INFORMATION TO USERS
This was produced from a copy of a document sent to us for microfilming. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the material submitted.
The following explanation of techniques is provided to help you understand markings or notations which may appear on this reproduction.
1.The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)”. If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure you of complete continuity.
2. When an image on the film is obliterated with a round black mark it is an indication that the film inspector noticed either blurred copy because of movement during exposure, or duplicate copy. Unless we meant to delete copyrighted materials that should not have been filmed, you will find a good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photo graphed the photographer has followed a definite method in “sectioning” the material. It is customary to begin filming at the upper left hand comer o f a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.
4. For any illustrations that cannot be reproduced satisfactorily by xerography, photographic prints can be purchased at additional cost and tipped into your xerographic copy. Requests can be made to our Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases we have filmed the best available copy.
University Microfilms International 300 N. ZEEB ROAD. ANN ARBOR, Ml 48106 18 BEDFORD ROW, LONDON WC1R 4EJ, ENGLAND 8015919
Pa y t o n , Jim m y Ja y
THE EFFECTS OF THE EQUAL YIELD FORMULA ON SCHOOL FINANCE IN OHIO
The Ohio Stale University Ph.D. 1980
University Microfilms International 300 N. Zeeb Road, Ann Arbor, MI 48106 18 Bedford Row, London WC1R 4EJ, England THE EFFECTS OF THE EQUAL YIELD FORMULA ON
SCHOOL FINANCE IN OHIO
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
Jim J. Payton, B.S., M.A.
*****
The Ohio State University
1980
Reading Committee: Approved By 'f Frederick D. Stocker j * g j/ f / (jS / Walter G. Hack i f Adser John J. Kennedy Educational/Foundations apd Research ACKNOWLEDGEMENTS
The completion of this study would not have been possible without the assistance of many people. At the risk of leaving some persons unmentioned I would like to thank:
- the members of my committee for th e ir guidance throughout the project;
- Bill Harrison, who helped me get started;
- Rodney Neff, Dan Brown and Phil Detamore of the Ohio
Department of Education who facilitated the data acquisition for this study;
- Dr. Bruce Gensemer who took the time to answer my questions and give me advice;
- Barb Smith who typed, typed and retyped; and
- the school district superintendents who returned the survey questionnaire.
Finally, I would also like to express my appreciation to my children, Keely and Drew, who did without a full-time father and to my wife, Glenda, who did without a full-time husband so that this study could be completed. VITA
December 19, 1946 . . . Born - Dayton, Ohio
EDUCATION
1969 ...... B.S. Otterbein College, Westerville, Ohio Major: Mathematics and Education Minor: Economics
1 9 7 1 ...... M.A. in Economics, The Ohio State University, Columbus, Ohio
WORK EXPERIENCE
1969-1971 ...... Research Assistant and Teaching Assistant, Department of Economics, The Ohio State Uni versity
1971-1973 ...... Tax Economist, Ohio Department of Taxation, Columbus, Ohio
1973-1975 Research Associate, Center for Human Resource Research, The Ohio State University, Columbus, Ohio
1975-1977 ...... Evaluation Coordinator, Ohio Commission on Aging, Columbus, Ohio
1976 ...... Consultant, Education Review Committee, Ohio General Assembly
1977-Present ...... Educational Assessment Consultant, Ohio Department of Education, Columbus, Ohio
Fields o f Study
Research Methodology School Administration Program Evaluation School Finance
i i i TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS...... ' ...... ii
VITA...... iii
LIST OF TABLES...... vi
LIST OF FIGURES...... viii
Chapter
I. INTRODUCTION AND BACKGROUND ...... 1
A Historical Development of School Finance ...... 1 Sources for Revenue for Education ...... 1 The Evolution of Funding Systems ...... 5 History of School Finance in Ohio ...... 10 The Present S tu d y ...... 21
II. LITERATURE RELATED TO THE PRESENT STUDY: A REVIEW . . . 36
Expenditure Determinants Studies ...... 37 Theoretical Models for Prediction and Analysis . . . 44 Evaluations of State Aid Formulas ...... 57 Ohio Evaluation Studies ...... 74
III. METHOD...... 85
Variation Analysis ...... 86 Fiscal Neutrality Analysis ...... 92 Incentive for Effort Analysis ...... 104
IV. RESULTS...... 108
Variation Measures ...... 108 Fiscal Neutrality Analysis Results ...... 116 Incentive for Effort Analysis ...... 127 Open-Ended Comments from Questionnaire ...... 137
V. SUMMARY AND CONCLUSIONS...... 140
Sum m ary ...... 140 Conclusions ...... 143 D iscussion ...... 146 Recommendations for Further Research ...... 152 iv TABLE OF CONTENTS
Page
APPENDIXES
A. Definition of Variables ...... 156
B. District Type Definitions ...... 160
C. Regression Tables and Survey Instrument ...... 162
BIBLIOGRAPHY ...... 167
v LIST OF TABLES
Table Page
1 Distribution of Elementary and Secondary School Funds by Source: United States, 1919-20 to 1975-76 ...... 4
2 Distribution of Ohio Elementary and Secondary School Funds by Source: 1970-71 to 1977-78 16
3 Percentage of Funds Wealth Neutral by Phase-In Level for Five District State Model ...... 79
4 Percentage of Funds Wealth Neutral for Selected Phase-In Levels: All 617 Districts ...... 79
5 Characteristics of Selected Variation Measures ...... 91
6 Combinations of Dependent and Independent Variables Used in Simple R e g re s s io n s . 95
7 Independent Variables Included in Multiple Regressions with Capacity Measures ...... 96
8 Variation Measures for Total Operating Expenditures Per Pupil for All Districts, 1974 to 1979 ...... 109
9 Percentage Change in Variation Measures Caused by Distribution of Basic State Aid Being Added to Other State Aid plus Local Revenue for all D istricts, 1974 to 1979 ...... 113
10 Measures of Variation Among School D istrict Types of Total Operating Expenditures Per Pupil, 1974 to 1979 115 2 11 R Values for Representative Simple Regressions, 1974 to 1979 117
