EXAMINING FUND BALANCE IN MICHIGAN SCHOOL DISTRICTS
By
Zainin Bidin

A DISSERTATION
Submitted to
Michigan State University
in partial of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
K-12 Educational Administration
2012

ABSTRACT
EXAMINING FUND BALANCE IN MICHIGAN SCHOOL DISTRICTS
By
Zainin Bidin
This research examines the financial profiles of 550 public school districts in Michigan
and highlights the association between school district fund balance and the following eleven
indicators: enrollment, percent enrollment change, percent of students receive free and reduced
lunch (FRL), percent of special education students, percent of English Language Learners (ELL),
per pupil foundation allowance, urban, property taxable value, pupil-teacher ratio, average
teacher salary, and business and administrative expenses as a percentage of current operating
expenditures from fiscal year 2001 to 2010. School districts financial profiles display a rising
number of districts in financial deficit and near deficit, which is defined as fiscal stress in this
study. Throughout the years, the mean of the district fund balance has steadily declined, and the
movement of school districts in and out of deficit and fiscal stress has been unpredictable.
Utilizing pooled regression analysis, an investigation was conducted to reveal the positive and
negative associations with the district fund balance. Four are found to be positively associated,
these include: foundation allowance, property taxable value, pupil-teacher ratio, and number of
ELL students, which is one of the high-cost student indicators. The remaining six are enrollment,
enrollment change, FRL, special education, salary, and business administrative expenditures and
has a negative association with the district fund balance. Compared to districts located in rural
areas, urban school districts have a lower fund balance, but for both school districts a higher fund
balance could be achieved with more funding. On the other hand, giving money to school
districts will not increase their fund balances. A plan must be put into place to spend the money

effectively, taking into account teacher’s salary, class size control, and business administrative
expenses. With all of these in mind school district managers will have a greater chance of
maintaining their fund balance and fiscal status. Three cross-section regression analyses for
fiscal years 2000, 2005, and 2010 with percent enrollment change in one and three years proved
to be consistent with the pooled regression analysis. These cross-section analyses are useful in
the short-term but will not estimate the fund balance accurately in five or ten years.

DEDICATION
To my wonderful family: my understanding and patient husband, Tohar Besah; my
precious children, Ikram, Iman, Izzah, Iddeen, and Ilham; my loving parents, Tiah and Bidin; and
all school children out there. Thank you for your inspiration, love, and support that make my
dreams come true.

iv

ACKNOWLEDGEMENTS
I would like to acknowledge the help and valuable support of my academic advisor who I
am deeply indebted to, Dr. David Arsen. Your invaluable expertise and guidance kept me on
track throughout the research process. I would also like to acknowledge and thank the rest of my
committee, Dr. James Fairweather, Dr. Susan Printy, and Dr. Amita Chudgar, for their
significant contribution to this research. Special thanks to CSTAT Department and Emre
Gonulate for sharing your knowledge in statistics with me. I am grateful to the Malaysia Ministry
of Education for funding my study and the College of Education for scholarships awarded to me.
My special appreciation and thanks also go to the support I received from various individuals
during this endeavor. Thank you God and thank you everyone.

v

TABLE OF CONTENTS

LIST OF TABLES…….…………………………………………………………………….….viii
LIST OF FIGURES…...……………………………………………………………………….….x
CHAPTER 1
INTRODUCTION……….…..……………………………………………………………………1
1.1 Overview………………………………………………………………………………1
1.2 Statement of the Problem…………….………………………………………………..4
1.3 Research Questions........................................................................................................4
1.4 Significance of Study………….………………………………………………………5
CHAPTER 2
LITERATURE REVIEW…………………………………………………………………………7
2.1 Proposal A…………………………………………………………………………….7
2.2 Challenges in Proposal A…………………………………………………………….10
2.2.1 State Economy Decline……….……………………………………………11
2.2.2 Declining Enrollment………………………………………………………14
2.2.3 Revenue-Cost Gap…………………………………………………………16
2.2.4 Capital Funding………….…………………………………………………18
2.3 Financial Consequences of Proposal A………………………………………………19
2.4 Fiscal Stress………………………………………………………………………….20
2.4.1 Definition…………………………………………………………………..20
2.4.2 Factors Associated with Fiscal Stress……………………………………...24
2.5 Conclusion…………………………………………………………………………...31
CHAPTER 3
METHODOLOGY………………………………………………………………………………33
3.1 Research Questions…………………………………………………………………..33
3.2 Data Sources…………………………………………………………………………34
3.3 Empirical Strategy…………………………………………………………………...34
3.3.1 Definition…………………………………………………………………..34
3.3.2 Dependent Variable………………………………………………………..35
3.3.3 Independent Variables……………………………………………………..36
3.3.4 Analysis……………………………………………………………………39
CHAPTER 4
RESULTS………………………………………………………………………………………..43
4.1 Descriptive Analysis…………………………………………………………………43
4.1.1 Trends in Fund Balance……………………………………………………45
4.1.2 Changes in Fiscally Stressed Districts……………………………………..47
4.2 Regression Analysis………………………………………………………………….53
4.2.1 Test of Assumptions……………………………………………………….54
vi

4.2.2 Pooled Regression Analysis……………………………………………….55
4.2.3 Cross-Section Regression Analysis………………………………………..58
4.3 Model Stability………………………………………………………………………62
4.4 Conclusion…………………………………………………………………………...63
CHAPTER 5
DISCUSSION……………………………………………………………………………………65
5.1 Summary of the Study……………………………………………………………….65
5.2 Research Findings……………………………………………………………………69
5.3 Interpretations of the Research Findings…………………………………………….71
5.4 Policy Implications…………………………………………………………………..75
5.5 Scope and Limitations of Study……………………………………………………...77
5.6 Future Research……………………………………………………………………...77
APPENDICES…………………………………………………………………………………...79
APPENDIX A
Michigan public schools with deficits for fiscal year
ending June 30, 2011……….……………….…………………………………………...80
APPENDIX B
District location definitions…………….………………………………………………..81
APPENDIX C
Test of assumptions……………………………………………………………………...83
REFERENCES………………………………………………………………………………….90

vii

LIST OF TABLES

Table 2.1

SAF Revenue Sources for K-12 Education in Michigan
Before and After Proposal A………………………………………………………8

Table 3.1

Fund Balance Indicators…………………………………………………………39

Table 4.1

Descriptive Statistics for All Variables………………………………………….44

Table 4.2

Descriptive Statistics for Urban…………………………………………………44

Table 4.3

District Fund Balance as a Percentage of
Total Expenditures, 2001-2010…………………………………………………..45

Table 4.4

Distribution of Districts by Fund Balance
As a Percentage of Total Expenditures…………………………………………..46

Table 4.5

Number of School Districts in Fiscal Stress, 2001-2010………………………...48

Table 4.6

Districts in Fiscal Stress* and Deficit**, by Number of Years …………………49

Table 4.7

Districts in Fiscal Stress Continuously from 2001 to 2010……………………...50

Table 4.8

Districts in Deficits from 2001 to 2010, by Number of Years…………………..51

Table 4.9

Number of District Movement of Fiscal Stress*
And Deficit**, 2001-2010……………………………………………………….51

Table 4.10

Regression Analysis of District Fund Balances, 2001-2010…………….............56

Table 4.11

Regression Analysis of District Fund Balances, 2000, 2005, 2010
(percent enrollment change in three years)……………………….……………...59

Table 4.12

Regression Analysis of District Fund Balances, 2000, 2005, 2010
(percent enrollment change in one year)………………….…...…………………61

Table 4.13

Paired Samples Test……………………………………………………………...63

Table A-1

Michigan public schools with deficits for fiscal year
ending June 30, 2011…………………………………………………………….80

Table C-1

Variance inflation factor – enrollment…………………………………………...83

Table C-2

Variance inflation factor – percent enrollment change.........................................83

Table C-3

Variance inflation factor – percent free and reduced lunch……………………...84
viii

Table C-4

Variance inflation factor – percent special education students...………………...84

Table C-5

Variance inflation factor – percent English Language Learners………………...85

Table C-6

Variance inflation factor – foundation allowance.................................................85

Table C-7

Variance inflation factor – property taxable value….…………………………...86

Table C-8

Variance inflation factor – pupil teacher ratio…………………………………...86

Table C-9

Variance inflation factor – average salary…..…………………………………...87

Table C-10

Variance inflation factor – business administrative expenditures….……………87

Table C-11

Variance inflation factor – urban………………………………………………...88

ix

LIST OF FIGURES

Figure 4.1

Number of Fiscally Stressed and Not
Fiscally-Stressed Districts, 2001-2010…………………………………………..47

Figure C-1

Histogram………………………………………………………………………...88

Figure C-2

Normal P-P plot of regression standardized residual…………………………….89

Figure C-3

Scatterplot………………………………………………………………………..89

x

1
INTRODUCTION

1.1

Overview
Over the past few years, the number of Michigan public school districts that have a

declining fund balance has been increasing. Many of them have deficits and are facing serious
fiscal problems. These fiscal problems have captured the interest of Michigan educators,
taxpayers, and lawmakers with regard to how the funding system works. The state’s education
funding system was intended to improve the distribution of funding across school districts.
However, the real per pupil revenue to fund schools has been declining since 2002, and this has
forced school districts to respond with budget cuts. Michigan’s K-12 education funding, which
relies on the state’s economic health, has been made even harder in the current recession.
As a majority of states have reformed their mechanisms through which schools are
funded, Michigan K-12 education has also gone through its own school finance reform. In 1994,
Michigan adopted a new reform for school financing called Proposal A, the primary goal of
which was to change the approach used in collecting and distributing revenues for education
provided by the state. School funding was reformed by creating a new system in which the
majority of revenues are raised by the state, and funding is distributed based on a per pupil
foundation allowance formula, with foundation levels determined by the state annually. Most
funds that support K-12 education are now controlled by the state, whereas local districts
maintain control over capital funding, which is raised through local property taxes.
Through the foundation allowance approach under Proposal A, the spending gap between
1

the highest-revenue and lowest-revenue school districts has steadily decreased since it was
implemented (Arsen & Plank, 2003). However, the reform resulted in a shift from local to a
centralized system of Michigan public school finance. Local districts are constrained with the
revenue levels set by the state to survive with their budgets and adjust their spending priorities.
Prior to Proposal A, Michigan’s public school funding relied heavily on local property
taxes for education revenues. Michigan’s property tax burden was more than 33 percent above
the national average before Proposal A was adopted (Michigan Department of Treasury, 2002).
Aside from being concerned with education inequalities, Proposal A promised Michigan voters
significant property tax relief. In order to compensate school districts for the loss of revenues, the
sales tax and other taxes were raised to provide the state with the additional revenues needed.
Since revenues for these taxes fluctuate with changes in the economy, school funding, which is
now directly linked to the state’s economy, becomes less stable. When the state’s economy is
facing a downturn, additional revenue to fund schools is no longer a possibility. Districts need to
make adjustments in their budgets. However, not all of them are able to strengthen their financial
condition. Some of them do not avoid falling into deficit, and the task is even more critical for
those districts that already confront financial pressures. For example, Detroit Public Schools’
general fund has been in deficit every year since 2008.
In 2010, the Michigan Department of Education (MDE) reported 43 school districts in
financial deficit in the state. Seven of them are charter schools and 36 are public school districts
(Appendix A). Even though this represents only five percent of the total of 848 education entities
(551 local districts, 240 public school academies, and 57 intermediate school districts), it was a
138 percent increase from 2005, when there were only 18 school districts in financial deficit.
According to the MDE, a district is considered a deficit district if it is not able to maintain a
2

positive fund balance at the end of the fiscal year. Deficit districts are then required to submit a
deficit elimination plan (DEP) to MDE. The MDE has established a procedure that requires
school districts to enter into a mandatory deficit reduction program when its audited financial
report to the state reveals a deficit fund balance in the General Fund, or deficits in any other
funds which surpass the surplus in the General Fund. Thus, not having a positive fund balance
has placed many school districts in fiscal stress. If the districts fail to eliminate their deficits,
their financial affairs will be intervened upon directly by the state under the Local Government
and School District Fiscal Accountability Act (PA 4 of 2011). PA 4 supplants the Public Act 72
(PA 72) of 1990 under which State officials are authorized to intervene in the financial affairs of
local governments that experience serious financial problems. It expands the authority of
emergency managers over academic matters, contracts and agreements, and other district
responsibilities.
It is a very demanding task for a deficit district to prepare the required components of the
elimination plan, especially if there is confusion about what caused its deficits. Knowing the
underlying factors would give insight to district management if there are signs of more deficits in
the near future. Being able to detect deficits ahead would enable districts to take preventive
actions. Not only should deficit districts take preventive actions, other financially well-managed
districts also need to do the same if they want to avoid deficits and not be in fiscal stress. Being
in fiscal stress places a district in a vulnerable position. Thus, an understanding of the
relationship between fiscal stress and its underlying factors should be of interest to all districts
and state policy makers.

3

1.2

Statement of the Problem
While the Michigan Department of Education (MDE) requires all school districts to

maintain a positive fund balance, the number of districts that end their fiscal year in deficit keeps
on increasing. Before these districts actually reported deficit in their fund balance, they must
have been experiencing growing fiscal stress, but many were unable to rectify the condition.
Local districts need adequate and reliable tools to anticipate instability in their fund balance.
There is little research on the association of districts’ fund balances and their budgeting and other
policy and community characteristics. Hence, it is essential to develop a model which can
identify factors that associated with the variation of fund balance and thus will be helpful to
school districts management to keep track with their fiscal status and prepare for any
shortcomings.

1.3

Research Questions
The focus of this study is to examine the indicators associated with fiscal stress in

Michigan public school districts. This study is designed to answer the following research
questions:
1.

How has the financial profile of Michigan school districts changed over time?
(a) How has the number of fiscally stressed and not fiscally stressed districts changed
over time?
(b) To what extent have individual districts moved in and out of fiscal stress?

2.

What are the indicators associated with districts’ fund balances?
(a)

Are the indicators within or beyond local district control?

(b)

Is the model developed from the above indicators stable over time?
4

Financial profiles vary across districts. In the past ten years, many districts have
maintained a good financial status, while some have fallen into deficit. With the current
recession that has forced the state to make cuts in K-12 education funding, more school districts
may fall into deficit. As captured in the first question, this study is interested to find out how the
number of public school districts in financial difficulties has changed over time.
The concern over fiscal stress should not be limited to districts having fund balance
below this level. An examination of the changes in all district’s fund balance that affect fiscal
stress will be useful to districts to know what they need to do to improve if their fund balance
may put them in fiscal stress in the near future, or to continue to operate with little or no
disruption. In order to do so, the factors that associated with the problem must be identified. The
model developed in this study must not only be able to identify the factors that associated with
fiscal stress accurately, it must be stable over time. This would help school districts to enhance
their planning and allow more options at an early stage. Thus, the second question of this study
explores the factors that account for school district’s fiscal stress by identifying the financial
indicators that associated with fund balance, which is inversely related to fiscal stress, and the
accuracy and stability of the model over time.

1.4

Significance of Study
School district management needs to have some basic conceptual and empirical tools in

order to identify factors associated with fiscal stress in their decision making. Identifying the
factors could be helpful in preventing financial deficits by providing information on how the
factors are related to the stress affect a district’s spending. Based on the information of the
financial fiscal condition provided by the model developed in this study, it is hoped that an
5

accurate judgment of a district’s fiscal stress will benefit decision makers at the district and also
state level to adjust the level of spending in a more timely and less disruptive manner for all
districts and not only for deficit districts.

6

2
LITERATURE REVIEW

The primary goals of Proposal A were to reduce property taxes and reduce the funding
disparities across school districts. There is a substantial research literature on Proposal A and its
impacts. This study limits its scope to the literature on the sources and distribution of revenue
under Proposal A, identifying a number of issues that have arisen from the reform and,
specifically the growing number of districts that confront financial deficits.
This section also reviews research on fiscal stress and methods used to analyze it.
Previous studies have proposed wide range measures of fiscal stress for local governments and
school districts. However, none of them has identified a generally accepted standard model - that
can fit the circumstances of school districts in other states or settings.

