social Context and Individual Behavior.” In Dogan,
Mattei and Rokkan, Stein (eds.). Quantitative Ecological
Analysis in the Social Sciences. Cambridge, N. J.: M.I.T.
Press, I969.

 

 

Shapiro, Sherman. ”Some SociorEconomic Determinants of Expenditures
for Education.“ Comparative Education Review, Vol. VI, I962.

 

Sharkansky, Ira and Hofferbert, Richard I. ”Dimensions of State Poli-
tics, Economics and Public Policy.” American Political Science
Review, Vol. 63, September, I969.

 

Shively, W. Phillips. ”Ecological Inference: The Use of Aggregate
Data to Study Individuals.” American Political Science Review,
Vol. 62, December, I969.

 

Sullivan, John L. ”A Note on Redistributive Politics.” Unpublished
paper, Iowa State University, I972.

Thomas, J. Alan. School Finance and Educational Opportunity in Michi-

an. Lansing, Michigan: Michigan Department of Education,
I968.

 

Urban Coalition. Public School Finance: Who Pays and Who Benefits.
Washington, D. C.: The Urban Coalition, I972.

 

Usdan, Michael. The Political Power of Education in New York State.
New York, N. Y.: The Institute for Administrative Research,
Teachers College, Columbia University, I963.

 

Van Arsdol, Maurice; Camilleri, Santo F.; and Schmid, Calvin. I'The
Generality of Urban Social Area lndices.” American Sociologj-
cal Review, Vol. 23, I958.

 

 

Wade, L. L. and Curry, R. L. A Logic of Public Policy: Aspects of
Political Economy. Belmont, Calif.: Wadsworth Publishing

 

 

 

Co., I970.

Walter, Benjamin and Wirt, Frederick. ”The Political Consequences of
Suburban Variety.“ Social Science Quarterly, Vol. 58, December,
l972.

Williams, Oliver. I'Life Style Values and Political Decentralization
in Metropolitan Areas.“ Social Science Quarterly, Vol. 48,
December, I967.

 

Williams, Oliver; et. al. Suburban Difference and Metropolitan Policies.

 

Philadelphia, Pa.: University of Pennsylvania Press, I965.

I65

Wirt, Frederick and Kirst, Michael. The Political Web of Schools.
Boston, Mass.: Little, Brown, and Company, I972.

 

Wood, Robert. Suburbia: Its People and Their Politics. Boston, Mass.:
Houghton Mifflin Company, I958.

 

Zimmer, Basil and Hawley, Amos. Resistance to Reorganization of School

 

Districts and Government in Metropolitan Areas. Providence,
R. I., Brown University Press, I966.

Documents

 

California Supreme Court. Serrano v Priest. 5 Cal. 3d 584, 487, p 2f,
l24l, 96. Rptr. 60l, I97I.

Journal of the House of Representatives, Lansing, Mi.: State of Michi-
gan, No. I08, July 29, l97l.

Gongwer News Service. Lansing, Michigan. April 5, I969; September I.
l97l; April l5, I97]; May 30, I972; October ll, I972; and
February IO, I973.

 

Mayor's Committee for Community Renewal.) A Profile of Detroit: I969.
Detroit: City of Detroit, Michigan, I970.

 

Michigan Chamber of Commerce. Special Report. April, l972.

 

Michigan Department of Education. Local District Report: Explanatory
Mateials for Assessment Report No. 6, I97l.

 

Michigan Department of Education. IOIZ Bulletins (I960-I973).

 

Michigan State American Federation of Labor and Congress of Industrial
Organizations. Press Release, October l7, I972.

Michigan Supreme Court. Butcher v the Township of Grosse Isle. State
of Michigan Supreme Court, No. 52935, June, I97I.

 

Office of the Governor. A Chronolggy of Educational Reform in Michigan.

 

Lansing, Michigan: Office of Planning and Coordination, I970.
Office of the Governor. Press Release, March I8, I972.

Office of the Governor. ”Special Message to the Legislature on Ex-
cellence in Education.” April I2, l97l.

I66

Interviews

 

Erickson, Kai. Director of Public Relations, Michigan Education Associa-
tion. November 6, I973, and February 9, I973.

McKerr, Robert. Associate Superintendent for Finance, Michigan Depart-
ment of Education. April 5, I973.

Pastula, Richard. Special Assistant to the Governor. February 9, l973.

Newspaper Editorials

 

Detroit Free Press. ”C and D would Bring Equity in School Finance.”
November 5, I972.

Detroit News. “No on Proposal C.“ October 4, I972.

 

Grand Rapids Interpreter. ”Clever Selling Job May Backfire on Proposal
C at Ballot Box.” October I9, I972.

 

Jackson Citizen Patriot. “Legislature Key to Proposal C.“ October 25,

I972.

 

LansinggState Journal. “Yes on C but a No for 0.” October 3i, I972.

 

Macomb Daily. “Busing Cuts Big Swath in Statewide Balloting.“ Novem-
ber I3, I972.

 

Muskegon Chronicle. ”Problems Remain, but Vote Yes on Proposal C.”
October 26, I972.

 

Saginaw News. “Proposal C Presents Tax Reform Opportunity.” November 2,

I972.

 

MICHIGAN STATE UNIV. LIBRQRIES
llIllWilliillllllllllilllllllllllllllllllllilllll
31293102115643