12 Regression Coefficients of Property Wealth Measures of Multiple Regression with TOPEPP or TOPEWP as the Dependent Variable, 1974 to 1979 120
vi LIST OF TABLES
Table Page
13 Gini Coefficients for Lorenze Curves Using TOPEX on Vertical Axis and Weighted and Unweighted Students Ranked by Various Wealth Measures on Vertical Axis, 1974 to 1979 ...... 122
14 Capacity Gini Coefficients Calculated by Subtracting Tax Rate Gini from Total Expenditure Gini for Various Wealth Measures, 1974 to 1979 . . . , . 126
15 Percent and Number Answering Questions That Answered "Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1976 ...... 133
16 Percent and Number Answering Questions That Answered "Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1977 ...... 134
17 Percent and Number Answering Questions That Answered "Yes" for Questions 1--5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1978 ...... 135
18 Percent and Number Answering Questions That Answered "Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1979 ...... 136
19 Results of Simple Regressions of Total Operating Expenditures Per Pupil (TOPEPP) or Various Wealth Measures, 1974 to 1979 163
20 Results of Simple Regressions of Total Operating Expenditures Per Weighted Pupil (TOPEWP) on Various Wealth Measures, 1974 to 1979 ...... 164
vi i LIST OF FIGURES
Figure Page 1 Response options a district at point P faces when a district power equalizing formula is implemented and the opportunity line shifts from XX to Z Z ...... 31
2 Lorenze curve and 45° l i n e ...... 98
3 Lorenze curve showing disequalization due to variation in total current operating expenditures per p u p i l ...... 101
4 Lorenze curve showing disequality due to local expenditures, that is, combined effect of AV and TR ...... 102
5 Lorenze curve showing disequality due to variation in AVon l y ...... 103
6 Lorenze curves showing disequalization due to distribution of Total Operating Expenditures to students in d istric ts ranked by Adjusted Equalized PropertyValue Per Pupil (AEPVPP) ...... 124
vi i i CHAPTER I
INTRODUCTION AND BACKGROUND
A Historical Development of School Finance
The purpose of this study was to assess the influence that the implementation of the Equal Yield Formula has had on the financing of elementary and secondary public education in Ohio, using a selected set of criteria. Before discussing the present research in detail, a brief historical development of school finance in the United States and a description of the history of funding of elementary and secondary education in Ohio will be presented. Also, a discussion of the legal challenges to school finance systems in California, New Jersey and Ohio will be presented. This should add some insight into the rationale for how and why we have arrived at the present status of school finance to be evaluated in this study.
Sources for Revenue for Education
In school year 1977-78, the amount spent on public elementary and secondary education in Ohio was $2991 million. Of this amount, ap proximately 47.5 percent came from local sources, 46.8 percent came from the state, and 5.7 percent came from the federal government (see
Table 2).
The combination of levels of government involvement in edu cation had its roots as far back as 1683 when the State of Massachu setts decreed that towns with 500 or more families were required to support two grammar and two writing schools. The towns provided the revenue, but the towns were divided into districts which had control over expenditures (Grubb & Michelson, 1974, p. 25). This system spread to other New England states and eventually throughout the North. Con trol, however, was eventually centralized at the town level for efficiency and coordination of policy (Grubb & Michelson, 1974, p. 26).
In the early eighteenth century, state level control began to develop in the form of state inspections of schools. By the beginning of the nineteenth century, state boards of education were being established and these boards gained power to establish standards with respect to buildings, curricula, days of school, textbooks, and teacher certification (Grubb & Michelson, 1974, p. 27).
Although this centralization of some control continued to develop in the northern states, state revenues accounted for a rela tively small portion of total expenditures for education prior to
1900 (see Table 2-1 in Grubb & Michelson, 1974).
This local autonomy was due to what Coons, Clune and Sugar- man (1970, p. 14) have referred to as "subsidiarity," This refers to the idea that decisions should be made and administration should take place at the level closest to the decision situation. Thus, the local community was thought to be the proper level for control of schools.
Fear of state control following state dollars sustained resistance to state funding.
In the South, following the Civil War, education was estab lished by state reconstruction governments which provided low levels of funding. With the implementation of literacy tests for voting and with industrialization, the need for improved education was apparent. Fear that states would support black schools and that the Fourteenth
Amendment would require equal distribution of funds, caused schools to turn to county and local taxes for increased funding (Grubb &
Michelson, 1974, p. 25).
Thus, while education in the North started on a local basis and moved to some slight state control before 1900, education in the
South started on a state basis and began to move toward more local control.
At the beginning of the twentieth century, several educators including Elwood Cubberley voiced concern over the fact that reliance on local resources had led to considerable inequality in spending on education among school districts. The desire for better schools and for a more equal distribution of spending led to increased funding from the state. Table 1 shows the relative shares of spending on education by the three levels of government. You will note that the state share increased from 16.5 percent in 1920 to 43.9 percent in
1976, and that the federal share had grown to 8.8 percent in 1976.