2.1

Proposal A
Proposal A was approved by voter referendum and was implemented in 1994, and it

continues to provide the basic framework for school finance in Michigan’s K-12 public schools.
Proposal A introduced a new foundation-allowance approach to school funding that replaced the
former district power equalization (DPE) plan. The new approach uses a mixture of state and
local taxes to fund the K-12 educational system, but with more reliance on state taxes. Local
school districts lost most control over the revenue available for their schools’ operation, and this
made Michigan school finance system highly centralized at the state level.
Under the reform, revenues come from a variety of sources. Table 2.1 summarizes the
7

changes in the revenue sources enacted when Proposal A was implemented. While sharply
reducing local property taxes, the Proposal A system relies on revenues from sales and a variety
of other taxes which are earmarked for the state’s School Aid Fund (SAF) to finance foundation
allowances. New revenue sources that are directly deposited to the SAF now include the twopercentage point increase in the sales tax from 4% to 6%, the 50 cent per package increase in the
cigarette tax rate, the 0.75 percent real estate transfer tax, the 14.4 percent of individual income
tax revenues (increased to 23.0 percent in 1997), and the 6-mill state education tax levied on
property. A percentage of state income tax collections is also earmarked to the SAF.
Table 2.1

SAF Revenue Sources for K-12 Education in Michigan
Before and After Proposal A

Revenue Source
Sales tax

Before Proposal A
60% of proceeds from the 4%
rate

Use tax
Income tax
Real estate transfer tax
Cigarette tax
$0.02 of the $0.25 tax
(per pack)
Tax on other tobacco
products
Liquor excise tax
Revenue from the 4% tax
Lottery
Net revenue
State tax on all
property
Local homestead
34 mills (average)
property tax
Local non-homestead 34 mills (average)
property tax

After Proposal A
60% from the 4% rate and 100%
from the 2% increase
All revenue from the 2% increase
14.4% of collections from the 4.4%
rate (down from 4.6%)
All revenue from the 0.75% tax
63.4% of proceeds from the $0.75
tax
Proceeds of the 16% tax (on
wholesale price)
Revenue from the 4% tax
Net revenue
6 mills
0
18 mills

Source: Arsen and Plank (2003). Michigan school finance under Proposal A; State control,
local consequences. The Education Policy Center at Michigan State University.
Proposal A established a classified property tax base comprised of homestead and nonhomestead property for the purpose of taxation. Homestead property is property that a taxpayer

8

declares as his or her primary residence. Except qualified agricultural property, all other
property, such as business property, rental housing, or vacation homes, are considered nonhomestead property. As noted, the new finance system included a new statewide uniform
property tax levy of 6 mills on both homestead and non-homestead property, with the revenue
going directly to the SAF. Districts are required to levy 18 mills on non-homestead property and
this is the main source of revenue that is generated locally. Most districts are not allowed to levy
additional mills to finance general operating expenditures. Revenues from the tax on nonhomestead property remain at the local level. These local revenues, which comprise part of the
districts’ foundation allowances, are offset dollar-for-dollar by corresponding decreases in state
foundation aid. Average levies on both types of property were 34 mills before the reform, so the
new provisions represent significant property tax relief for Michigan voters (Cullen & Loeb,
2004).
The new funding system allows districts that were high-spending in 1994 designated
“hold harmless” districts, to levy additional property taxes on homestead property (with local
approval). With this local revenue, these districts are allowed to remain at their spending levels
as before Proposal A. However, they are only allowed to increase their revenue every year by the
amount of the annual increase in the statewide basic foundation grant, which applies to other
districts as well. During the first three years after Proposal A was adopted, local voters could
approve up to three additional mills to supplement the funding provided by the state. However,
since 1997 local school districts are not allowed to seek any enhancement mills unless they join
with other districts in their Intermediate School District (ISD). A majority of the electors in the
ISD must approve the enhancement mills, and the revenues must be equally distributed across
districts on a per-pupil basis.
9

Proposal A has reduced school funding inequities across districts, even though it has not
eliminated them. The foundation allowance in the funding system reduces the fiscal disparities
between school districts by establishing a minimum per pupil spending level for all school
districts. In addition, state assistance is provided for districts lacking sufficient local revenues
for the allowance. The Michigan Department of Treasury (2002) reported that in 1994, the ten
lowest revenue school districts had weighted average per pupil revenues of $3,476, while the ten
highest-revenue school districts had weighted average per pupil revenues of $9,726, nearly three
times more than the ten lowest-revenue school districts. By 2003, the ten lowest revenue school
districts had a weighted average per pupil foundation allowance of $6,700, while the ten highest
revenue districts had a weighted average foundation allowance of $11,389, less than twice as
much as the minimum foundation allowance (see also Arsen & Plank, 2003; Cullen & Loeb,
2004).
The centralization of school funding under Proposal A has given the state the authority to
determine operating funding levels for local school districts. Under such a highly centralized
system, public schools as a whole should benefit from equities in the resources received. School
funding becomes more equitable when the state constrains revenue growth in high-spending
districts and provides more revenue to low-spending districts, and when districts do not rely on
local property for school resources.
2.2

Challenges in Proposal A
Even though some of the objectives of Proposal A have been achieved, the reform is not

without its problems. Proposal A has not yielded gains for all school districts. It has affected
different school districts in different ways. Due to the unique situation of every district, the way

10

the public school revenues are collected and distributed seems not to give the same value of
educational equality or equity to all children across districts (Addonizio, 2003; Arsen & Plank,
2003; Zimmer & Jones, 2005; Izraeli & Murphy, 2007). The system has created issues for local
districts that no longer have much control over their operating revenue and that are reliant on the
state to determine school revenue levels. The next section discusses some of the issues, relating
to the state’s economic crisis, declining enrollment, revenue-cost gap, and capital funding.
2.2.1 State Economic Decline
Centralization directly linked Michigan education finance to the performance of the state
economy. The school funding system created by Proposal A depends on a mixture of the state’s
sales, income, and property taxes, which ultimately depend most on the health of the state
economy. When Michigan was one of the richest states, Proposal A could distribute more
revenues and improve equalization among school districts. However, after the state lost
employment every year from 2000 to 2010 this was no longer a sure thing.
During the initial years following Proposal A’s adoption, state operating revenue growth
was strong, even though per-pupil revenue growth was moderate because of rising student
enrollments. Unfortunately, Michigan’s sustained economic decline has diminished sales and
income tax revenues that school funding relies upon most. After being funded at a higher perpupil foundation allowance since Proposal A was adopted (except in 1999), schools were
stressed financially with a flat funding level of $6,700 in 2002 and 2003. The foundation
allowance was cut by $74 in 2004 but went back to $6,700 in 2005. Then, House Bill 4887 was
signed in 2005 to increase the minimum per-pupil foundation allowance to $6,875 for the 2006
school year. The increase was a relief to schools during this hard time, but it still did not solve
the issue of unstable revenues that rely on the state’s economy and the financial problems facing
11

Michigan schools.
Proposal A promised a minimum per pupil foundation allowance, more equity among
local school districts, and lower property taxes. While taxpayers benefit from the property tax
reduction, it puts limits on local school districts’ ability to generate additional revenues locally,
specifically from the property tax. Indirectly, it has made the “local” revenues “state-controlled”
revenues by largely eliminating local tax millage elections to raise operating revenues. To
replace the reduced property taxes in school revenue sources, Proposal A increased sales tax
revenues to fund school district operating costs. Since sales tax revenue is more cyclically
sensitive than property tax revenue, this substitution put Michigan school funding in a less stable
situation.
Before Proposal A was adopted, earmarked taxes (both constitutional and statutory) used
to be the total state-generated revenue going to the SAF, with a portion coming from the K-12
General Fund allocation (discretionary). Since the reform, the general fund plays a much smaller
role in percentage terms because nothing in Proposal A requires general fund contributions to the
SAF. The composition of the SAF has changed, and the revenues earmarked for the SAF have
been insufficient to fund the state’s funding commitments to local schools. The difference, or
deficit, has been covered by revenues from the state’s general fund budget. Since the single
largest source for the fund comes from income tax revenues, the reduced income tax after the
implementation of Proposal A, even only by 0.2 percent (down from 4.6% to 4.4%), affected the
growth and the stability of Michigan education funding. In addition, the current recession has
reduced the portion of income tax revenues.
According to a report of the Citizens Research Council of Michigan (CRC, 2010), the

12

reduction in state dedicated funding in 2009 was partially offset by the availability of temporary
funding provided through the 2009 federal “stimulus” legislation. Significant levels of nonrecurring federal monies were used in 2009 and 2010. Michigan receive additional federal
resources, through another new “stimulus” program, to support the 2011 SAF budget. The report
mentioned budget cutbacks, including a 2009/10 per-pupil reduction of $154, which also raises
important questions regarding the revenue system Michigan uses to support its K-12 education.
Even though the report emphasized that state and local operating revenue declines in 2009 and
2010 school years would not be fully reflected in the level of education services delivered, the
cutbacks might spread cost-cutting in school districts, which in return, affect the service
provided.
The CRC Report (2010) also stated that the fiscal effects of Michigan’s heavy reliance on
substantial amounts of one-time and temporary funding are that they produce revenue “cliffs” in
some years and bring about an imbalance between current resources and current spending. As a
result, Michigan taxpayers have been increasingly concerned about public schools, which are
termed as being “under-funded”, and about the spending pressures that might have outpaced the
resources available. The new funding structure under Proposal A had allowed Michigan to
support current per-pupil expenditure levels that were above the national average, but it is
evident that Michigan’s support for K-12 education is now declining. Since 2007, per pupil
expenditure levels have fallen to the national average (The National Center for Education
Statistics).
Even though revenue stability is often desired, the stability and income elasticity of
revenues that fund Michigan public schools can no longer promise equalization in Michigan
school finance. Since the major source of revenue for school funds under Proposal A largely
13

depends on the state income and sales taxes, the reform makes school funding more sensitive to
state economic conditions. The pace of per pupil revenue growth under Proposal A, which
initially improved equalization among school districts, has slowed in recent years, and the
revenues earmarked for the SAF have declined due to the economic downturn (Arsen & Ni,
2010).
2.2.2

Declining enrollment
Under Proposal A, the amount of money that the state allocates to each school district

depends on two main factors: the value of the district’s per-pupil foundation allowance, and the
number of pupils enrolled in the district’s schools. The revenues available to the school district
increase when the per-pupil foundation allowance increases, or when enrollment increases. Per
pupil foundation allowances have increased in all districts since 1994, but some districts have
enjoyed large increases while others have not. Meanwhile, enrollment has increased in some
school districts, but fallen in others. Thus, the revenues allocated to school districts do not
increase (or decrease) uniformly. Besides the per-pupil foundation allowance, this is due to the
uneven changes in school enrollments across districts.
Arsen and Plank (2003) note that while analyses of Proposal A have generally focused on
changes in the value of foundation allowances, enrollment changes have had equally large effects
on district revenues. Their findings displayed the interaction effects between changes in
enrollments and changes in the per pupil foundation allowance and total revenues for selected
districts. From 1994 to 2001, urban districts received mid-range per pupil funding increases since
they fell in the middle range of Michigan districts. Yet, even during the initial stage, declining
enrollment in central city districts produced declining total foundation revenues. Meanwhile,
high-income suburbs enjoyed the fastest revenue growth, despite their small increases in per14

pupil funding, because they experienced the most rapid enrollment growth.
Growing enrollment districts have benefited from Proposal A’s revenue distribution,
while declining or static districts have fared less well. People movement from one place to
another causes some districts gain enrollment, while some lose enrollment. Besides this, the
Michigan school of choice policy also plays a role in influencing district enrollment. This policy
is designed to allow local school districts to enroll non-resident students and count them in
membership without having to obtain approval from the district of residence, and permits local
school districts to enroll students who reside in other local school districts within the same
intermediate school district (www.michigan.gov/mde). This practice removes resources of
funding from the losing enrollment districts, though it helps raise revenues in preferred districts.
A study by Arsen and Ni (2010) found that the differences in enrollment have generated
inter-group variations in total revenue decline since 2002. They stated that between 2002 and
2008, the decline in real foundation revenue in Michigan’s central city districts was nearly four
times the rate of decline in high-income suburbs. Revenues decrease much faster than costs,
specifically in the current economic situation. Some costs are fixed in the short run, and districts
or schools cannot simply reduce them as much or as fast as revenue declines. Even though
declining enrollment eases fiscal pressures in one sense, because it diminishes the number of
students by which the SAF has to be divided, schools with declining enrollment suffer from
being under-funded due to less revenue received.
Proposal A has no provision to help districts where enrollment is declining, and local
voters have no option to ask for additional millage to offset revenue lost due to the enrollment
drop. As a result, the funding system makes it harder for schools to provide an adequate
15

education for the students who remain enrolled. Addonizio (2003) concluded that urban districts
with high costs and high proportions of children from low-income families appear to be
substantially underfunded. This problem is intensified by declining student enrollments in these
districts. If school enrollment and the number of high cost students are evenly distributed across
the state, the foundation allowance received would have given even revenues to all school
districts.
2.2.3

Revenue-cost gap
After Proposal A was implemented, the state has increased (or decreased) every district’s

per-pupil foundation grant (with minor exceptions) by the same dollar amount since 2001. Until
today, three-fourths of all Michigan school districts now receive the same per pupil foundation
allowance, while the remaining districts receive somewhat more. However, when looking across
districts, not all districts incur the same per pupil costs in delivering education to different types
of students. Every district is unique in its needs and problems, but the foundation allowance does
not fully take into account the differential cost of educating students, for instance, students
whose English is their second language (ELL), high school students compared to elementary
school students, special education students, and students who come from a low social economic
status family and who live in poverty. Such high-cost students are disproportionately found in
urban schools, where the general consumer prices are also relatively high. Categorical grants
offset only some but not all of the higher costs for special education, ELL, and low income
students. Hence, the revenues districts receive should at least take into account the differences in
the costs of the education provided to their students.
The foundation grant under Proposal A is awarded at the district level, so the foundation
amount is the same for each school within a district, no matter how different the costs of the
16

schools are. Thus, this creates a gap between district’s revenues and costs, specifically in districts
with more high-cost students. Other than the same foundation amount that results in a gap
between revenues and costs, the reduced amount of the categorical grants under Proposal A also
makes the situation worse. Categorical grants are provided to schools by the state or federal
government and may be spent only to pay for services needed by, for instance, low-income
students and students with disabilities. The reform reduced the number of categorical grant
programs in Michigan, which further reduced the total revenue received by districts. Since the
costs to provide education for high-cost students who are mostly grouped in the categorical grant
programs are higher in some parts of the state, the gap is even larger in some districts. The
difference of revenues received and a faster increase of the costs to deliver education,
particularly in recession, has forced many schools not to cater to high-cost students. Schools with
high concentrations of them are left with daunting responsibilities.
Addonizio (2003) mentioned that the adequacy of educational resources in high-cost and
high-poverty districts has not been carefully examined since Proposal A was adopted. He was the
first to examine the issue of adequacy and the costs related to it in an empirical way since
Proposal A was adopted. He used data of school spending, achievement, and demographic from
thirty Michigan urban school districts to determine the possible costs of an adequate education.
Using an education cost formula, he displayed dramatic differences in costs when two relatively
similar districts (Kalamazoo and Ypsilanti) were identified as the benchmark districts. He found
educational insignificance of reaching equity benchmarks that do not reflect inter-district
differences in educational costs, needs, and economic efficiency. Even though this paper does
not discuss education equity and adequacy, it is evident that districts that have to pay higher
costs for educational needs are being underfunded when compared to benchmark districts. One
17

could argue that being underfunded does not mean a district would be not efficient in delivering
education, but Addonizio’s findings on MEAP test are clear.
2.2.4

Capital Funding
While Proposal A has greatly centralized operational funding for Michigan’s K-12

schools, it leaves the responsibility of capital funding to local districts. Proposal A did not
address funding for capital facilities. Expenditures for school capital continue to rely entirely on
local property tax revenues.
Typically, public school facilities in most states are funded with funds borrowed from the
sale of general obligation bonds that are issued by local districts (subject to voter approval),
where voters approve a certain increase in local property tax levies, for a specified period, to pay
for the facilities. Under Proposal A, school facilities in Michigan remain a local responsibility, as
it used to be before the reform. They are funded with distinct property tax levies from the mills
assessed to fund school operations, and local districts have wide discretion to set these tax rates
(millage). The tax rate required to generate funds that are needed by a district depends primarily
on the per-pupil value of taxable property in the local district. Hence, property-wealthy districts
could generate more revenue with relatively low tax rates, whereas property-poor districts must
tax themselves at much higher rates to raise equivalent funds.
Arsen and Davis (2005) found that the inequities in capital funding are greater than
inequities in operational funding ever were, because of the difference between property-wealthy
and property-poor districts. High-spending districts are able to fund their capital spending for
better school resources to educate their students, while low-spending districts just remain in their
old building structures, which then need more operating budgets for maintenance expenditures.

18

Old school facilities may require renovation for the purpose of increasing the number of students
and for their safety, health, accessibility for the disabled, and a conducive learning environment.
Due to the incapability to issue school bond loans in property-poor districts, schools in these
districts would only be able to repair their existing facilities, which then pull the money out from
their operating budgets.

2.3 Financial Consequences of Proposal A
In general, Proposal A has brought positive changes into the Michigan’s school finance
system. However, it has affected different school districts in different ways. Some districts are
better off, while others are not. It is evident that the growth of both total and per-pupil real
revenue for Michigan school districts has slowed since the passage of Proposal A, even though
the total revenues available have increased. The state’s economic decline has indeed depressed
the SAF revenues and will continue to do so in the current recession. Since 2002, annual
increases in per-pupil foundation allowances have not kept pace with inflation, and this has
forced school districts to find their best way to survive and take the consequences. One
alternative is to improve efficiency. The challenge is especially critical for those districts that
already confront financial difficulties as a result of Proposal A.
When education revenues do not meet the rising operating costs that school districts face,
they must go through budgetary changes and initiatives in managing a more equitable
distribution of school funding to meet the challenges. This becomes more complex when revenue
decisions are set centrally, but nearly all spending determinations remain at the local level under
the current reform. Being unable to control the revenues received, what school districts can
19

attempt to do is control their spending. School districts are pressured to make changes in their
budgets in order to restore their financial stability. The task is even harder for those districts that
are already in deficit. All districts try to avoid having operating deficits in their General Fund at
the end of the fiscal year because it is an indicator used by the Michigan Department of
Education (MDE). Under current MDE procedures, as soon as a district discovers a deficit fund
balance, it should notify MDE, and MDE will then notify the district of its obligation to submit a
deficit elimination plan (DEP) to MDE.
Until 2002, Proposal A led most school districts to improve their financial positions as
measured by their fund balances. Yet for the school fiscal year that ended June 30, 2010, there
were 43 districts (36 traditional school districts and 7 public school academies) with operating
deficits (Michigan Department of Education, 2011). Since additional revenue is no longer a sure
thing in the current recession, not only deficit districts but also other districts are under pressure
to keep their fund balances in good shape. It is important to have a better understanding of the
causes of the deficits so that school districts can sustain their fiscal health. Hence, districts
should make adjustments to strengthen their financial condition and avoid falling into fiscal
stress.
2.4

Fiscal Stress

2.4.1

Definition
There are various ways to define fiscal stress1. Most definitions describe what the entity

experiences as a result of financial problems (Trussel and Patrick, 2009). Fiscal stress, as defined

1

Fiscal stress and fiscal distress are used interchangeably in this literature, in reference to
financial difficulties in State and local governments.
20

by Chapman (1998), occurs when there is a change in one set of several variables, but there are
no offsetting changes in other variables. The variables include changes in revenue flows, changes
in citizen tastes and preferences, changes in demographic characteristics of the jurisdiction, or
changes in responsibilities of the jurisdiction. In any one of these variables, stress occurs because
either changes in revenues received or expenditures demanded do not offset. Stress can occur
through a slowly or quickly worsening condition and can create long-run and short-run pressures,
or both at the same time. Chapman implied that the effects of stress must be examined in a longrun framework so that the complexity of both the stress and the responses to it can be
understood.
Bradbury (1982) defined fiscal stress by distinguishing between “budgetary fiscal
distress” and “citizen fiscal distress”. She associated budgetary fiscal distress with short-term
problems for local governments in balancing their budget. The problems are indicated by the
locality’s surplus or deficit as a fraction of its budget. An alternative indicator of budgetary
distress is the locality’s interest on general debt relative to general revenues, which would
eventually show as high debt-related costs and generate an unbalanced budget. Her concept of
citizen fiscal distress refers to longer-term difficulties for a local government in providing
adequate local public services at a reasonable cost to its citizens. This fiscal distress is indicated
by a high tax burden or high costs for basic services for local citizens.
In charitable nonprofit organizations, Tuckman and Chang (1991) referred to fiscal stress
as financial vulnerability and defined it as a constant decline in program spending. A nonprofit
organization is financially vulnerable if it is likely to reduce its service offerings immediately
when it experiences financial distress. The distress might be an economic downturn or loss of a
major donor. Their model to examine the financial vulnerability of this type of organization
21

includes four indicators: equity ratio, revenue concentration, administrative cost ratio, and
operating margins.
In the government sector literature, there are also alternative definitions of fiscal distress.
The U.S. Government Accountability Office (GAO) defines a fiscally distressed local
government as one in which citizens and businesses bear substantially higher tax burdens in
order to obtain levels of public services. Different communities tax their citizens and businesses
at different levels to obtain similar public services. This happens because neither fiscal
circumstances nor the needs for public services are uniform across communities. Thus, the
distress makes it harder for poorer communities to provide adequate services on their own.
There is no single definition of fiscal stress. Different researchers and government
entities define it in different ways. Across the nation, states also have their own definitions and
measures in classifying their fiscal stress. As for the state of Michigan, the Michigan Department
of Treasury (MDT) has developed fiscal indicator scores for all local governments in the state
(Michigan Department of Treasury, 2002). The state ranks local governments based on scores: 04 points represent a fiscally neutral position and requires no state action; 5-7 points indicate a
"fiscal watch" condition and a cause for concern; and 8-10 points indicate fiscal stress and being
placed on a watch list for the current and following years, and they receive consideration for
state review. The ratings are based on key factors from nine categories: population growth, real
taxable valuation growth, large real taxable value decrease, General Fund expenditure as a
percent of taxable valuation, General Fund operating deficits, prior General Fund operating
deficits, size of General Fund balance, fund deficits in current or previous years, and general
long term debt as a percent of taxable value. If despite these preventive measures, local
governments, including school districts, fall into a financial crisis, they are then governed by the
22

provisions of the Local Government Fiscal Responsibility Act, commonly known as Public Act
72 (PA 72) of 19902, which was supplanted by PA 4 of 2011. Under PA 4 of 2011, the governor
may appoint a financial manager whose authorities would supersede those of the elected school
board and superintendent.
Even though there is no specific definition by the MDE for fiscally stressed school
districts, it does require school districts to have a positive fund balance in order for them to avoid
falling into deficits. Fund balance, the difference between assets and liabilities, is an important
component to give an early warning if a district is falling into deficits. It measures a district’s
fiscal condition by comparing the liabilities and assets of the district. A positive fund balance
shows means a district has more assets than liabilities and a negative fund balance means there
are more liabilities that assets. A gap between spending obligations and revenues over time,
which has been an issue in Michigan public schools, would cause a structural deficit in school
districts’ General Fund, even when the economy is recovering from recession.
Being distinct from profit organizations, school districts only rely on revenues from local,
state, and federal sources. Raising taxes might help school districts have more revenue sources,
but Proposal A puts limits on property taxes that school districts can generate to fund current
operations. Current operating expenditures, which include all instructional programs and
supporting services, are deducted from the revenues to achieve fund balance. Hence, any changes
in all these must be taken into account in defining the status of a district’s financial condition as
being fiscally stressed or fiscally healthy. Considering the importance of fund balance in
determining school district’s financial status, this study uses this variable to be examined and
2

The Local Government Fiscal Responsibility Act is the primary State statute under which State
officials are authorized to intervene in units of local government that experience serious
financial problems, or financial emergencies.
23

associated with fiscal stress.
2.4.2

Factors Associated with Fiscal Stress
Any entity that faces or will face fiscal problems would be in a better position to monitor

their financial status if they could identify the factors that contribute to fiscal stress. A small set
of studies have identified ways to analyze the factors contributing to fiscal stress for public
school districts (De Luca, 2006; Ammar et al., 2005; Manca et al., 1999) and local governments
(Trussel & Patrick, 2009; Chapman, 2008; Coe, 2008). However, none of these studies has
identified the same measures or benchmark in analyzing fiscal stress.
De Luca (2006) investigated components of a school district that contribute to fiscal
stress for 605 public school districts in Ohio. Her study used Ohio’s three classification of fiscal
stress categories based on an examination of the overall revenue-expenditure ratio and a review
of the degree of compliance with respect to the financial records submitted to the state: fiscal
caution, fiscal watch, and fiscal emergency (Ohio Revised Code Section 3316.03 (A)). She then
grouped districts according to district type based on a “Typology of Ohio School Districts” and
district enrollment size, which was replicated from previous studies by Tamari (1978) and Brown
(1993). The study included eleven ratio variables which were analyzed using a separate stepwise
discriminant analysis to determine factors significantly associated with fiscal stress. The study
found no single set of characteristics for all school districts that were associated with fiscal
stress.
Anticipating that no single indicator or set of indicators was likely to accurately
associated with school districts at risk of future financial crises, Ammar et al. (2005) applied a
financial condition indicator system (FCIS). Their framework used fuzzy rule-based systems
24

(FRBSs) to combine different units of the measures into an overall evaluation of financial
condition. The FCIS developed in this study included four components: short-run financial
condition – to measure a district’s ability to pay its bills and balance the budget without
extraordinary measures; long-run financial condition – to measure a district’s capacity to finance
adequate services over the long run without tax and debt burdens; economic condition – to
reflect the impact of the local economy on the district’s capacity to raise taxes and on the
expenditures needed to reach adequate student performance; and student performance – to
measure whether students are meeting state standards. A total of fifty different measures of these
components were analyzed to develop an early warning system that assesses a school district’s
financial condition. Their study found that a strong short-run and a good long-run financial
condition of some districts result from higher fund balances and liquidity ratios. They
recommended that individual districts use the results of the FCIS for comparisons with districts
of a similar size or need and resource capacity category.
Manca et al. (1999) examined three school districts in California and Nevada regarding
the circumstances they faced before and during a declaration of fiscal insolvency. They also
examined key factors that contributed to regaining fiscal solvency, and statutory provisions in
state law that helped resolve fiscal problems. Their qualitative study identified fourteen common
factors which occurred in the districts, including declining fund balances over a period of three
or more years; overspent revenues in the General Fund; overstated revenue income; the lack of
budget tracking procedures and internal controls; understaffed Business Services; no long-range
fiscal planning; breakdowns in communication; superintendents without a clear understanding of
the financial position of the school districts; and out-of-date board policies and procedures
related to fiscal management. They agreed that modifications in state funding should be made in
25

order to avoid fiscal insolvency and takeover by the state.
Besides school districts, local governments may also experience fiscal stress. Trussel and
Patrick (2009) presented a model incorporating financial risk factors associated with the
likelihood of fiscal stress in local governments. They merged the 1998-2005 files of 2,633 local
governments in Pennsylvania to create a cross-sectional time-series database which they
examined using logistic regression methods. Their framework limited the risk factors of fiscal
stress to four categories: revenue concentration – reliance on intergovernmental revenues and tax
revenues; administrative expenditures – all general purpose expenditures made from the General
Fund; debt usage – total liabilities and total liabilities to revenues ratio; and entity resources –
total revenues and revenue growth. This study found that the risk factors of inter-governmental
revenues, administrative cost ratio, debt level, and revenue growth are significantly related to the
probability of distress with the predicted signs, while others are not statistically significant. Even
though the model they developed can determine the likelihood that a local government will
experience fiscal distress based on a few indicators, it cannot identify the underlying causes of
the distress.
Chapman (2008) mentioned that state and local governments will continue to face
uncertainties to provide adequate services and infrastructure and will become increasingly
constrained unless they have significant tools to stabilize their finances and maintain their fiscal
sustainability in the long run. In his study on state and local fiscal sustainability, he discussed
state and local governments’ current revenues and expenditures; three types of pressures that
affect state and local sustainability - cyclical, structural, and intergovernmental; and three
specific trends that are likely to affect the fiscal future of state and local governments –
medicaid, pensions and retiree health benefits, and infrastructure; and finally he offered three
26

areas of potential solutions – revenue, expenditure, and administrative and management. He used
measures from state and local revenues and expenditures in the years 1977, 1992, and 2004 for
comparison purposes. The author concluded that the fiscal system needs to change in order to
have fiscal sustainability. He warned that state and local fiscal sustainability will disappear if
there are no changes in both revenues and expenditures in the fiscal system. Besides the three
specific trends mentioned, this study ignored another crucial area, education, which very likely
affects the fiscal future of state and local governments.
In preventing local government fiscal stress, states not only need to detect the stress but
also to take necessary actions to remediate it. A survey by Kloha et al. (2005) of state officials in
New Hampshire, Florida, Maryland, North Carolina, Ohio, Pennsylvania, Kentucky, New Jersey,
and New Mexico discussed how states detect distress in local governments and the best practices
that some states use to prevent fiscal emergencies. The survey revealed the system used by the
states to measure a local government’s fiscal condition. All the measures were extracted from the
annual operating budget and financial reports. The results showed how states vary in providing
assistance for local units due to their staff size and the state’s authority. Strong-authority states
intervened by having local governments take corrective action and by increasing their revenues
and/or decreasing costs; weak-authority states could only recommend changes; and no-authority
states did not intervene in local financial affairs. The study found that states should have a
system that detects fiscal distress, assists identified troubled localities, and strongly intervenes if
a crisis situation arises in order to prevent fiscal stress in local governments. However, in
establishing such a system, the author emphasized that states should take into account what they
use to measure local governments’ fiscal condition.
The literature reviewed here provides helpful tools in identifying early warnings of fiscal
27

problems in order to identify and prevent fiscal stress. Some apply similar methods while others
are different. However, none of the studies suggests a standard model that could be used by all
local governments and school districts. The models may or may not be applicable to other local
governments or school districts. The use of one state classification of fiscal stress categories
might fit only school districts in that state. Some states do not have any measures to classify their
school districts if they should be in fiscal stress, and if they do, the measures should not be
limited to the variables used in the studies reported in this literature. There is no single measure
that fully captures the fiscal condition of a local government or school district.
Even though different methods are applied to incorporate different measures in order to
have not only one indicator or set of indicators to identify fiscal stress, many quantitative
methods are particularly sensitive to measurement. Hence, small changes in variable
measurement can lead to large changes in classification, thus classifying some districts in the
wrong category. Other than the figures that we could extract from financial reports, a qualitative
study such as Manca et al. (1999) is also important to give insight into identifying school
districts which are not fiscally healthy. Identifying factors that associated with school districts’
fiscal stress needs further study and empirical work. A comparison of changes in revenue and
expenditure is another way to measure how local governments or school districts maintain their
fiscal health. The analysis by Chapman (2008) shows the percentage increase or decrease in
revenue and expenditure in local governments. However, the comparisons of local revenues and
expenditures across states at three specific years in a 27 year-period are not quite relevant due to
the unique characteristics of how different states generate different levels of revenue and
expenses, especially over such a long time.
Even though past research highlights the importance of an analysis of fiscal status for
28