Local revenue was still the largest share on average. However, recent estimates by the U. S. Office of Education and National Education
Association indicate that 1978-79 school year will be the first year in which state aid is the largest share. The 1978-79 estimated shares are state (47.4 percent), local (43.7 percent), and federal (8.8 per cent) . 4
TABLE 1 Distribution of Elementary and Secondary School Funds By Source: United States, 1919-20 to 1975-76
Percentage of Total by Source Total Financial Support School Year (In millions of dollars) Local3 State Federal
1919-20 970.1 83.2 16.5 .3 1929-30 2,088.6 82.7 16.9 .4 1939-40 2,260.5 68.0 30.3 1.8 1941-42 2,416.6 67.1 31.4 1.4 1943-44 2,604.3 65.6 33.0 1.4 1945-46 3,059.8 63.9 34.7 1.4 1947-48 4,311.5 58.3 38.9 2.8 1949-50 5,437.0 57.3 39.8 2.9 1951-52 6,423.8 57.8 38.6 3.5 1953-54 7,866.9 58.1 37.4 4.5 1955-56 9,686.7 55.9 39.5 4.6 1957-58 12,181.5 56.6 39.4 4.0 1959-60 14,746.6 56.5 39.1 4.4 1961-62 17,527.7 56.9 38.7 4.3 1963-64 20,544.2 56.3 39.3 4.4 1965-66 25,356.9 53.0 39.1 7.9 1967-68 31,903.1 52.7 38.5 8.8 1969-70 40,266.9 52.1 39.9 8.0 1971-72 50,003.6 52.8 38.3 8.9 1973-74 58,230.9 50.1 41.4 8.5 1975-76 70,802.8 47.4 43.9 8.8
SOURCES: U. S. Department of Health, Education and Welfare, National Center for Education Statistics, Statistics of State School Systems; and Revenues and Expenditures for Public Elementary and Secondary, 1975-76 (preliminary data).
aIncludes a very small amount of gifts, tuition and trans portation fees from patrons.
This reliance on local sources could have resulted from the fear that increased state dollars would mean increased state control.
It should be noted, however, that the correlation between state aid and state control has not been substantiated in the research on this topic
(see Fowlkes & Watson, 1975: Grubb & Michelson, 1974; Levin & Cohen). 5
It is well known that large disparities did exist and still do exist in per pupil expenditures among states and among school dis tricts within states. The plans or formulas posed by school finance theorists were intended to, in part, decrease these expenditure dis parities among school districts. The plans have not been allowed to work as intended. As Elchanan Cohn (1974, p. 23) has stated:
The structure, funding, or encumbering provisions of the various legislative acts in the states often serve to di lute or distort effectiveness of the programs as originally intended.
Grubb and Michelson (1974, p. 29) expressed this idea as follows: ...The commitment to equalization of revenues has always been tempered by other principles which have weakened the equalization effects of state aid-- among them the stimu lation of additional local revenue for schooling, the maintenance of "lighthouse districts" of purportedly special excellence, the maintenance of local control, and the provision of across-the-board property tax relief.
It is important, however, to be familiar with the basic plans for funding education which were developed by these theorists and which have dominated most of the school finance thinking of the twentieth century since these will be referred to throughout this dissertation.
The brief description of these plans that follows will help to put the present status of school finance in Ohio in context.
The Evolution of Funding Systems
In recognition of the unequal distribution of spending on education among school d is tric ts , Elwood Cubberley stated in his doctoral dissertation in 1905 that state aid should be distributed "in definite relation to the needs of the community and to the efforts which it makes to provide good schools and to secure the attendance of children at them." (see Cubberley, 1905, p. 4). Cubberley proposed that state general aid should be dis tributed based on a criteria involving the number of teachers employed by the district (effort) and the average on aggregate daily attendance
(needs). This general aid was to be supplemented by equalizing grants to those districts that were not able to meet state established mini mum standards even though they had taxed themselves at the maximum rate allowable by law. This system is what is now referred to as the Flat Grant Plan.
Although Cubberley's plan was adopted by New York State, several researchers that followed Cubberley criticized his plan on the grounds that it did not really equalize the distribution of state funds. Districts with higher wealth could afford to hire more teachers and thus received higher state aid subsidies based on the effort cri teria than could the poorer districts. This is the same criticism leveled against the present adjustments in the Ohio state aid system for number of teachers, teacher experience and training, and number of' educational service personnel.
Cubberley's objectives of equalization of educational op portunity and reward for local effort were thought to be mutually inconsistent by George D. Strayer and Robert M. Haig. This idea was reflected years la ter by James S. Coleman when he stated in the forward to (Coons, Clune & Sugarman, 1970):
The history of education since the industrial revolution shows a continual struggle between two forces: the desire by members of society to have educational opportunity for all children, and the desire of each family to provide the best education it can afford for its own children. Neither of these desires is to be despised; they both lead to in vestment by the older generation in the younger. But they can lead to quite different concrete actions. 7
Strayer and Haig emphasized the equalization of educational opportunity in fiscal terms as the major objective of state aid. In
1923, they developed a plan which is now known as the Strayer-Haig
Minimum Foundation Plan which embodied the following elements:
1. A satisfactory minimum education program is developed.
2. The cost of such a program is determined.
3. The property tax rate needed to finance such a program in the district with the highest property valuation is calculated.
4. Each district in the state is required to tax at this
"charge off" rate needed in the wealthiest district to finance the minimum program.
5. Each district applies this charge off rate to its property valuation to determine how much of the cost of the minimum program can be paid for locally.
6. The state grants to each district the difference between what it can raise locally at the charge off rate and the cost of the minimum program.
7. There is no maximum tax rate. D istricts may fund a more expensive program than the minimum.
The crucial element here is the level of the minimum program since this plan really only equalizes up to the minimum level. Wealthy districts can increase expenditures with less millage than it would take for a poorer district to increase its expenditures to a comparable level. If the minimum level is too low, poorer districts may not be able to provide an adequate educational program without an excessive tax rate. Paul R. Mort was an advocate of the Strayer-Haig Minimum
Foundation Plan and did much to spread its use throughout the states.