business entities, states, local governments, and school districts, few recommend appropriate
solutions for them to get out of fiscal stress. Some studies recommend a state intervention policy
in local districts to help them recover. However, without a proper understanding of the factors
that associated with the stress, it is hard to identify the best strategies to adopt.
Despite using various analyses to develop models to associate fiscal stress, none of the
studies reviewed measure fund balance as a key factor to fiscal stress for school districts or local
governments. School districts universally organize their accounting systems on a fund basis as
defined by the national General Accounting Standards Board. A fund is established for
accountability purposes to demonstrate that financial resources are being used only for permitted
purposes. All school districts in Michigan have a General Fund and other smaller funds to
account for specific activities. A fund balance is increased when fund revenues exceed fund
expenditures for a fiscal period, and is decreased when fund expenditures exceed fund revenues.
A positive fund balance represents a financial resource available to finance expenditures of the
following fiscal period. A deficit fund balance can only be recovered by having revenues exceed
expenditures in the following fiscal period.
The reviewed studies did make use of fund balance in examining fiscal stress, but not as a
factor being affected by other variables. The funding of Michigan public school districts depends
on student enrollment, but no study in the literature used this factor as an input into a model of
fiscal stress. De Luca (2006) grouped her sample by size, but each group was composed of a
different number of districts. Michigan public school districts represent a wide range of sizes,
and the difference results in a non-uniform enrollment and different educational needs and
demands. There was no district grouping based on their fund balances per pupil in the studies
discussed here. Determination of an appropriate fund balance is a critical factor in district
29

financial planning and budgeting processes.
The MDE requires all districts to maintain a positive fund balance at the end of the fiscal
year. A school district in Michigan is considered to have a deficit fund balance if its General
Fund balance before reserves is negative, or other funds have negative fund balances that in total
exceed the General Fund balance (http://michigan.gov/mde). In the fiscal year ending June 30,
2005, there were 18 school districts, plus four public school academies, in financial deficit. In
2009, there were 41 school districts with operating deficits, but eight districts successfully
eliminated their deficits in the end of the year. In 2010, from the total of 33 districts that began
the year with deficit, five districts were successful in reducing the deficit to the level indicated in
their deficit elimination plan, three districts were able to reduce their deficit, but not to the extent
contained in their deficit elimination plan, and 25 districts ended the year with a greater deficit.
In addition, ten districts began the fiscal year with a positive fund balance, but incurred a deficit
during the year. This made a total of 43 public school districts and public school academies that
ended fiscal year 2010 in deficits. Local districts should not wait until their fund balance falls
into deficit to know they are in fiscal stress. They should be able to detect the stress before that
by examining the variations in their fund balance.
Since Michigan school districts differ in their size and revenue received, fund balance as
a share of total expenditures is used as a standardized measure in this study. Total expenditures is
taken into account because it influences the amount of school district fund balance. School
districts receive fund for only eleven months in a year but have to pay for a twelve-month
expenses. Thus, they should have at least one twelfth or 8% of fund balance as a percentage of
the total expenditures to pay for the twelfth month operations. For the purpose of this study,
fiscal stress is defined as the inability of a Michigan public school district to maintain a sufficient
30

fund balance (8% of fund balance as a share of total expenditures) in the General Fund at the end
of the fiscal year.
2.5

Conclusion
Like a number of other states, Michigan continues to struggle with the issue of school

funding in the current recession. The centralization of Michigan’s school finance system under
Proposal A highlights some fundamental problems and makes it harder for school districts to
improve efficiency with limited resources. The central role of student enrollment in state aid
appears to have aggravated the financial burden of declining enrollment districts. The basis for
distributing revenues to school districts does not reflect the actual cost of educating different
students. Even though less enrollment reduces the number of students for whom the State Aid
Fund has to be divided, it contributes less revenues to these districts and makes them suffer from
being under-funded. Expenditures must be made, even with less revenue, to continue operating
for the benefit of the children. Districts with these problems will continue to have financial
difficulties.
In the current economy, every school district has been trying to improve its efficiency by
making full use of its revenues and by cutting costs to avoid a negative fund balance and being
classified as a deficit district. The responses might include cost reduction in programs or services
which in the end affect fund balance. It is important to know how school districts with different
levels of fund balances make budgetary changes in response to student enrollment, and what they
need to offset their limited revenue and expenditures to keep them in fiscal health. Being able to
understand the variation in fund balance would place school districts in a better position to
monitor their fiscal status and take preventive actions before it gets worse. It also offers valuable
information to the state government on a district’s financial status. An early warning system
31

could allow decision makers to adjust the funding available to balance their budgets, rather than
wait for deficit districts to have more deficits, and more districts to have deficits.
Previous studies support the need to analyze the fiscal status of business entities, states,
and local governments. Indeed, such forecasting models do exist for most of these areas. Even
though substantial research has found a variety of characteristics that associated with fiscally
stressed entities, there is no single model that is suitable for school districts in general, or for
Michigan districts in particular. There is also reason to doubt the external validity of these
studies to other entities in different places. While various measures have been proposed which
could identify the causes of fiscal stress, fund balance could be a proxy of other risk factors
associated with fiscal stress. The purpose of this study is to identify such factors. It examines the
fiscal status of Michigan school districts based on the level of their fund balance. The association
of fund balance and other factors that are related to districts’ characteristics and management
strategies is investigated. It is hoped that this study will provide insight to Michigan public
school administrators of how to understand the variation of their fund balance and factors
associated with it, which then can be related to their fiscal status. It is also hoped that the
findings and suggestions in this study will be a supplement to district management in
understanding their fiscal status to avoid future problems.

32

3
METHODOLOGY

The concern over districts’ fiscal stress must not be limited to having a positive fund
balance at the end of fiscal year. The underlying factors that bring those districts to the stress
point must be understood and addressed. It is hoped that a model to associate the likelihood of
fiscal stress in Michigan school districts can be developed from the factors identified in this
study. This section describes the research questions, the conceptual framework and hypotheses,
data sources, variables, and analyses of the study.

3.1

Research Questions
This study is designed to answer the following research questions:

1.

How has the financial profile of Michigan school districts changed over time?
(a) How has the number of fiscally stressed and not fiscally stressed districts changed
over time?
(b) To what extent have individual districts moved in and out of fiscal stress?

2.

What are the indicators associated with districts’ fund balances?
(a)

Are the indicators within or beyond local district control?

(b)

Is the model developed from the above indicators stable over time?

It is useful to envision two sets of factors that affect districts’ financial positions. Some of
the factors are subject to district-level management, while others depend on district
characteristics or circumstances that are largely outside the district’s control. For example, pupilteacher ratio, average teacher salary, and business and administration expenses are factors that
33

are subject to district-level management, whereas the enrollment, foundation allowance, number
of high-cost students, number of economically disadvantaged students, and district property
wealth are factors that are outside the district’s control.
3.2

Data Sources
The Michigan Department of Education (MDE) requires all public school districts and

charter schools each year to collect and submit data about students, personnel, and individual
schools. Non-financial district-level data, which include enrollment, foundation allowance,
property tax value, pupil-teacher ratio, teachers’ salary, and total business and administrative
expenditure are accessed from the Michigan Department of Education (MDE). For number of
students who receive free and reduced lunch and special education students, data are from the
Michigan Center for Educational Performance and Information (CEPI), which assembles and
maintains the data in the Financial Information Database (FID). Data for number of English
Language Learners and status of community type are accessed from the National Center for
Education Statistics (NCES). The Michigan School Business Officials (MSBO) provides data for
fund balance. Data for 550 local public school districts for fiscal years 2000-01 to 2009-10 are
collected and analyzed in this study.

3.3

Empirical Strategy
This section describes the definition of dependent and independent variables, and the

analyses that will be used in this study to find answers to the research questions.

3.3.1

Definition
For the purpose of this study, fund balance as a share of total expenditures is used as an
34

indicator to explain fiscal stress in school districts. Fund balance is the difference between assets
and liabilities of a fund. A positive fund balance means there are more assets than liabilities and
a negative fund balance means there are more liabilities than assets. Total expenditures are taken
into account in computing district fund balance in this study. This is due to the importance of this
variable in determining the fund balance and it is also a standardized measure for all districts,
which differ in their size and total revenue received. School districts are defined as being in
deficit if they have negative fund balance as a share of total expenditures at the end of fiscal
year.
As for now, districts receive funds for only eleven months, but they must pay for the
expenses of twelve-month operations. This generates a minimum of one-twelfth or 8% of fund
balance as a share of total expenditures that a district should at least have to cover their expenses
for the twelfth month. Thus, district’s status of being in fiscal stress is defined as the inability of
a school district to maintain a sufficient fund balance of 8% as a share of total expenditures at the
end of the fiscal year. Districts that have fund balance more than 8% of total expenditures are
defined as not fiscally stressed. Districts that have negative fund balance (below zero percent of
total expenditures) are classified as deficit districts and are also examined in this study. These
definitions are used to answer research question one. Only district fund balance as a share of
total expenditures is used as the dependent variable to answer research question two, without
classifying them as in fiscal stress or deficit.
3.3.2

Dependent Variable
This study examines district’s fund balance as a share of total expenditures as a

dependent variable. Fund balance (FB*) is used to describe this variable. For each school district,
the dependent variable is computed as:
35

Fund balance (FB*) = (district’s fund balance/General Fund expenditures) 100.
3.3.3

Independent Variables
This study examines factors that might contribute to school district fiscal stress pertaining

to issues of Michigan school funding as discussed in the literature review. The factors that
contribute to a district’s financial condition are categorized as district characteristics - which are
largely beyond local district control - and district management and those which are subject to
local district control. Five main components make up the district characteristics: district size
(enrollment), enrollment changes, concentration of high-cost students (percent of students who
receive free or reduced lunch, percent of special education students, and percent of English
language learners), district per pupil foundation allowance received by the state, community type
(e.g. urban), and district property wealth (taxable value per pupil). The three components of
district management are pupil-teacher ratio, average teacher salary, and business and
administration expenses as a percentage of current operating expenditures (business and
administration/current operating expenditures). The following is the description of each variable:
District enrollment (enroll) – determined by the number of pupils legally enrolled and reported
to the CEPI in the fall every year. Districts with high enrollment receives more revenues from
the state because Michigan’s school funding system makes enrollment a key factor in
determining districts revenues. It is directly related to revenues received by districts as well as
their expenditures. It is unclear, a priori, whether this variable will have a positive or negative
association with the dependent variable.
Percent of enrollment change (%enroll ) – since enrollment has not changed evenly across
districts, declining and many, it is important to know the net effect of if this variable on district’s
36

revenues and expenditures and hence their fund balances. One-year and three-year changes in
enrollment are used in the analyses. A positive association is expected between this variable with
the dependent variable.
The following variables are related to issues of educating high-cost students in different
regions of the state and different levels of district wealth.
Percent of students who receive free or reduced lunch (%FRL) – the percent of students who
are eligible for the National School Lunch program to receive free-or reduced-priced lunch. This
eligibility is determined by students’ family income in relation to the federally established
poverty level and is used as a proxy for students’ socioeconomic status. This variable is expected
to have a negative association with the variation of fund balance.
Percent of special education students (%SpecEd) – the percent of students with disabilities
who are entitled to receive special education services through their local school district under the
federal Individuals with Disabilities Education Act (IDEA). Since special education students cost
more to districts, but not uniformly distributed across districts, it is important to know if there is
an association of this variable with the variation of fund balance. It is expected to have a negative
association with fund balance.
Percent of English language learners (%ELL) – the percent of students whose first language
is not English and who are in the process of learning English under the MDE Title III Program.
Programs for this group of students cost more to districts, but their numbers are not even across
districts. This variable is expected to associate negatively with the dependent variable.
Per pupil foundation allowance (FA) – per pupil payment by the state for general school

37

operations created as part of the Proposal A. The state distributed larger funding increases in
formerly low-revenue districts to narrow the per-pupil funding gap. Although the gap has been
narrowed, there are still gaps between revenues received and the cost of educating students. It is
essential to know if this variable has an influence on the fund balance. This variable is expected
to have a positive association with fund balance.
Community type (urban) – the location of school districts as categorized by the state according
to the population density. Different regions have different regional cost of living. In general,
urban areas tend to have higher cost of living compared to suburban and rural areas. Thus, it is
expected to see a negative association between fund balance and districts that are located in
urban.
Property taxable value per pupil (SEV) – the value used to calculate the property taxes for a
property since Proposal A was adopted. SEV reflects property wealth and the fiscal capacity of a
district. This variable is expected to have a positive association with the dependent variable.
Pupil-teacher ratio (PTR) –defined as a district’s Fall Pupil Count, excluding adult education
participants, divided by the total number of K-12 teachers. This variable reflects how districts try
to reduce their expenditures if the class size is big. It is expected to have a positive association
between this variable and fund balance.
Average teacher salary (salary) – full-time and prorated portions of regular teachers’ salaries
for teaching services provided to pupils, divided by enrollment. This variable is used to see how
districts’ spending on teacher salaries is associated with their fund balance. It is expected to show
a negative association between them.

38

Business and administration expenses as a percentage of current operating expenditures
(BusAdm/COE) – the percentage of the total cost of general administration, school
administration, business services, central services, and other support services from the total of
current operating expenditures. This variable shows how districts manage their expenditures and
is expected to inversely influence the fund balance.
Table 3.1 summarizes the independent variables with the expected sign of how they
would affect the dependent variable:
Table 3.1

Fund Balance Indicators
Indicator

1. District enrollment
2. Percent of enrollment change
3. Percent of students who receive free or reduced price lunch
4. Percent of special education students
5. Percent of English language learners
6. Foundation allowance
7. Urban
8. Taxable value
9. Pupil-teacher ratio
10. Average teacher salary
11. Business and administration expenses as a percentage of current
operating expenditures
3.3.4

Expected Sign for
Fund Balance
?
+
+
+
+
-

Analysis
Descriptive statistics for all variables are generated for all school districts to answer

research question number one. Financial data for a ten-year period from fiscal year 2001 to 20103
are evaluated in order to determine how the number and composition of school districts in
financial difficulty has changed over time. Districts are grouped into two groups based on their
3

Academic years refer to calendar years, e.g. academic year 2000-2001 refers to calendar year
2001.
39

fiscal status – fiscally stressed if their fund balance is below 8% and not fiscally stressed if their
fund balance is above 8%.
A dummy variable is created for school district’s location. District locations are defined
as urban or not according to the definitions of the National Center for Education Statistics
(NCES). The definitions are used as guidelines for the urban (Appendix B): 11 - City, Large; 12
- City, Midsize; 13 - City, Small; 21 - Suburb, Large; 22 - Suburb, Midsize; 23 - Suburb, Small;
31 - Town, Fringe; 32 - Town, Distant; 33 - Town, Remote; 41 - Rural, Fringe; 42 - Rural,
Distant; 43 - Rural, Remote. All districts with locale codes that start with 1 or 2 are defined as
urban and districts with locale codes that start with 3 or 4 are defined as not urban
Before analyses are conducted, the data are tested for the assumptions of collinearity,
normality, and homoscedasticity. To address research question number two, data for fiscal years
from 2001 to 2010 are pooled and analyzed using multiple linear regression method. The
dependent variable is regressed against the eleven predictors. For each school district, the
estimated regression equation is:

FB * =

+
+
+

+
+

+
+

+
+

+

+

+
(1)

Three one-year models are also examined at three time periods, fiscal years 2000, 2005,
and 2010 with multiple linear regression analysis. These three points of time seem reasonable,

40

especially in light of the significant increase in the number of deficit school districts from 2000
to 2005 and 2010. The regression equations analyzed for the three years are the same as Equation
(1) but at three different snapshots of time. Because one year enrollment change could be
relatively small, three one-year models are also examined for the three fiscal years with the
predictor of percent of change in enrollment is calculated three years back.
In order to see how accurate and stable the model, the next step is to check the accuracy
of the estimated fund balance compared with the actual fund balance. The estimated coefficient
and intercept values for one year are applied to the actual values of independent variables in
another year. For this purpose, the coefficients and intercept from fiscal year 2000 estimated
model is applied to the fiscal year 2010 equation. The results of this estimated dependent
variable of 2010 are then compared with the actual dependent variable of 2010 data. The purpose
is to see how accurate the model is after ten years. The estimated regression equation is:

̂

=

+

+

+

+

+

+

+
+

+

+

+
+
(2)

41

The difference between the estimated and the actual 2010 amount of stress is calculated
for each district. The actual estimates of 2010 are also compared with the estimated coefficients
of 2005 in order to see how accurate the estimation is if it is after five years. Then, the same
process is done for the estimated data of fiscal year 2005 by using the coefficients and intercept
from fiscal year 2000.