In fact, it was this plan that was the basis for state aid to edu cation in Ohio from 1935 to 1975. Mort refined measures of d is tric t financial need, defined minimum programs, and developed a weighted pupil technique to be included in the Foundation Plan (see Cowle,
1978, p. 15).
Mort supported the idea that local leeway was essential so that light house districts could develop innovative educational pro grams. Idealistically, his hope was that the successful new programs would be disseminated to other districts and eventually be funded statewide as part of the minimum foundation program. (Jones, n.d.).
Harlan Updegraff, a school finance theorist and a contem porary of Strayer and Haig, followed Cubberley in emphasizing the concept of local effort. As R. L. Johns (1971, pp. 6-7) summarized
Updegraff's major ideas:
The efficient participation of citizens in the responsibility of citizenship should be promoted by making the extent of the state's contribution dependent upon local action... Efficiency in the conduct of schools should be promoted by increasing the state grant whenever the true tax rate is increased and by lowering it whenever the local tax is decreased.
Updegraff's basic plan has evolved into what is at present called the Percentage Equalizing Plan. Under this plan, equalization aid is granted on a matching basis as a percentage of local expendi tures per pupil. Relative property wealth is taken into account so that districts with relatively low property wealth will receive a higher matching percentage. Several states have adopted some form of this plan in combination with various categorical programs. In 1930, Henry C. Morrison, a professor at the University of
Chicago, was one of the f ir s t school finance theorists to advocate
full state funding. He argued that there should be a statewide uni fied school system with full state funding using state income tax.
His primary objective in advocating such a plan was equalization of educational opportunity (1930).
Since 1930, this plan has received increased support, but at the present only Hawaii has a statewide system of education.
The above-discussed plans have dominated school finance thinking up to recent years. It was not until 1970 that the Power
Equalizing Plan became prominant. This plan, proposed by J. E.
Coons, W. H. Clune III, and S. D. Sugarman (1970), has as its basic tenet the idea that the amount of state aid a school d is tric t re ceives should be solely a function of the tax rate voters are willing to approve for education. Under this plan, state aid would be determined as follows:
1. The property valuation per pupil needed to provide an adequate basic level educational program if a stipulated tax rate is employed is determined. This valuation per pupil figure will be the
"guaranteed valuation per pupil."
2. The difference between the guaranteed valuation per pupil and the district's actual valuation per pupil is calculated.
This figure will be positive if the guaranteed valuation per pupil exceeds the actual, and negative if the actual valuation per pupil exceeds the guaranteed valuation per pupil. 10
3. This difference in valuation per pupil is multiplied by the district's tax rate to determine the amount of state aid to be granted to the district. However, in the case where the difference found in 2 above is negative, the amount calculated in this step will be returned to the state by the district.
Thus, in this plan millage has "equal revenue raising power" regardless of which school district it applies to.
Various forms of this plan have been recently adopted in several states. However, political realities and budget constraints have caused states to limit millage for which a certain level of valu ation is guaranteed, and have caused them to eliminate the requirement that districts with per pupil valuations higher than the guaranteed . level return funds in excess of the guaranteed level to the state.
A modified form of the district power equalizing plan was adopted in Ohio in 1975, and i t is the effects of moving from the
Strayer-Haig type plan to this power equalizing type plan which is the subject of this study. However, a brief history of school finance in
Ohio will first be presented to explore the major events that lead up to the adoption of this plan in Ohio.
History of School Finance in Ohio
The origin of school finance in Ohio goes back to the Land
Ordinances of 1785 and 1787 which provided that the income from the sixteenth section of each township in the Northwest Territory was to be used for education (this section draws heavily.from Ohio Dept, of
Education, 1976). 11
In 1803, Section 2 of Article VI of the Ohio Constitution required the legislature to provide for a thorough and efficient statewide system of public schools. In turn, the General Assembly delegated much of the responsibility for operating schools to local school d is tric ts .
Up to 1930, public schools in Ohio were funded solely with real property taxes and "school lands" revenue. In 1935, the School
Foundation Program was initiated and funded with a three-percent retail sales tax. At this point, the state assumed approximately half of the cost of public schools. This program was based on the Strayer-Haig
Minimum Foundation Plan.
Since the 1930s, various revisions have been made in the foundation program in order to increase the level of funding and the equality of the distribution of funds among school districts. An example of such a set of changes was made in 1971 with the enactment of Amended Substitute House Bill 475.
Prior to implementation of Am.Sub. H.B. 475, the foundation support a school district would receive was calculated as follows.
First, the number of approved classroom units was multiplied by the estimated cost of salaries and retirement payments for teachers. To this product another per-unit allowance for additional classroom ex penses and an allowance for pupil transportation costs were added.
From this sum, the 17% mill charge off rate times the district's tax duplicate was subtracted. The district would receive this remainder or $3,400 per classroom unit, whichever was larger. To this figure, an additional $1,000 per unit for special education was added. 12
With the enactment of Am.Sub. H.B. 475, the foundation sup port was calculated as follows. The basic program support was based on a per-pupil allowance of $600 for each pupil in grades one through twelve instead of the classroom unit calculation. The $3,400 per unit guarantee was replaced by an adjustable per pupil guarantee with the amount of the guarantee being based on the top valuation per pupil of the district. The charge off rate was raised from 17% to 20 mills as was the minimum operating millage rate required for a district to participate in the program. After the charge off calculation, an a l lowance for nonpublic students and a municipal overburden allowance for special expenses incurred by districts with high percentages of low income students were added. These changes and the significant increase in state aid to education were accompanied by the enactment of a state income tax to help fund education and other state services.
The revisions in the formula designed to make the distri bution of funds to school d is tric ts more equitable were muted by the re a litie s of politics which resisted the notion that any school d is trict would have to give up some of its state aid. Thus, several guarantees were enacted to assure that districts would get a guaranteed amount of aid if their foundation calculation was lower than the guarantee.
In 1973-74, only 45.1 percent of all state aid was being paid to school d is tric ts based on the foundation calculation with the rest being allocated according to guarantees. Only two districts out of 617 were not receiving some kind of guarantee. In 1974-75, two supplemental f la t grants of $20 and $40 paid to each school d is tric t 13 further reduced the role of the foundation formula so that only 35.1 percent of all state aid could be attributed to the operation of the formula (Gensemer, 1977, p. 2). Also, the large increases in the level of state aid to schools was somewhat neutralized by inflation.
Due to the ineffectiveness of the foundation program to distribute funds in a manner that would keep the level of expendi tures from being so closely tied to local property wealth, and due to erosion of funding increases by inflation, the 111th General Assembly made sweeping revisions in the Ohio school finance laws which became known as the "Equal Yield Formula." These revisions, which became effective for the 1975-76 school year, changed the basis of the foundation program from a Strayer-Haig type approach to a power equalizing type approach. However, several modifications were made in the pure power equalizing format.
The amount of state aid a school d is tric t receives is based on the following principles.
1. Qualifying Millage
A district must have an operating tax rate of 20 mills to participate.
2. Local Yield Per Pupil Per Mill
This is calculated to determine how much revenue for each pupil can be generated with one mill.
3. The Guaranteed Yield
The state guarantees that each district will have available to it $48 per pupil per mill for the first 20 mills and $42 per pupil per mill for the next 10 mills between 20 and 30 m ills. There is no guaranteed yield for millage above 30. Another way to say this is- that the state guarantees $960 per pupil for the f ir s t 20 mills and $1,380 per pupil for districts levying at least 30 mills. 14
As an example, if a d is tric t's local yield per pupil is $38, the state will provide the additional $10 for the first 20 mills to make a total guaranteed yield of $48 per pupil per mill. The state will also provide the additional $4 per pupil per mill for each mill levied between 20 and 30 mills. Thus, regardless of the property wealth of a district, it is guaranteed a certain yeild for each mill levied. If this basic aid calculation is less than the basic state aid amount received for fiscal year 1977, then the fiscal year 1977 amount will be guaranteed to the district.
4. Adjustments
Adjustments are added to or substracted from the basic support figure based on pupil-teacher ratios, training and experience of classroom teachers, and pupil to edu cational service personnel ratios. Districts are guaranteed the fiscal year 1977 unadjusted basic state aid figure plus the present year's adjustments.
5. Categorical Funding
To the basic support figure are added allowances for approved vocational education and approved special education units.
Also added to the basic support are salary allowances for employee services beyond the regular school term, allowances for pupil transportation, and special al lowances to d istric ts with high concentrations of dis advantaged students, which replaced the municipal overburden allowance.
6. Guarantees
The reappraisal guarantee provides that districts will . receive no less than they received in the year prior to being reappraised. This guarantee applies for one year.
For any district that is consolidated with another, a period of three years is allowed in which the con solidated district is guaranteed at least the sum the combined districts received in the year of consoli dation.
Not enough money was appropriated to fully fund this plan in the 1975-76 school year, so the basic support amount was phased in over four years. The difference between the basic support calculated under 15 the Equal Yield Formula and the fiscal year 1975 basic support was multiplied by the phase-in percentage. This product was then added to the 1975 basic support amount to determine the present year's basic support entitlement. The phase-in percentages of the respective years were 17 percent (1975-76), 26 percent (1976-77), 46 percent
(1977-78), and 100 percent (1978-79).
In addition to looking at the changes in the form of the foundation program, it is important to also note the level of funding and the distribution of funds for education by source. As noted earlier, in 1935 the State Foundation Program provided about half of the funds for elementary and secondary education in Ohio. This per centage remained fairly stable until 1946 when i t began to decrease and local sources provided an increasing proportion of funding. By
1963-64, the percentage of operating funds provided by the state was only 31 percent. As shown in Table 2, by 1971-72, the s ta te 's share of funding had gone back up to 41.8 percent with the enactment of the state income tax. The state's share rose to a high of 47.9 percent in
1974-75. Along with these changes, the federal share increased from.
2.2 percent in 1963-64 (not shown in Table 2) to 5.7 percent in
1977-78.
It should be noted that the state's share of funding consists of foundation formula aid and categorical aid. Since 1970-71, the per cent of state funding that was categorical grew from 9.4 percent to
31.7 percent in 1977-78. Thus, a good portion of the increase in state funds was in categorical programs. 16
TABLE 2 Distribution of Ohio Elementary and Secondary School Funds By Source, 1970-71 to 1977-78
Percentage of Total by Source Total Financial Support School Year (In millions of dollars) Local State Federal
1977-78 2991.0 47.5 46.8 5.7 1976-77 2816.7 47.8 46.1 6.1 1975-76 2623.0 47.3 46.7 6.0 1974-75 2459.7 46.8 47.9 5.3 1973-74 2199.6 48.7 46.0 5.3 1972-73 2094.2 51.1 44.0 4.9 1971-72 1943.1 53.2 41.8 5.0 1970-71 1573.8 58.8 35.5 5.7
SOURCE: Ohio Department of Education, Division of School Finance.