42

4
RESULTS
This chapter presents results of the statistical analyses to answer all the research
questions. It is divided into two main sections. The first section addresses the first research
question with descriptive analyses for all variables in this study for all districts over ten years.
Districts are grouped based on their fund balance and changes in their status of being in fiscal
stress. The second section of this chapter presents the pooled regression analyses results to
answer the second research question. Cross-section regression analyses are also examined and
compared at three snapshots of time. Then, actual and estimated data values are compared for
differences to check for model stability. Charter and private schools are excluded from the
original data.

4.1

Descriptive Analysis
This section provides answers to the research question 1:

How has the financial profile of school districts changed over time?
a.

How has the number of fiscally stressed and not fiscally stressed districts changed over
time?

b.

To what extent have individual districts moved in and out of fiscal stress?
Tables 4.1and 4.2 provide descriptive information for all variables. The fund balance

variable is measured as a share of the General Fund expenditures. Since school districts differ in
their size and total revenue they receive, fund balance as a percentage of total expenditures
represents a standardized measure. The statistics show a large range between the minimum and
43

the maximum fund balance throughout the ten years, ranging from a low of -48.65% to a high of
313.86%, with mean at 18.26%. Enrollment has a big gap from a low of 2 students to a high of
164,506 students. Of the three indicators of high-cost students, percent of students who receive
free and reduced lunch has the largest variance from a minimum of 0 to 100% and a mean of
35.59.
Table 4.1

Descriptive Statistics for All Variables

Log Fund balance/expenditures
Fund balance/expenditures
Enrollment
% of change in enrollment
% of free and reduced lunch
% of special education
% of English Language Learners
Foundation allowance/pupil
Pupil-teacher ratio
Business admin.
expenditure/COE
Average salary
Property taxable value/pupil

Minimum
0.77
-48.65
2
-70.00
0
0
0
6000
2
2.68

Maximum
2.56
313.86
164506
187.50
100.00
40.00
69.70
15876
43
31.84

Mean
1.82
18.26
2929.51
-.92
35.59
12.80
1.56
7188.87
20.59
13.09

SD
0.11
24.32
6547.63
6.76
19.25
4.26
4.82
937.39
3.70
2.66

16974
26992

96215
20024051

52029.86
257110.40

8575.65
643704.45

Based on the definitions of the NCES, all districts with locale codes that start with 1 or 2
are grouped as urban and districts with locale codes that start with 3 or 4 are grouped as not
urban. From the total of 550 school districts examined, 170 districts (30.9%) are located in urban
and 380 districts (69.1%) are not located in urban.
Table 4.2

Descriptive Statistics for Urban
Frequency

Urban
Not Urban
Total

Percent

Valid Percent

170
380
550

30.9
69.1
100.0

30.9
69.1
100.0

44

Cumulative
Percent
30.9
100.0

4.1.1

Trends in Fund Balance4
Total expenditure is a crucial part that determines fund balance of a school district at the

end of fiscal year. As the number of Michigan public school districts with negative fund balance
has been increasing for the past few years, it is good to know how their fund balance as a
percentage of total expenditures vary in the last ten years. Table 4.3 displays the comparisons of
fund balance (as a percentage of total expenditures) in all districts from fiscal year 2001 to 2010.
Table 4.3

District Fund Balance as a Percentage of Total Expenditures, 2001-2010
N

2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

550
550
550
550
550
548
549
549
550
550

Minimum

Maximum

Mean

SD

-8.40
-6.00
-7.87
-12.73
-15.89
-31.16
-43.73
-48.65
-39.57
-38.24

313.86
218.15
238.34
252.06
223.95
187.35
218.30
259.68
243.36
248.54

20.10
21.10
20.56
19.49
18.01
16.48
16.96
17.55
16.58
15.77

26.74
24.47
24.14
23.00
21.61
21.50
23.55
25.81
25.31
25.96

Over the ten-year span, fund balance as a percentage of General Fund total expenditures
ranged from a low of -6%(2002) to a high of 313.86% (2001). The mean range is between 15.77
(2010) and 21.10 (2002). The mean of fund balance has generally declined since fiscal year
2003, the year that marked the beginning of a sustained slowdown in state foundation allowance
growth.
Table 4.4 displays the distribution of school districts in five groups defined by their fund
balance. The first group in the table represents school districts with negative fund balance. The
4

Fund balance as a percentage of total expenditures
45

number of deficit districts declined between 2001 and 2003 and increased thereafter. This group
experienced the largest increase in the number of districts (26 districts) in the ten years. Whereas,
districts with fund balances greater than 24% of spending had the largest decrease. This group
had declined by 55 districts in those years. Districts with fund balances of 0 – 8% of total
expenditures had the second largest increase. This group had an increase of 19 districts in ten
years. In general, more school districts are having a declining fund balance. Looking at the
highest group (> 24%), not all 30 districts that had a decline in their fund balance in 2004 went
down to the next lower group (16 – 24%). This group only had an increase of 22 districts. Even
though it is not known which districts made the new number in this group, it shows that some of
the districts from the highest group had much lower fund balance and went further down to the
other lower groups. The same movement can also be seen in fiscal years 2006, 2009, and 2010.
Table 4.4 Distribution of Districts by Fund Balance as a Percentage of Total Expenditures
Year

Number of Districts
< 0%
Freq.
%

2001
10
2002
9
2003
5
2004
9
2005
13
2006
19
2007
21
2008
20
2009
30
2010
36
Increase 26
(decrease)
in 10 years

1.8
1.6
0.9
1.6
2.4
3.5
3.8
3.6
5.5
6.5
4.7

0 – 8%
Freq.
%

8 – 16%
Freq.
%

16 – 24%
Freq.
%

> 24%
Freq.
%

120
101
105
104
121
140
131
126
132
139
19

182
171
182
187
187
199
196
198
202
196
14

108
129
115
137
140
112
127
119
107
104
(4)

130
140
143
113
89
78
74
86
79
75
(55)

21.8
18.4
19.1
18.9
22.0
25.5
23.9
23.0
24.0
25.3
3.5

46

33.1
31.1
33.1
34.0
34.0
36.3
35.7
36.1
36.7
35.6
2.5

19.6
23.5
20.9
24.9
25.5
20.4
23.1
21.7
19.5
18.9
(0.7)

23.6
25.5
26.0
20.5
16.2
14.2
13.5
15.7
14.4
13.6
(10.0)

4.1.2 Changes in District Status
This study defines fiscal stress as a situation where a district’s fund balance falls below
8% of total operating expenditures. The cutoff point is set at 8% because this is the least amount
of fund balance school districts need to have in order to avoid borrowing to cover current
operations. The state disburses school aid to local districts in eleven monthly payments. If they
are to avoid short term borrowing, districts must have fund balance of at least one twelfth or 8%
of total operating expenditures to pay for the twelfth month expenses. Based on this definition,
school districts are grouped in either being in fiscal stress or not in fiscal stress as summarized in
Figure 4.1 and Table 4.5.
Figure 4.1

Number of Fiscally Stressed and Not
Fiscally-Stressed Districts, 2001-2010

For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation.
47

Table 4.5

Number of School Districts in Fiscal Stress, 2001-2010

2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

Stressed
Not Stressed
Frequency
%
% Change Frequency
%
% Change
130
23.6
420
76.4
110
20.0
-15.25
440
80.0
4.71
110
20.0
0.00
440
80.0
0.00
111
20.2
1.00
430
79.8
-0.25
134
24.4
20.79
416
75.6
-5.26
158
28.7
17.62
390
70.9
-6.22
152
27.6
-3.83
397
72.7
2.54
146
26.5
-3.99
403
73.3
0.83
161
29.3
10.57
389
70.7
-3.55
175
31.8
8.53
375
68.2
-3.54

Total
N
550
550
550
550
550
548
549
549
550
550

The number of fiscally-stressed districts increased between 2004 and 2010. The biggest
increase occurred in fiscal year 2005. As the number fiscally stressed districts increased, the
number of districts not in fiscal stress decreased. Even though the magnitude of the increasing
number of fiscally-stressed districts remains modest, the overall trend is worrisome for state and
local education officials.
A further analysis of fiscally stressed districts is displayed in table 4.6. The table also
shows a comparison with districts that have a negative fund balance or deficit. The purpose is to
identify the number of districts that have had below zero fund balance and those that have been
slightly above zero fund balance. This would indicate district’s strategies in managing their
finances when they do not have too much fund left but still manage to avoid deficits.

48

Table 4.6 Districts in Fiscal Stress* and Deficit**, by Number of Years
Number of years

Number of districts
Fiscal stress
Deficit
0
261
499
1
53
10
2
37
16
3
26
4
4
37
6
5
23
6
6
26
3
7
19
2
8
17
4
9
15
0
10
36
0
* where fiscal stress = fund balance/total expenditures < 0.08
** where deficit = fund balance/total expenditures < 0
From the total of 550 school districts, 261 districts or less than half never have fiscal
stress since 2001. One hundred and thirty-six districts or 25% of total districts have been in fiscal
stress for 5 years and more. This group of districts should be on the state watch list. 36 districts
are in a severe situation with ten consecutive years of being in fiscal stress. They have not been
in good fiscal health since 2001. In contrast, only 15 districts have had negative fund balance for
five years and more from fiscal year 2001 to 2010 and 499 districts never have negative fund
balance. It is good to know that more than 90% of total districts never experienced deficits in the
ten years. There are 238 districts with fund balance below 8% which nevertheless still manage to
keep their fund balances from turning negative. For at least once in those ten years, there are 289
districts that have had fund balance below 8%, while there are only 51 districts have had fund
balance below 0%. Out of this total, for five years and more, 136 districts have been below 8%
of fund balance, whereas only 15 districts have been below 0% of fund balance. No district
maintained a negative fund balance for all ten years.

49

Despite the fact that the MDE requires all districts with negative fund balance to enter
into a mandatory deficit reduction program, some districts are still not improving even after few
years having fiscal stress. Table 4.7 lists all of the districts that have been in fiscal stress for ten
years. Clearly, some districts are chronically in precarious financial condition, even for years.
Table 4.7

Districts in Fiscal Stress Continuously from 2001 to 2010

Districts
Hastings Area School District
Bay City School District
Gladstone Area Schools
North Dickinson County School
District
5. Beecher Community School District
6. Ironwood Area Schools
7. Hillsdale Community Public Schools
8. Litchfield Community Schools
9. Hancock Public Schools
10. Portland Public School District
11. Hale Area Schools
12. Jackson Public Schools
13. Grand Rapids City School District
14. Comstock Park Public Schools
15. Hartland Consolidated Schools
16. Armada Area Schools
17. Fraser Public Schools
18. New Haven Community Schools

Districts
19. Mason County Central School District
20. Meridian Public Schools
21. Berkley School District
22. Hazel Park City School District
23. Madison Public Schools
24. Ewen-trout Creek Consolidated
Schools
25. Allendale Public School District
26. Mendon Community School District
27. Durand Area Schools
28. Morrice Area Schools
29. Detroit City School District
30. Allen Park Public Schools
31. Garden City School District
32. Inkster City School District
33. Redford Union School District
34. Wyandotte City School District
35. Woodhaven Public Schools
36. Southgate Community School District

1.
2.
3.
4.

In addition, districts that have had deficits for five years and more should be on the state
watch list. More attention should be focused on those districts that have been in deficit for eight
years since 2001. They could get into more deficits in the coming years if some precautions are
not taken. Table 4.8 lists all the districts that have had deficits for five years and more.

50

Table 4.8

Districts in Deficits from 2001 to 2010, by Number of Years

Number of Years
5

6
7
8

Districts
Hancock Public School
Clintondale Community Schools
Muskegon Heights School District
Vanderbilt Area Schools
Willow Run Community Schools
Inkster City School District
Ironwood Area Schools
Dollar Bay-Tamarack City Area Schools
Garden City School District
Beecher Community School District
Ecorse Public School District
New Haven Community Schools
Madison Public Schools
Ewen-Trout Creek Consolidated Schools
Redford Union School District

I turn now to examine school districts movement of in and out of fiscal stress and
deficits. The movements reported are the number of change in district status, which is divided
into two groups: (1) fiscally stressed districts - report movements from fiscally stressed to not
fiscally stressed and from not fiscally stressed to fiscally stressed and (2) deficit districts – report
movements from positive to negative balance and from negative to positive balance. Table 4.9
displays the number of districts with the number of change in their status.
Table 4.9

Number of District Movement of Fiscal Stress* and Deficit**, 2001-2010

Not fiscally stressed
Fiscally stressed
Deficit not deficit
not fiscally stressed
fiscally stressed
Number
Number
Number
Number
Number
Number
of change of district of change of district of change of district
1
147
1
174
1
25
2
24
2
33
2
1
3
4
3
4
* where fiscal stress = fund balance/total expenditures < 0.08
** where deficit = fund balance/total expenditures < 0

51

Not deficit
Number
of change
1
2

deficit
Number
of district
33
10

The column for fiscal stress shows 147 districts have one movement in their status from
fiscally stressed to not fiscally stressed, 24 districts with 2 movements and 4 districts with 3
movements in the same direction. What these districts have done to move out from fiscal stress
should be followed by other districts that are in the same situation. Whereas, there are 174
districts that have moved once from not fiscally stressed to fiscally stressed, 33 districts have
moved twice and 4 districts have moved three times. These districts, especially districts that have
moved three times, should be watched closely so they will not have fiscal stress again. Four
districts have the most number of movements (3) of in and out of fiscal stress. There is no district
that has experienced moving in and out of fiscal stress more than three times over the ten years.
However, it should be reminded that 36 districts, as presented in Table 4.7, are in fiscal stress for
ten consecutive years and have not moved out from the stress.
Districts that have moved three times from not stress to stress in the ten years and should
be on the watch list not to fall into stress again are Ashley Community Schools, Montabella
Community School District, Walkerville Public Schools, and Posen Cons School District.
Whereas, districts that recovered three times from stress are Fennville Public Schools, Jonesville
Community Schools, Baldwin Community Schools, and Waterford School District. Even though
these districts are able to move out three times from fiscal stress, it should be reminded that they
have also been in fiscal stress for at least three times too.
For deficit districts, 33 districts have moved once from not deficit to deficit and ten
districts move twice in the same direction. These 43 districts should be monitored closely too and
they must do their best to not enter deficits again. Meanwhile, 25 districts have recovered once
and one district has recovered twice from deficit when they moved out from negative to positive
fund balance. Even though these 26 districts were in deficit, they should be models to other
52

districts as how they got out of it. There is no district that has more than two changes in their
status in both movements, but there are 15 districts (Table 4.8) that have been in deficit for five
years and more.

4.2

Regression Analysis
A regression analysis was conducted to address the research question 2:

What are the indicators associated with districts’ fund balances?
(a)

Are the indicators within or beyond local district control?