In spite of the long line of revisions and increases in funding, the Ohio Foundation Program has come under legal attack for not providing equal educational opportunity for all children and for not providing a thorough and efficien t system of public schools. This has come in the wake of a groundswell of school finance litig atio n fol lowing the landmark Serrano v. Priest decision by the California
Supreme Court in August, 1971 (Serrano v. P rie st, 1971). In this decision, the Court held that the s ta te 's school finance laws were un constitutional on the basis of the plaintiff's complaints which had been dismissed by a trial court. The decision was made based on the premise that due to wide disparities in per pupil expenditures which were a function of local district property wealth, students in poor districts had been denied an equal opportunity for an education. The
Court ruled that this violated the equal protection clauses of both the federal and state constitutions. 17
In Robinson v. Cahill (Robinson v. C ahill, 1973), the New
Jersey Supreme Court held the New Jersey school finance system to be unconstitutional based on state constitution provisions only. Unlike
Serrano and other school finance cases, the New Jersey decision was not based on the equal protection argument, but on a state constitutional provision pertaining to education which requires the legislature to provide for the maintenance and support of a thorough and efficient system of public schools. The Court determined that the New Jersey school finance system failed to provide equal educational opportunity and thus, failed to provide a thorough and efficient system of public schools for all children. (For details of Robinson v. Cahill,
Serrano v. Priest and other cases, see Augenblick, 1978; Browing &
Long, 1974; Ohio General Assembly, 1972, pp. 1-17; Schoettle, 1972).
These two cases are cited as examples because the c rite ria upon which each case was decided are encompassed in the case filed in
Ohio. In April, 1976, the complaint was filed in Hamilton County
Common Pleas Court. The plaintiffs, which include the members of the
Cincinnati City Schools Board of Education, School Superintendent and Clerk-Treasurer, and a few parents and students in the Cincinnati
City Schools, asked the Court to declare the School Foundation Program unconstitutional in violation of Article VI, Section 2; Article VI,
Section 3; Article I, Section 2; and Article II, Section 26 of the
Ohio Constitution.
Specifically, the plaintiffs alleged that the School Found ation Program was in violation of the Ohio Constitution on the three counts listed below. (This section draws heavily from Complaint and
Action for Declaritory Judgment, 1976). 18
1. The Foundation Program places upon each school d is tric t the burden of raising most of the current operating funds needed to provide an adequate education for each school-aged child. The uneven distribution of property wealth among school districts enables some school d is tric ts to raise enough local revenue with low or moderate tax effort while other low property wealth districts must make exces sive tax effort to provide for just a minimum expenditure level. This system thus makes the level of educational opportunity a student will have a function of the whim of the voters in his or her school d is tric t.
The Foundation Program fa ils to provide funding commensurate with differing education needs of students, fails to take into account varying d is tric t needs based on "cost of doing business" d ifferen tials, and fails to account for d ifficu lties of d is tric ts in passing school levies which have other local tax burdens imposed on their voters
(municipal overburden). This system is, therefore, not able to pro vide a thorough and efficient system-of common schools, and is in violation of Article VI, Section 2 of the Ohio Constitution which man dates the General Assembly to secure a thorough and efficien t system of common schools through the state. The General Assembly also has failed to fulfill its responsibility to provide for the organization, administration, and control of a public school system supported by public funds as mandated by Article VI, Section 3 of the Ohio
Constitution.
2. Since the Foundation Program operates such that greater educational opportunity is afforded to children in some school dis tricts than to children in other school districts, these statutes 19 deprive the residents of Ohio of equal protection and benefit of the laws, in violation of Article I, Section 2 of the Ohio Constitution.
3. Since the Foundation Program operates such that equal educational opportunity is not provided to all school children, it operates so unevenly that it violates Article II, Section 26 of the
Ohio Constitution which states that all laws of a general nature shall have uniform operation throughout the state.
After over a year of litigation, the trial court ruled in favor of the plaintiffs on all counts, but the decision was appealed to the First Appellate D istrict of Ohio Court of Appeals. In September
1978, the Appeals Court upheld the decision of the tria l court on all counts except the violation of the "thorough and efficient clause," that is, Article VI, Section 2 of the Ohio Constitution. This decision was subsequently appealed to the Ohio Supreme Court where i t was heard in April 1979. On June 13, 1978, the Ohio Supreme Court, in a 6-1 decision ruled that the Equal Yield Formula was not unconstitutional.
(The rest of this section draws heavily from Bd. of Edu. v. Walter
(1979), 58 Ohio St. 2s.)
In affirming the Court of Appeals judgment that the Ohio statutory system of financing elementary and secondary education does not violate Section 2, Article VI (thorough and efficient clause), the.
Ohio Supreme Court concluded that the Equal Yield Formula's guaranteed floor of $960 per pupil at 20 mills was sufficent to assure that each d is tric t has the means to comply with state minimum standards. The court stated that a thorough system could not mean one in which part or any number of school d istric ts of the state were starved for funds and that an efficient system could not mean one in which part or any number 20
of school d istricts in the state lacked teachers, buildings, or equip
ment. Furthermore, the court found that the evidence did not support
the plaintiffs' contention that "educational deprivation" had occurred.
In deciding on this count, the court recognized the deference accorded
to the General Assembly in education matters, recognizing that the
court had no responsibility to second guess the wisdom or the policy
of the General Assembly.
In the dissenting opinion, Justice Locher cited the findings
of fact and legal conclusions of the trial court and the Court of Ap
peals as evidence of wide disparities in expenditures among school
d is tric ts which resulted in many school d istric ts not meeting minimum
state standards. This evidence showed, he concluded, that a thorough
and efficient system had not been provided by the state.
In order to decide whether the formula violated the Equal
Protection Clause of Article I, Section 2 of the Ohio Constitution,
the Court had to determine if there is a rational basis for unequal
educational opportunity. If a rational basis exists, the Equal Pro
tection Clause would be assumed to have not been violated unless the
discrimination impaired the exercise of a fundamental right or estab
lished a suspect classification. If this were the case, the state would have to show that this discrimination was due to a compelling
state interest.
The court ruled that the-Ohio Constitution did notexplicitly
or implicitly guarantee a right to education and, therefore, the right
to an education was not a fundamental right. Thus, the court only
needed to determine if there is a rational basis for the system. The court found that local control of schools is a rational basis for the disparities in expenditures per pupil among Ohio school districts and, therefore, reversed the judgment of the Court of Appeals that the Ohio statutory system for financing elementary and secondary education violated Section 2, Article I of the Ohio Constitution.
Justice Locher, in the minority dissenting opinion, stated that the right to education is a fundamental right because i t is explicitly guaranteed by the Ohio Constitution, because of its nexus to the rights to vote and to free speech, and because of its relationship to ones success in a democracy. He could find no com pelling state interest for discrimination with respect to this fundamental right. Justice Locher further argued that Section 2,
Article VI placed the primary responsibility for education at the state level and not at the local level and, therefore, local control was not a rational basis for the present system.
Regardless of the decision as to the constitutionality of the Equal Yield Formula, the questions regarding the effects of the formula on school finance in Ohio should be studied to provide infor mation for future decision-making of various in terest groups and of the Ohio General Assembly in formulating school finance legislation each biennium. It is with this history and this perspective toward future revisions in the system that we turn now to a description of the present study.
The Present Study
As will be noted in Chapter 2 in the review of the Ohio
Legislative Service Commission study and the study by Bruce Gensemer, 22
in spite of large amounts of increased state funds for education during
the f ir s t half of the 1970s, Ohio's school finance system was in need
of revision. While Ohio ranked 19th among the states in 1975 in
personal income per capita, i t ranked only 41st in terms of tax
effort for public elementary and secondary education. Ohio teachers'
average salary was about $1,000 below the national average (Gensemer,
1977, p. 7). Operating expenditures per pupil among school d is tric ts
ranged from $673 to $3,173. Wide disparities in tax valuation per
pupil ranging from $3,464 to $227,802 in 1975 coupled with an average
annual inflation rate for school costs of eight percent contributed to
the difficulty of passing school levies (Ohio Department of Education,
1975, and Gensemer, 1977, p. 4). Taxpayers in many d istric ts were
forced to tax themselves much more than taxpayers in other districts
to obtain the same level of expenditures as other districts. In
1975, only 28 percent of the new operating levies were passed by the
voters. Save harmless guarantees had supplanted the operation of the
foundation formula so much that by the 1973-74 school year, only 45.1
percent of all state aid was allocated according to the formula and only 2 of the 617 school d is tric ts were not receiving some kind of guarantee (Gensemer, 1977, p. 2).
In 1975, the 111th Ohio General Assembly made substantial
revisions in the school finance system in Ohio. As discussed in the
previous section these revisions, which became effective in the
1975-76 school year, changed the foundation program from a Strayer-
Haig minimum foundation-type plan to a modified power equalizing or guaranteed yield approach. The implementation of the new law came 23
with a set of expectations regarding its effects on the distribution
of school operating revenues in Ohio. Some of these expectations were
based on economic theory while others were based on the experiences
of other states. The Equal Yield Formula was intended to provide
incentive for low tax-low property wealth districts to increase their
millage rates, to reduce the variation among districts in operating
expenditures per pupil, and to move the system toward fiscal
neutrality.
Problem Statement. The reseach problem of this study was to determine,
after four years under the Equal Yield Formula, if these expectations
had been met. Specifically, the goals of the study were to determine
after four years under the Equal Yield Formula if variation in total
operating expenditures per pupil among districts had been reduced, if
the Ohio school finance system had moved toward fiscal neutrality,
and if the "incentive" millage had had any effect on levy proposals.
These expectations were formulated into hypotheses which were tested using a variety of analysis techniques employing actual data for school years 1973-74 through 1977-78, preliminary estimates for school year 1978-79, and some data from the 1970 U. S. Census.
The key variables and transformations are described in Appendix A.
It should be noted that the terms "equity" and equal educational opportunity are used frequently in this study. From a theoretical point of view, i t would perhaps be best to measure equity in terms of educational outputs or at least educational resource inputs. However, due to the technical difficulties in accurately 24 measuring such inputs and outputs, and due to the lack of data especially for output measures, equity and equal educational oppor tunity were measured using school operating revenues per unit of educational need (for a discussion of equity measures see Berne,
1977, pp. 2-14; Cohn, 1974, pp. 25-31; Grubb & Michelson, 1974, pp. 3-11).
The following pages state the hypotheses derived from the expectations and discuss the rationale for these hypotheses. The methodology for testing these hypotheses, the data, and the data analysis techniques employed are described in Chapter III.
Hypothesis I: The variation in operating expenditures per pupil among school d is tric ts will be found to have been reduced under the
Equal Yield Formula.
Under a pure power equalizing formula, state funds would be allocated to districts based on the difference between the actual local yield per pupil per mill and the guaranteed yield per pupil per mill. If a district's actual local yield is greater than the guaran teed yield, the district would turn back this excess to. the state for redistribution. This system would do much to reduce the variation in expenditures per pupil. Although this recapture provision was not a part of the Equal Yield Formula, variation in expenditures per pupil were still expected to decrease since basic aid would be distributed in an inverse relationship to local district property wealth, and because low tax-low property wealth districts would be able to raise the same amount per pupil per mill as wealthier districts up to the 30 mill limit. As stated in the Education Review Committee's 1974 staff report:
Under a Guaranteed Yield Plan1, there would be strong incentives for poor-property-wealth districts to in crease their tax rates. This is because for each additional mill of local tax effort, they would receive additional state dollars.
More will be said about this "incentive effect" when
Hypothesis III is discussed.