(b)

Is the model developed from the above indicators stable over time?
The association between the predictor variables and the dependent variable is examined

to give an insight whether it is generated by district management or factors that are beyond local
district control. The dependent variable is fund balance as a percentage of total expenditures
(FB*), which is inversely related to fiscal stress. The predictor variables include eight variables
which are (1) beyond local district control: enrollment (in thousands), percent of change in
enrollment, percent of students who receive free and reduced lunch (FRL), percent of special
education students, percent of English Language Learners (ELL), foundation allowance (in
thousands), urban, and property taxable value per pupil (in thousands); and (2) within local
district control: pupil-teacher ratio, average salary (in thousands), and business administrative
expenditures as a percentage of current operating expenditures (COE). The first test on normality
shows the data are not normally distributed and the presence of outliers. Consequently, all
variables except urban, are transformed into log. A dummy variable is used to identify urban. All
districts in city and suburban are defined as urban and all districts in city and rural are defined as
not urban. The assumptions of normality, collinearity, and homoscedasticity are first tested
53

before regression analysis is conducted.
There are missing financial data for District 23490 Oneida Township School District in
fiscal year 2007 and 2008. Since data for ELL only started in 2002, this variable does not have
any data for fiscal year 2001 for all districts. A mean imputation is applied to replace them with
the mean of the other nine-year data5.
4.2.1 Test of Assumptions
First, the assumption of multicollinearity is tested. The correlations among pairs of the
variables do not show high correlations. The highest correlation is 0.65, which is between urban
and enrollment. Futher, the variance inflation factor (VIF) is used to detect multicollinearity in
the data. The VIF results for multicollinearity test are displayed in Appendix C. The VIF
represents the multiplicative increase in the variance (squared standard error) of the estimator
due to

being correlated with other predictors (Agresti & Finlay, 2009). The threshold of VIF is

three and any VIF values of more than this value is considered as evidence of serious
multicollinearity. Since none of the independent variables has a VIF greater than three, it
indicates that there are no clear multicollinearity problems.
The next assumption tested is a test of normality. A histogram of standardized residuals
shows a roughly normal curve (Appendix C). The histogram has a slightly longer tail on the left
which indicates a not normal distribution. Yet, the normal probability plot shows the values fall
nearly on the straight line. Thus, the normality assumption appears to be satisfied.
The next assumption to test is homoscedasticity. Based on the result of scatter plots, there
is no clear heteroskedasticity presents. This shows that the variance of residual error is constant
5

The results were insensitive to alternative specifications of the imputation.
54

for all values. According to the results, it is generally decided that all assumptions are met and
the regression results could be examined and interpreted as they are.
4.2.2 Pooled Regression Analysis
The data from fiscal year 2001 to 2010 are pooled across districts to run an ordinary least
squares (OLS) regression model to investigate which predictor variables are associated with the
dependent variable. The regression results report two models: (1) without year dummies and (2)
with dummies for years. Both models provide statistically significant estimation of fund balance:
(1) F = 344.963, p < 0.01 with

= .410 and (2) F = 195.310, p < .01 with

= .418. Table 4.10

displays the results.
The first result with no year dummies show a positive association between fund balance
and the predictors of percent of enrollment change, percent of ELL, foundation allowance, and
property taxable value/pupil. However, percent of enrollment change and ELL do not have a
significant association with fund balance. As expected, an increase in percent of enrollment
change, foundation allowance, and property taxable value increase district fund balance. ELL
shows an opposite association with fund balance from what is expected.
There is a negative and significant association of fund balance with the predictors of
enrollment, percent of FRL, special education, urban, pupil teacher ratio, salary, and business
administrative expenditures/COE. Two of the variables that describe high-cost students - percent
of FRL and special education - significantly decrease district fund balance. School districts
located in urban have less fund balance compared with those in rural areas. All of three variables
that related to district management - pupil teacher ratio, salary, and business administrative
expenditures – reduce district fund balance if there is an increase in these numbers. As indicated
55

by pupil-teacher ratio, the result shows that districts have higher fund balance if they have bigger
class size.

Table 4.10

Regression Analysis of District Fund Balances, 2001-2010
B
2.469

SE
.060

B
2.238

SE
.067

-.050**

.006

-.058**

.006

.001

.024

-.013

.024

% free and reduced lunch

-.039**

.004

-.029**

.004

% special education

-.162**

.006

-.167**

.006

.006

.004

.005

.004

Foundation allowance/pupil

.198**

.035

.317**

.040

Urban

-.015**

.003

-.018**

.003

Property taxable value/pupil

.024**

.006

.032**

.006

Pupil-teacher ratio

-.057*

.023

.002

.024

Average salary

-.157**

.027

-.129**

.027

Business admin. expenditure/COE

-.239**

.016

-.247**

.016

dyear1

.031**

.006

dyear2

.036**

.006

dyear3

.028**

.006

dyear4

.025**

.005

dyear5

.020**

.005

dyear6

.008

.005

dyear7

.007

.005

dyear8

.007

.005

dyear9

.001

.005

Constant
Enrollment
% enrollment change

% of English Language Learners

.410
p < .01**

p < .05*

56

.418

Further, dummy variables are added for years with fiscal year 2010 as a reference year.
The first five fiscal years, 2001 – 2005, are significantly associated with fund balance. This
indicates that compared to fiscal year 2010, fund balance in fiscal years from 2001 to 2005 has a
significant increase. Though not significant, the next five fiscal years also have a positive
association with district fund balance compared to fiscal year 2010, but at a declining rate.
After the year dummies are added, percent of enrollment change and pupil teacher ratio
show the opposite association with fund balance. Enrollment change is now negatively
associated with fund balance, while pupil teacher ratio is positively associated although both of
them are not significant. Other predictors that have a positive association with fund balance are
percent of ELL, foundation allowance, and property taxable value. Not like foundation
allowance and property taxable value, ELL was not expected to have a positive association with
fund balance. Besides enrollment change, there is also a negative association between district
fund balance with enrollment, percent of FRL, special education, urban, salary, and business
administrative expenditures.
Of all the variables that are beyond district control, only percent of ELL, foundation
allowance, and property taxable value help increase district fund balance. Fund balance is
negatively associated with other district characteristics that are beyond their control. District
management on the variables that are within their control, which are pupil-teacher ratio, average
salary, and business administrative expenditures, help increase fund balance.

57

4.2.3

Cross-Section Regression Analysis
Three one-year models are also examined at three time periods, fiscal years 2000, 2005,

and 2010. These three points of time seem reasonable, especially in light of the significant
increase in the number of deficit school districts from 2000 to 2005 and 2010. Because one year
enrollment change could be relatively small, the three models are also examined for the same
fiscal years with percent enrollment change calculated over three years. The regression equation
tested for each of the three years is:

=

+

+

+
+

+

+
+

+
+

+
+

+

Table 4.11 displays the results of the regression analysis for fiscal years 2000, 2005, and
2010 using percent enrollment change in three years. For comparison, Table 4.12 presents the
results of regression analysis of the same three fiscal years, but with a one-year change in
enrollment.
All models in both tables provide statistically significant estimation of fund balance at
p<0.05. The variation in fund balance explained by the predictors decreases slightly. They are
38.7% in 2000, 38.5% in 2005, and 29.8% in 2010 with percent enrollment change in three years
and 38.6% in 2000, 38.6% in 2005, and 31.2% in 2010 with percent enrollment change in one
year.

58

Compared with pooled regression analysis, the results in Table 4.11 show consistency.
Enrollment, FRL, special education, ELL, foundation allowance, urban, property taxable value,
and business administrative expenditures/COE show the same direction of association with
district fund balance. However, only enrollment, percent of special education, and business
administrative expenditures/COE are consistent in their significance and direction in associating
with fund balance.
Table 4.11

Regression Analysis of District Fund Balances, 2000, 2005, 2010 (percent
enrollment change in three years)

Constant
Enrollment
% enrollment change
% free and reduced lunch
% special education
% of English Language
Learners
Foundation
allowance/pupil
Urban
Property taxable
value/pupil
Pupil-teacher ratio
Average salary
Business admin./COE

2000
B
(SE)
1.678**
(.302)
-.093*
(.038)
.103
(.070)
-.004
(.016)
-.260**
(.033)

2005
B
(SE)
1.612**
(.283)
-.064
(.037)
-.034
(.073)
-.071**
(.024)
-.285**
(.037)
.024
(.022)
.478*
(.228)
-.061**
(.019)
.023
(.035)
.0004
(.136)
.086
(.159)
-.174*
(.088)
.385

1.240**
(.222)
-.034
(.021)
.089*
(.041)
.793**
(.188)
-1.012**
(.185)
-.495**
(.076)
.387

p < .01**
p < .05*
59

2010
B
(SE)
3.040**
(.309)
-.103**
(.035)
.023
(.059)
-.137**
(.033)
-.118**
(.037)
.021
(.019)
.156
(.215)
-.034
(.018)
.057
(.031)
-.220*
(.108)
-.303*
(.130)
-.371**
(.079)
.298

Besides a not significant association between enrollment and fund balance in 2005, the
result indicates that district size is becoming more relevant in 2010 to determine the amount of
fund balance it has. However, percent enrollment change is not significantly associated with fund
balance in the three years. It has a negative association with fund balance only in fiscal year
2005. Higher enrollment change is associated with higher fund balance in fiscal years 2000 and
2010, but not in 2005.
Except for ELL students, the other two indicators for high-cost students - percent of FRL
and special education – result in a significant lower fund balance if these number of students
increase. Percent of ELL students does not have a significant association with fund balance even
though they are positively associated. This shows this group of students seems less relevant
related to fund balance. Foundation allowance and property taxable value are also becoming not
significant recently even though they seem to have a positive influence on district fund balance.
Urban school districts appear to have lower fund balance than rural school district in all three
years, though being significant only in 2005.
How school districts manage their spending on pupil-teacher ratio, salary, and business
administrative expenditures are found inconsistent in influencing the fund balance they have. In
general, it appears that district management indicators of salary and business administration are
only significant when they have a negative association with fund balance. It shows that the more
they save on salary and business administration, the higher fund they have in the end of fiscal
year. However, district control over class size shows both significant positive and negative
association with fund balance in 2000 and 2010. Bigger class size, as indicated by pupil teacher
ratio, helps district increase their fund balance in 2000 but not in 2010.

60

Table 4.12

Regression Analysis of District Fund Balances, 2000, 2005, 2010 (percent
enrollment change in one year)

Constant
Enrollment
% enrollment change
% free and reduced lunch
% special education
% of English Language
Learners
Foundation
allowance/pupil
Urban
Property taxable
value/pupil
Pupil-teacher ratio
Average salary
Business admin./COE

2000
B
(SE)
1.666**
(.304)
-.093*
(.038)
.114
(.084)
-.004
(.016)
-.260**
(.033)

2005
B
(SE)
1.668**
(.289)
-.061
(.037)
-.063
(.071)
-.071**
(.024)
-.286**
(.037)
.024
(.022)
.483*
(.228)
-.061**
(.019)
.022
(.035)
-.013
(.125)
.087
(.159)
-.175*
(.088)
.386

1.241**
(.224)
-.034
(.021)
.086*
(.041)
.791**
(.191)
-1.004**
(.186)
-.498**
(.075)
.386

2010
B
(SE)
2.830**
(.289)
-.095**
(.034)
.180**
(.053)
-.121**
(.032)
-.107**
(.037)
.017
(.019)
.081
(.207)
-.030
(.018)
.060*
(.030)
-.269*
(.106)
-.282*
(.124)
-.354**
(.018)
.312

p < .01**
p < .05*
Looking at both Table 4.11 and Table 4.12, the magnitude of all coefficients does not
show a consistent increase or decrease, but more predictors are becoming significant lately in
2010. Compared with results that use three-year change in enrollment, a one-year change in
enrollment becomes significant in 2010 and positively associated with district fund balance. This
indicates a higher fund balance with higher enrollment change. On the other hand, enrollment is
still negatively associated with fund balance in all three years and only significant in 2010.
61

Though not significant, percent of ELL is consistently showing a positive association
with fund balance if compared with results with three-year enrollment change and pooled
regression. Percent of FRL, special education, urban, and foundation allowance are consistent in
the direction and significance of association with fund balance. Property taxable value is
becoming significant recently.
District control on pupil-teacher ratio does not show a consistent result. It has a negative
association with fund balance in 2005 and 2010, but only significant in 2010, and a positive and
significant association with fund balance in 2000. District management in reducing salary helps
increase fund balance in 2000 and 2010 but not in 2005. District should pay more attention in
reducing business administrative expenditures since this variable shows a consistent and
significant influence on fund balance.

4.3

Model Stability
Finally, this section examines the stability of the models developed from the cross-

section regression analysis in the three years to estimate fund balance. The actual fund balance
for 2010 is compared with the fund balance estimated from the regression equation for the same
year but functioned with the coefficients from 2005 and 2000. Then, the actual fund balance for
2005 is compared with the fund balance estimated from the regression equation for the same year
but functioned with the coefficients from 2000. This is done with regression model that uses a
one-year enrollment change. Then, paired sample t-test is performed to compare the differences
between the two variables in each pair. Table 4.13 displays the significance of the difference
between the actual and estimated values. The mean difference is the actual fund balance
subtracts the estimated fund balance.
62

Table 4.13

Paired Samples Test

Model
Year
Used Estimated
2000
2000
2005

2005
2010
2010

Paired Differences
Mean
SD
SE
mean
-.0274
.1932
.2036

.1576
.1529
.1430

.0067
.0065
.0061

t
95% Confidence
Interval of the
Difference
Lower Upper
-.0406 -.0142 -4.071
.1804
.2060 29.071
.1917
.2156 33.380

Df

Sig.

549
549
549

.000
.000
.000

From the table, it appears that when coefficients in model 2000 is used in 2005, the
estimated fund balance tends to be higher than the actual fund balance, which is described by
negative mean value. The actual fund balance tends to be higher than the estimated fund balance
when coefficients in models 2000 and 2005 are used in 2010. The results show there are
significant differences in actual and estimated values of fund balance, which indicate that the
models at the specified years are not stable in estimating fund balance five or ten years later. The
confidence interval suggests that we can be 95% confident that the actual fund balances are
within the estimated limits.
Despite the significant differences between the actual and estimated fund balances, these
models provide statistically significant estimation of fund balance in each year, as discussed in
the cross-section regression analysis. It can be concluded that the models are significant to
estimate fund balance in the respective years, but not to estimate fund balance five or ten years
later.

4.4

Conclusion
In conclusion, the analysis of school districts financial profile reveals that in general, the

mean of fund balance (as a percentage of total expenditures) has been declining. From fiscal year
63

2001 to 2010 few districts moved out of fiscal stress and deficits, while there are some districts
that entered fiscal stress and deficits. There are a number of districts that should be on the state
severe list and watch list based on the level of their fund balance.
The investigation of factors associated with fund balance through pooled regression
analysis show that enrollment, percent of students who receive free and reduced lunch, percent of
special education students, foundation allowance, urban, property taxable value, salary, and
business administrative expenditures/COE are significantly associated with fund balance, while
other variables are not significantly associated with fund balance. Although not significant,
district fund balance unexpectedly increases if there is an increase in the number of ELL
students, which is one of the high-cost student indicators. Increase in foundation allowance,
property taxable value, and pupil teacher ratio also helps increase district fund balance.
Enrollment, enrollment change, FRL, special education, salary, and business administrative
expenditures have a negative association with district fund balance. Fund balance of a school
district which is located in urban is lower than those located in rural areas.
In general, district fund balance can be improved by controlling their spending in the
areas that are within their control. Even though lower expenses in district characteristics help
increase their fund balance, these variables are beyond their control. Compared to fiscal year
2010, fund balance in fiscal years from 2001 to 2005 has a significant increase. Three crosssection regression analyses for fiscal years 2000, 2005, and 2010 show consistent results with the
pooled regression analysis. The three-year models are significant to estimate fund balance in the
respective years, but not to estimate fund balance five or ten years later.

64

5
DISCUSSION

The final chapter of this study presents the researcher’s interpretations of the findings.
The following sections in this chapter summarize and restate the entire study, the research
questions, and the results. Interpretations, implications, and recommendations are offered at the
conclusion of this chapter.

5.1

Summary of the Study
Financial difficulties have put many school districts in fiscal stress. Therefore, these

districts must learn ways to improve fiscal health in order to achieve a positive fund balance.
Imbalanced budgets put every school district at risk of a negative balance, so in an attempt to
counteract this growing trend, the Michigan Department of Education (MDE) has made
requirements that all school districts must meet to maintain a positive fund balance. If these are
not met, the district will be classified as a deficit district, and placed into a mandatory deficit
reduction program. Despite the requirements, the number of districts that end their fiscal year in
deficit and or near deficit has increased. It is a very demanding task for local districts to confront
financial pressures because funding is centralized at the state level under the current school
funding system, Proposal A.
Proposal A has given the state the authority to determine operating funding levels for
local school districts. The majority of revenue is raised by the state and funding is distributed
based on a per pupil foundation allowance formula that is annually determined by the state. Since
it was adopted, school funding inequities across districts have been reduced, though they have
65

not been eliminated. While this system has reduced inequities, it has created problems for local
districts through lack of control over their operating revenue and increased reliance on the state
to determine school revenue levels. Factors like economic crisis, declining enrollment, gap in
revenue and cost, and capital funding have worsened the problems. For example, during the
fiscal year that ended June 30, 2010, 43 districts closed their accounts with operating deficits.
The current economy that has forced the state to make cuts in K-12 education funding will lead
to more school districts falling into deficit.
Before school district is required to enter into the mandatory deficit reduction program or
assigned an emergency manager, it is essential to have an early warning system that can alert
district to the indicators associated with the variation in their fund balance. As defined in this
study these indicators either positively or negatively affect fiscal stress. Previous studies do not
reach a single definition for fiscal stress, but this study defines fiscal stress as having a fund
balance below 8% of the total expenditures. Districts must pay for the expenses of twelve-month
operations, and as of now, they receive funds for only eleven months. This established minimum
of 8% or a one-twelfth minimum percent of fund balance as a share of total expenditures covers
their expenses for the twelfth month. Since school districts differ in their size and total revenue
they receive, fund balance as a percentage of total expenditures represents a standardized
measure.
The literature review centered on the conceptual and practical development of Michigan
school funding system through Proposal A, details how problems arise, how fiscal stress has
been defined by different entities, and methods of predicting factors of fiscal stress. Even though
substantial research has found a variety of characteristics of fiscally stressed entities and
provided models to estimate fiscal stress, there is no single model that is suitable for school
66

districts in general, or for Michigan districts in particular. The objectives of this study are to
analyze the financial profile of Michigan public school districts, especially those districts that
have been in deficit since 2001, and to investigate the indicators associated with fund balance.
The indicators can be grouped into two sets of factors, those that are subject to district-level
management, and those that depend on districts characteristics or circumstances that are largely
outside the district’s control. The research questions designed in this study are:
1. How has the financial profile of school districts changed over time?
(a) How has the number of fiscally stressed and not fiscally stressed districts changed
over time?
(b) To what extent have individual districts moved in and out of fiscal stress?
2. What are the indicators associated with districts’ fund balances?
(a)

Are the indicators within or beyond local district control?

(b)

Is the model developed from the above indicators stable over time?

This study utilizes a ten-year data set from fiscal year 2001 to 2010 to answer the
research questions. Data are collected from the Michigan Department of Education (MDE), the
Michigan Center for Educational Performance and Information (CEPI), the National Center for
Education Statistics (NCES), and the Michigan School Business Officials (MSBO). This study is
comprised of 550 public school districts.
Fund balance* (percentage of fund balance as a share of total expenditures) is the
dependent variable examined with eleven independent variables. The independent variables are
(1) districts characteristics which are enrollment, change in enrollment, percent of students who
receive free or reduced lunch (FRL), percent of special education students, percent of English
Language Learners (ELL), district per pupil foundation allowance received by the state, urban,
67

district property taxable value per pupil and (2) district management which is measured by pupilteacher ratio, average teacher salary, and business administration expenses as a percentage of
current operating expenditures (COE). These variables are selected based on their importance in
determining school district fund balance. Dummy variables are created for years to see if the
years have significant associations with fund balance.
The first part of the analysis uses descriptive statistics to answer the first research
question. Districts are divided into two groups based on their fiscal status, which is based on the
percentage of their fund balance as a share of total expenditures. Districts that are below 8% of
fund balance as a percentage of total expenditures are coded as fiscally stressed and districts with
more than 8% of fund balance as a percentage of total expenditures are not fiscally stressed.
Districts are further differentiated by the number of years they have been in stress and the
number of movement of being in and out of fiscal stress. Comparisons are made between the
number of fiscally stressed districts and deficit districts.
The second part of the analysis employs multiple regression analysis to address the
second research questions. A pooled regression model is employed to analyze the data for ten
fiscal years from 2001 to 2010. Since change in enrollment could be relatively small after one
year, the cross-section regression model is employed with change in enrollment is calculated
three years back. Finally, the models are checked for stability to estimate fund balance by
plugging the intercept and coefficients from one year in another year to compare the estimated
and actual fund balance. Intercepts and coefficients from fiscal year 2000 and 2005 are used to
estimate the fund balance in 2010. These estimates are then compared with the actual fund
balance in fiscal year 2010. The purpose is to verify the accuracy of the models from the
previous two years after five and ten years.
68

5.2

Research Findings

Research Question 1
This question explores the descriptive analysis of school districts fund balance for ten
fiscal years from 2001 to 2010. An analysis of the descriptive analysis found the fund balance
among school districts to have a significant difference (-6% to 313.86%). Enrollment, an
indicator which plays an important role in the school funding, shows an extremely varied range
from 2 students (Bois Blanc Pines School District) to 164,506 students (Detroit Public Schools).
The fund balance shows a decline since fiscal year 2003, as the foundation allowance declined
and remained for three years. In ten years, the number of districts with fund balances below than
16% has increased, while the number of districts with fund balances above 16% has decreased.
This indicates an increase in numbers of districts being in fiscal stress from fiscal year 2004 to
2010. The overall health of school districts’ financial status is decreasing, though the magnitude
is relatively small.
During this 10-year period, 289 districts have been in fiscal stress at least once during this
10-year period. From this total, 136 districts have been in fiscal stress for five years or more.
Two hundred and sixty-one districts have never been in fiscal stress, and 36 districts have been
in fiscal stress for ten consecutive years. One hundred and seventy-five districts have moved to
not fiscally stressed status, while 211 districts have moved to fiscally stressed status. To further
explore the status of the school districts, the definition of fiscal stress is changed from 8% to 0%.
Using this criteria, it is found that 15 districts have had fund balance below 0% for five years or
more, and 4 districts should be on the state severe list, as they have been below 0% fund balance
for eight years. Forty -three districts have regressed, moving from positive to negative fund
balance, while 26 districts have progressed after moving from negative to positive fund balance.
69

In summary, 289 districts have been in fiscal stress at least once and 51 districts have had a
negative fund balance at least once during the ten years. It can be concluded that many districts
managed to keep their fund balance just above the zero level but not near 8% or above.
Research Question 2
This question examines the association between district fund balance and factors within
and beyond district control. To answer this question, a pooled multiple regression analysis is
performed. Eight indicator variables are found significant in the association with fund balance:
these consist of: enrollment, percent of FRL, percent of special education students, foundation
allowance, urban, property taxable value, salary, and business administrative expenditures/COE.
Indicators that do not have a significant association with fund balance are percent of enrollment
change, percent of ELL, and pupil teacher ratio do not have a significant association with fund
balance. A positive association is found between fund balance and percent of ELL students,
foundation allowance, property taxable value, and pupil teacher ratio. A negative association is
found between fund balance and enrollment, percent enrollment change, FRL, special education,
urban, teacher’s salary, and business administrative expenditures/COE. Compared to fiscal year
2010, the first five fiscal years (2001-2005) have a significant association with the fund balance.
School districts which are located in urban areas are associated with lower fund balances
compared to those not located in urban areas. One of the high-cost student indicators, percent of
ELL, is found to have a positive association with fund balance. Although the numbers are not
significant, the number of this group of students helps increase district fund balance.
The additional analyses of cross-section regression for fiscal years 2000, 2005, and 2010
pertaining to percent enrollment change is calculated over one year, and includes three years
back to emphasize the change. The results have shown consistency in certain variables in
70

estimating fund balance. School districts located in urban seem to have lower fund balance than
rural school districts. Looking at the factors that are beyond school district control, such as:
percent of ELL students, foundation allowance, and property taxable value, these help school
district increase their fund balance. Enrollment change in one year is becoming more relevant
and significant lately. School districts management of spending on class size, teacher’s salary
and business administrative expenditures is significantly associated with fund balance. However,
pupil teacher ratio shows that bigger class sizes do not reduce district spending, thus does not
help the district increase their fund balance.
The test for model stability for fiscal years 2000 and 2005 shows both models are
significant to estimate fund balance in the respective years, but deviate from the actual balance
when five or ten years has passed.

5.3

Interpretations of the Findings
Based on the findings, there were big amount of gaps in all the variables, showing that

districts differ significantly in their characteristics. District’s fund balance as a percentage of
General Fund total expenditures indicates that some districts have very high deficit, while some
do not have problems in maintaining their fund balance. In general, district fund balance has
been declining gradually for this 10-year time period (fiscal years 2001 to 2010). Districts that
have declining fund balance are increasing at a higher rate than districts with increasing fund
balance.
The statistics indicate that enrollment and high-cost students are not uniformly distributed
across districts. Enrollment as low as 2 students in one school district could affect its fund
balance and fiscal stress compared to a district with a total of 164,506 students. The function of
71

enrollment is undoubtedly important in Michigan school funding to determine the revenue
received from the state every year. However, the district (Detroit Public School District) that had
the most number of students at one point - which has declined by half over these ten years – is
not among the districts that have deficits for more than five years. As supported by the results
(pooled regression) presented in this study, change in enrollment can be thought to be not
significant to influence fund balance.
Looking at the difference between the number of districts in deficit and fiscal stress,
many districts are able to keep their fund balance in the range of 0 – 8%. These districts are
employing good management strategies to cope with the financial difficulties. Yet, with the
uncertain economy and declining foundation allowance in recent years, it is risky for districts to
stay in this zone of fund balance. They must do their best to push their fund balance to a higher
level. Placing these districts on the state watch list, and districts in deficit for more than five
years on the state severe list and need state intervention could be helpful to alert the management
of their fiscal status.
The movement of districts in and out of fiscal stress indicates that the financial
difficulties faced by school districts are still unpredictable. Even though the number of fiscally
stressed districts has been less than not fiscally stressed in the period of this study, the number
has yet been increasing. Experience from districts that have moved from being fiscally stressed
to not fiscally stressed and districts that have never been in fiscal stress at all should be learned
by other districts. Their management strategies should be examples for improvement.
The research findings in this study show that district fund balance is associated with
factors which are within and beyond district control. Based on the pooled regression analysis, the
72

following indicators do not have a significant association with fund balance: percent of
enrollment change, percent of ELL students, and pupil teacher ratio. Though not significant, how
these variables are associated with fund balance must still be considered. Generally, higher
enrollment reflects larger funding received from the state. Although more expenses are incurred
when there are more students, some costs per pupil are saved, which could increase district fund
balance. This study proves that fund balance is inversely associated with enrollment. A theory as
to why these factors are inverted could be that districts have to spend more with higher
enrollment and the spending increases at a higher rate than the funding received. Thus, higher
enrollment is associated with lower fund balance.
When there is a downturn in economy, people will possibly move out from urban areas
and cause enrollment in urban to decline, which subsequently leads to declining fund balance in
urban districts. Change in enrollment is one component that is directly related to enrollment, and
also to fund balance. As stated by the Citizen Research Council of Michigan Report (2010),
because the marginal cost to educate a student is much less than the average cost to educate a
student, a district cannot eliminate costs as quickly as it loses funding when it is losing students,
and similarly, a district does not incur substantial new costs for accepting one (or even a few)
choice students, but it does receive additional state aid for that student.
Both enrollment and high-cost students are not uniformly distributed across districts.
These groups of students do cost more to districts and most of them are concentrated in urban
schools which have a higher cost of living. In an attempt to have a higher fund balance, districts
would begin considering eliminating or consolidating services with their neighbor districts,
especially those related to high-cost students. This study found that two of the high-cost
indicators, free and reduced lunch and special education programs reduce district fund balance.
73

However, percent of ELL is unexpectedly found to help increase district fund balance. The actual
costs of ELL programs vary across districts even with the same services provided. Despite this,
this group of students is positively associated with district fund balance. Districts receive
additional revenue for ELL while they receive per-pupil foundation allowance for the same
students at the same time. This additional revenue would help district maintain their fund
balance. Hence, districts should consider keeping this program, welcome more students and
enjoy the benefits of a higher fund balance.
Property taxable value and foundation allowance per pupil are found to help districts
increase their fund balance. Foundation allowance determines the amount of funding a district
receives. Even though districts are no longer allowed to levy additional mills to finance their
general operating expenditures, revenues from the tax on non-homestead property remain at the
local level. These local revenues comprise part of the district foundation allowance despite being
offset dollar-for-dollar by corresponding decreases in state foundation aid. Thus, districts with
higher property taxable value enjoy higher fund balance.
However, giving more money to school districts does not necessarily raise their fund
balance. The way the money is spent is important to help school districts maintain their fund
balance and fiscal status. This can be achieved through the efficiency of district management on
how they control class size, teachers’ salary, and business administrative expenses. As stated
earlier in the paper, less spending in these areas increases district fund balance.
Pupil teacher ratio, which indicates class size, is one of the district management strategies
to control their spending. Though not significant, this study found that bigger class size helps
district increase their fund balance. Some costs are saved with big class size, such as number of
74

teachers needed, teachers’ salary and number of classrooms. As evidenced in this study, higher
salary reduces district fund balance. Since this variable is related to class size, it can be
concluded that school districts that are able to reduce salary by having bigger class size can push
their fund balance to a higher level. However, if districts choose this strategy as one of the ways
to maintain their fund balance, they must consider the outcomes of having big classes, for
instance teaching and learning, teacher motivation, and student achievement. If districts do not
want to risk the outcomes, another area of spending that district could reduce is their business
administrative expenditure. This study found that districts that spend more on this expenditure
have lower fund balance. Districts would rather cut down their spending on business
expenditures than on teacher salary, especially if they are bound by a teachers’ contract and not
able to reduce their salary or increase class size.
In conclusion, district fund balance is associated with factors that are within and beyond
district control. Even though their fund balance would be better with efficient district
management on controlling their spending in the related areas, district characteristics which are
beyond district control also determine the level of their fund balance.

5.4

Policy Implications
Even though the findings of this study may not be generalized to other states, this

knowledge about the factors associated with district fund balance does offer valuable information
to policymakers’ effort in improving district financial condition. First, the part of enrollment is
undeniable in school funding. When students move from a district to another, one district loses
students, and the other gains. District’s fiscal health is better with more students because they
75

can have higher fund balance, but the task of educating students must continue even with a small
number of students, such as enrollment of two students in a district like Bois Blanc Pines School
District. Since these two components are equally related, the state should reevaluate the policy
concerning this issue. The school of choice policy would be one thing to consider if a district is
to fulfill and maintain the task of educating the children in a promising financial condition.
Under this policy, participating districts compete for students. This practice takes away districts’
resource of funding, though it helps raise revenues in favored districts. Continuing this policy
would make districts work harder to improve the operations of their educational institutions in
order to attract more students, and subsequently improve their fiscal health. Meanwhile, state
should assist the losing districts to restore their financial lost and continue providing education to
their children in a good fiscal health.
School districts receive additional revenue from local, state and federal through the
Categorical Grant for special education students. Even though these students are protected under
the federal Individuals with Disabilities Education Act (IDEA), the state reimburses only 28.6%
of the total costs allowed and incurred by districts. This low reimbursement rate from the state
does not pay for the full amount of the actual cost incurred. The problem is worse when the
number of these students is not uniformly distributed across districts. Both the total cost allowed
to be reimbursed, and the actual cost incurred should be taken into account for reimbursement by
the state. This consideration can help school districts reduce or eliminate financial deficits and be
out of fiscal stress.
Districts are classified as deficit if they have negative fund balance and required to
submit a Deficit Elimination Plan (DEP) to the Michigan Department of Education (MDE). If
they fail to eliminate their deficits according to the DEP, their financial affairs will be intervened
76

upon directly by the state under the Local Government and School District Fiscal Accountability
Act, which is not generally favored by local management. Before districts really are far in deficit,
a warning should be issued about the state of their finances. It is suggested that MDE produce a
standard measure that can be a reference for local districts to alert them about their financial
status. The standard could be a list of statuses, such as state watched list or state severe list based
on the amount of fund balance a district has.

5.5

Scope and Limitations of Study
The study covers all public school districts in the state of Michigan. Private and charter

schools are not included in this study. Since education funding systems differ widely across
states, the results may not be generalized to other states. Some of the data utilized in this study
are long-term data that could reflect the long-term situation of Michigan school funding and
some reflect only a snapshot of financial and demographic data at specific points of time. The
definition of “fiscal stress” in this study was created by the author based on the literature review
and the MDE requirements for school districts to have a positive fund balance at the end of the
fiscal year. All these factors may limit the application of the model developed in the study to
future time at different places with different policy.

5.6

Future Research
The research findings in this study are based on the descriptive statistics and statistical

analysis proposed by the author. It is undoubtedly that other statistical analyses will be helpful in
investigating the factors associated with school district fund balance. Continued research is
needed, especially with analysis that can examine how district level and time factor affect the

77

fund balance. It is hoped that future research will continue investigating and making
recommendations not only to local district management, but also to the state policymakers on
how school districts should use their resources efficiently to improve their financial status.

78

APPENDICES

79

Appendix A
Table A-1

Michigan Public Schools with Deficits for Fiscal Year Ending June 30, 2011
District

Deficit

1. Benton Harbor Area
Schools
2. Bellevue Community
Schools
3. Vanderbilt Area School
District
4. Les Cheneaux Community
Schools
5. Ewen-Trout Creek
Consolidated School Dist.
6. Hancock Public Schools

$15,090,065

7. Owendale-Gagetown Area
Schools
8. Hale Area Schools

35,703
134,801

9. Hudson Area Schools

302,622

415,651
82,983
89,275
775,033
1,557,539

10. Brighton Area Schools

7,246,901

11. East Detroit Public
Schools
12. Clintondale Community
Schools
13. Mt. Clemens Community
Schools
14. New Haven Community
Schools
15. Republic-Michigamme
Schools
16. Ishpeming Public School
District
17. Muskegon Height School
District
18. White Cloud School
District

8,218,403
5,538,622
2,467,182
643,591
84,592
535,433
4,319,698
190,418

80

District

Deficit

19. Pontiac School
$12,228,315
District
20. Avondale School
2,453,732
District
21. School District of the
3,927,017
City of Hazel Park
22. West Bloomfield School 1,722,193
District
23. Oak Park School
6,632,252
District
24. Covert Public
3,241,823
Schools
25. Ypsilanti Public Schools 3,755,042
26. Willow Run Community 2,803,557
Schools
27. Detroit Public Schools 327,299,271
28. Garden City School
District
29. Hamtramck Public
Schools
30. Highland Park City
Schools
31. School District of the
City of Inkster
32. Redford Union Schools
33. River Rouge School
District
34. Westwood Comm.
School District
35. Ecorse Public School
District
36. Southgate Community
Schools

456,066
4,077,367
7,211,206
9,300,981
1,573,788
3,055,907
5,472,777
27,838
1,731,524

Appendix B
District location definitions by the National Center for Education Statistics (NCES):
11 - City, Large: Territory inside an urbanized area and inside a principal city with population of
250,000 or more.
12 - City, Midsize: Territory inside an urbanized area and inside a principal city with population
less than 250,000 and greater than or equal to 100,000.
13 - City, Small: Territory inside an urbanized area and inside a principal city with population
less than 100,000.
21 - Suburb, Large: Territory outside a principal city and inside an urbanized area with
population of 250,000 or more.
22 - Suburb, Midsize: Territory outside a principal city and inside an urbanized area with
population less than 250,000 and greater than or equal to 100,000.
23 - Suburb, Small: Territory outside a principal city and inside an urbanized area with
population less than 100,000.
31 - Town, Fringe: Territory inside an urban cluster that is less than or equal to 10 miles from an
urbanized area.
32 - Town, Distant: Territory inside an urban cluster that is more than 10 miles and less than or
equal to 35 miles from an urbanized area.
33 - Town, Remote: Territory inside an urban cluster that is more than 35 miles from an
urbanized area.
41 - Rural, Fringe: Census-defined rural territory that is less than or equal to 5 miles from an
urbanized area, as well as rural territory that is less than or equal to 2.5 miles from an urban
cluster.

81

42 - Rural, Distant: Census-defined rural territory that is more than 5 miles but less than or equal
to 25 miles from an urbanized area, as well as rural territory that is more than 2.5 miles but
less than or equal to 10 miles from an urban cluster.
43 - Rural, Remote: Census-defined rural territory that is more than 25 miles from an urbanized
area and is also more than 10 miles from an urban cluster.

82

Appendix C
1.

Assumption Test of Multicollinearity
Table C-1

Variance inflation factor – enrollment
Coefficients
Collinearity Statistics
Tolerance
VIF
.927
1.079
.708
1.411
.585
1.709
.875
1.143
.422
2.370
.459
2.177
.309
3.235
.331
3.022
.775
1.290
.699
1.430

% enrollment change
% free and reduced lunch
% special education
% English Language Learners
Foundation allowance
Property taxable value
Pupil teacher ratio
Average salary
Business Admin. expenditures
Urban
Dependent Variable: enrollment

Table C-2 Variance inflation factor – percent enrollment change
Coefficients

% free and reduced lunch
% special education
% English Language Learners
Foundation allowance
Property taxable value
Pupil teacher ratio
Average salary
Business Admin. expenditures
Urban
Enrollment
Dependent Variable: % enrollment change

83

Collinearity Statistics
Tolerance
VIF
.698
1.432
.563
1.778
.813
1.229
.423
2.366
.460
2.172
.315
3.178
.328
3.049
.745
1.342
.528
1.892
.390
2.562

Table C-3

Variance inflation factor – percent free and reduced lunch
Coefficients

% special education
% English Language Learners
Foundation allowance
Property taxable value
Pupil teacher ratio
Average salary
Business Admin. expenditures
Urban
Enrollment
% enrollment change
Dependent Variable: % free and reduced lunch

Table C-4

Collinearity Statistics
Tolerance
VIF
.673
1.486
.831
1.203
.424
2.356
.466
2.145
.304
3.287
.326
3.068
.771
1.297
.530
1.886
.398
2.511
.932
1.073

Variance inflation factor – percent special education
Coefficients

% English Language Learners
Foundation allowance
Property taxable value
Pupil teacher ratio
Average salary
Business Admin. expenditures
Urban
Enrollment
% enrollment change
% free reduce lunch
Dependent Variable: % special education

84

Collinearity Statistics
Tolerance
VIF
.818
1.222
.424
2.359
.461
2.171
.314
3.180
.330
3.029
.765
1.307
.529
1.889
.406
2.463
.927
1.079
.831
1.204

Table C-5 Variance inflation factor – percent English Language Learners
Coefficients
Collinearity Statistics
Tolerance
VIF
Foundation allowance
.423
2.363
Property taxable value
.462
2.166
Pupil teacher ratio
.305
3.275
Average salary
.325
3.076
Business Admin. expenditures
.746
1.341
Urban
.531
1.885
Enrollment
.420
2.382
% enrollment change
.927
1.079
% free reduce lunch
.709
1.410
% special education
.565
1.768
Dependent Variable: % English Language Learners

Table C-6

Variance inflation factor – foundation allowance
Coefficients

Property taxable value
Pupil teacher ratio
Average salary
Business Admin. expenditures
Urban
Enrollment
% enrollment change
% free reduce lunch
% special education
% English Language Learners
Dependent Variable: Foundation allowance

85

Collinearity Statistics
Tolerance
VIF
.566
1.767
.341
2.929
.410
2.439
.759
1.318
.547
1.827
.390
2.562
.929
1.077
.699
1.431
.565
1.769
.816
1.225

Table C-7

Variance inflation factor – property taxable value
Coefficients

Pupil teacher ratio
Average salary
Business Admin. expenditures
Urban
Enrollment
% enrollment change
% free reduce lunch
% special education
% English Language Learners
Foundation allowance
Dependent Variable: Property taxable value

Table C-8

Collinearity Statistics
Tolerance
VIF
.339
2.953
.335
2.989
.750
1.334
.537
1.862
.390
2.563
.929
1.077
.705
1.419
.564
1.773
.818
1.223
.520
1.925

Variance inflation factor – pupil teacher ratio
Coefficients
Collinearity Statistics
Tolerance
VIF
.527
1.898
.748
1.337
.529
1.892
.397
2.520
.959
1.043
.695
1.440
.582
1.719
.817
1.224
.474
2.111
.512
1.955

Average salary
Business Admin. expenditures
Urban
Enrollment
% enrollment change
% free reduce lunch
% special education
% English Language Learners
Foundation allowance
Property taxable value
Dependent Variable: Pupil teacher ratio

86

Table C-9

Variance inflation factor – average salary
Coefficients
Collinearity Statistics
Tolerance
VIF
.786
1.272
.531
1.883
.397
2.518
.935
1.070
.696
1.436
.571
1.751
.814
1.229
.532
1.880
.473
2.115
.493
2.030

Business Admin. expenditures
Urban
Enrollment
% enrollment change
% free reduce lunch
% special education
% English Language Learners
Foundation allowance
Property taxable value
Pupil teacher ratio
Dependent Variable: Average salary

Table C-10 Variance inflation factor – business admin. expenditures
Coefficients
Collinearity Statistics
Tolerance
VIF
Urban
.530
1.888
Enrollment
.406
2.463
% enrollment change
.927
1.079
% free reduce lunch
.718
1.393
% special education
.578
1.731
% English Language Learners
.814
1.228
Foundation allowance
.429
2.329
Property taxable value
.462
2.164
Pupil teacher ratio
.305
3.277
Average salary
.343
2.916
Dependent Variable: Business admin. expenditures

87

Table C-11

Variance inflation factor – urban
Coefficients
Collinearity Statistics
Tolerance
VIF
.517
1.936
.927
1.079
.697
1.436
.563
1.775
.817
1.224
.437
2.289
.467
2.142
.304
3.288
.327
3.060
.747
1.339

Enrollment
% enrollment change
% free reduce lunch
% special education
% English Language Learners
Foundation allowance
Property taxable value
Pupil teacher ratio
Average salary
Business admin. expenditures
Dependent Variable: Urban

2.

Assumption Test of Normality
Figure C-1

Histogram

88

Figure C-2

Normal P-P Plot of Regression Standardized Residual

3. Assumption Test of Heteroskedasticity
Figure C-3

Scatterplot

89

REFERENCES

90

REFERENCES
Addonizio, M.F. (2003). From fiscal equity to educational adequacy: Lessons learned from
Michigan. Journal of Education Finance, Spring (28), 457-484.
Agresti, A., & Finlay, B. (2009). Statistical methods for the social sciences. Upper Saddle River:
Prentice Hall, Inc.
Ammar, S., Duncombe, W., Jump, B. (2005). A financial condition indicator system for school
districts: A case of New York. Journal of Education Finance. 30 (3), 231-258.
Arsen, D. & Davis, T. (2005) Taj Mahals or decaying shacks: Patterns in local school capital
stock and unmet capital need. Peabody Journal of Education. 81(4), 1-22.
Arsen, D. & Ni, Y. Responding to tough economic times: Education finance and policy in
Michigan. Working paper, Michigan State University. April 2010.
Arsen, D., Plank, D. (2003). Michigan school finance under Proposal A: State control, local
consequences. East Lansing: Education Policy Center at Michigan State University.
Boyd, D., Lankford, H., Wyckoff, J. (2002). School district expenditures and fiscal stress.
Education Finance Research Consortium. New York State Education Department.
Bradbury, K. L. (1982). Fiscal distress in large U.S. Cities. New England Economic Review.
November/December 1982 33-44.
Brown, K. (1993). The 10-point test of financial condition: Toward an easy-to-use assessment
tool for smaller cities. Government Finance Review. 9: 21-26.
Center for Educational Performance and Information. Data and reports. Retrieved February 23,
2012, from http://www.michigan.gov/cepi
Chapman, J. I. (1998). “The continuing redistribution of fiscal stress: The long run consequences
of Proposition 13”. Working Paper, Lincoln Institute of Land Policy.
Chapman, J. I. (2008). State and local fiscal sustainability. Public Administration Review. 115131.
Coe, C. K. (2008). Preventing local government fiscal crises: Emerging best practices. Public
Administration Review. 68(4), 759-767.
Cullen, J. B., Loeb, S. (2004). School finance reform in Michigan: Evaluating Proposal A.
Chapter 7 in J. Yinger (Ed.) Helping children left behind: State aid and the pursuit of
educational equity. Cambridge: MIT Press.

91

Citizens Research Council of Michigan (2010). State and local revenues for public education in
Michigan. Report 363. http://www.crcmich.org/PUBLICAT/2010s/2010/rpt363.pdf
Citizens Research Council of Michigan (2010). Public education governance in Michigan.
Report 359. http://www.crcmich.org/PUBLICAT/2010s/2010/rpt359.pdf
De Luca, B. M. (2006). Models for predicting school district fiscal stress: One size does not
it all. Journal of Education Finance. 31(4), 420-432.
Dielman, T. E. (1989). Pooled cross-sectional and time-series data analysis. New York: Marcel
Dekker, Inc.
General Accounting Office (GAO) (1990a). Distressed communities: Public services declined
in California as budget pressures mounted. (GAO-HRD-90-95). Washington, DC.
Izraeli, O., Murphy, K. J. (2007). The impact of Proposal A on school financing, equity, and
quality of public schools in the state of Michigan. Journal of Education Finance. 33(2), 111-129.
Kloha, P., Weissert, C, Kleine, R. (2005). Developing and testing a composite model to predict
local fiscal distress. Public Administration Review 65. (3): 313-323.
Manca, D., Noonan, D. W., Matranga, M. (1999). Recipe for financial disaster: Three districts
tell the story. Journal of Education Finance. 24(3), 379-388.
Michigan Department of Education. http://www.michigan.gov/mde
Michigan Department of Education. Buletin 1014. Retrieved February 7, 2012, from
http://www.michigan.gov/mde/0,4615,7-140-5235_6539-21514--,00.html
Michigan Department of Education. Michigan public schools with deficits. Retrieved April 12,
2011, from http://michigan.gov/documents/mde/Item_B_School_Districts_Deficit_Status
Michigan Department of Treasury. (2002). School finance reform in Michigan Proposal A:
Retrospective. Retrieved August 14, 2010, from http://www.michigan.gov/treasury
Michigan School Business Officials. School finance. Retrieved February 23, 2012, from
http://www.msbo.org/school-finance
National Center for Education Statistics. http://nces.ed.gov/
Olson, R. S., & LaFaive, M. D. (2007). A Michigan school money primer for policymakers,
school officials, media, and residents. Midland: Mackinac Center for Public Policy.
Ott, R. L., & Longnecker, M. (2001). Statistical methods and data analysis. Pacific Grove:
Thomson Learning, Inc.

92

Smith, C. A. (1985). Forecasting school district fiscal health. PhD Dissertation, Ohio State
University, Columbus.
StatSoft Electronic Statistics Textbook. Multiple regression. Retrieved July 19, 2012, from
http://www.statsoft.com/textbook/multiple-regression/
Trussel, J. M., Patrick, A. (2009). A predictive model of fiscal distress in local governments.
Journal of Public Budgeting, Accounting & Financial Management. 21(4), 578-616.
Tuckman, H. P., Chang, C. F. (1991). A methodology for measuring the financial vulnerability
of charitable nonprofit organizations. Nonprofit and Voluntary Sector Quarterly. 20(4), 445-460.
Valdez, E. A. Validating the regression model. Retrieved July 19, 2012, from
http://www.math.uconn.edu/~valdez/math3621s11/M3621Week5-S2011.pdf
Zimmer, R., Jones, J. T. (2005). Unintended consequence of centralized public school funding in
Michigan education. Southern Economic Journal. 71(3), 534-544.

93