The experiences of other states that made a similar tran sition from a Strayer-Haig formula to a guaranteed yield type formula can be cited as further basis for this hypothesis. For example, the
Yang study reviewed in Chapter 2 indicated that variation in expendi tures per pupil in both Kansas and Illin o is was reduced after implementation'of a guaranteed yield type formula (Yang, 1975).
Hypothesis II: The Ohio system of funding public education will be found to be more fiscally neutral after implementation of the Equal
Yield Formula than it was before implementation of the Equal Yield
Formula.
Since the judicial decisions of Serrano v. Priest,
Rodriquez v. San Antonio (1971) and similar cases in other states, the idea of fiscal or wealth neutrality has become a major objective of state school finance reform proposals. The lower courts declared that local expenditures on education may not be a function of the taxable wealth of the local community. 26
If we look at education as a consumer good in our market
basket, it is apparent that the concept of fiscal neutrality dis
tinguishes education from other consumer goods and services in the
economy. Most demand schedules for other goods and services are a
function of consumers' income or wealth. However, in this case, the
courts have singled out education as an extremely important service,
the availability of which should not be left up to the random distri
bution of property wealth among local school districts.
Thus, fiscal neutrality refers to the relationship between
expenditures on education by school districts and their fiscal ability
or wealth, as measured by property, income, or some combination of the
two. The funding system moves toward fiscal neutrality as the re
lationship between district expenditures and district wealth becomes weaker.
Stephen Barro (1974, p. 32) described two notions of fiscal
neutrality— ex ante and ex post. In the ex post notion, the actual
level of education support must not be correlated with wealth,
regardless of pricing patterns and educational preferences of school
district voters. This strict definition of fiscal neutrality is sup
ported by those who argue that educational opportunity should not be a
function of taxpayers' willingness to approve school levies (see
Complaint and Action for Delaratory Judgment, 1976). The concern here
is that equalizing fiscal capacity among school d istric ts does not guarantee that every school d is tric t will provide equal educational opportunity. 27
The ex ante notion requires that the abilit.y of a district to support schools should not depend on local wealth. Under this requirement, a unit of effort produces the same support in every district. Other factors such as voter preferences for education may cause actual expenditures to vary. This notion of fiscal neutrality reflects the strong preference for local control in this country.
Coons, Clune and Sugarman (1970) proposed the d is tric t power equalizing formula as a means to satisfy the requirements of this ex ante definition of fiscal neutrality. Research reviewed in
Chapter 2 indicated that the states of Michigan, Illin o is, Kansas, and Colorado all experienced some movement toward fiscal neutrality after implementing new equal yield-type formulas (see Hickrod et a l .,
1976; Vincent & Adams, 1978; Yang, 1975).
The expectation of movement toward ex ante fiscal neutrality can be inferred from the Ohio General Assembly's Education
Review Committee's objective:
To move as expeditiously as possible to a Guaranteed Yield Program which will give most school d istric ts equal ab ility to raise revenues (Education Review Committee Staff, 1974, p. 7).
In the present study, both ex post and ex ante notions of fiscal neutrality will be explored. The task is to see if the re lationship between district expenditures and district fiscal capacity has been weakened from the pre-Equal Yield Formula years to the f ir s t four years of the Equal Yield Formula. 28
Hypothesis III; School districts with millage between 20 and 30 mills, and with yields per pupil per mill less than $42 will be found to have been stimulated to spend more local dollars under the Equal Yield
Formula than they would have under the old foundation plan.
As stated in the discussion of Hypothesis I, one of the reasons variation in per pupil expenditures was expected to be reduced under the Equal Yield Formula was its "incentive for effort" effect.
Under the Equal Yield Formula, every d is tric t is guaranteed $48 per pupil in combined state and local revenue for each of its f ir s t 20 mills levied and $42 per pupil for each mill between 20 and 30 mills.
Therefore, a district with millage between 20 and 30 mills and with a local yield per pupil per mill of, say, $32 would be granted $16 per pupil for each of its f ir s t 20 mills and $10 per pupil for each mill between 20 and 30 m ills. Each additional mill up to 30 mills would bring the district an additional $10 per pupil from the state.
Under the previous foundation plan, each additional mill would yield only local' revenue and no state matching funds.
Under the equal yield plan, these state matching funds were expected to provide school d is tric ts (especially those with low property wealth) incentive to increase their millage rates.
The rationale for this expectation and, thus, for
Hypothesis III can be explained through use of microeconomic consumer u tility maximization theory. This type model has been used many times in attempting to analyze local governments' fiscal behavior (see
Barlow, 1967; Cotton & O'Brien, 1968; Henderson, 1969; Hirsch, 1970;
Thurow, 1966; Wilde, 1971; Wilensky, 1970). In this analysis, the 29
school d is tric t is considered to be the consumer of educational ser
vices and other goods and services at a given set of prices. Under
the Equal Yield Formula, funds are distributed on a matching grant
basis with each d is tric t having its own matching ratio depending on
its local yield per pupil per mill.
How much of a particular good a consumer purchases depends
in part on the consumer's income and the price of the good. If the good is not an inferior good, an increase in income will cause an increase- in the consumption of the good in question and other goods.
How much more of each good the consumer will purchase with the increase in income is measured by the consumer's income e lasticity of demand for each good. Thus, if the income-elasticity of demand for education were high, an increase in income would result in a larger percentage increase in the purchase of education.
Most goods have downward si oping demand schedules which indicate that if the price of a good increases, the quantity demanded decreases. Or if the price goes down, the quantity demanded increases.
Thus, i f the price of education is lowered (raised) the d is tric t will move down (up) the demand curve and purchase more (less) education than it did at the previous price.
In looking at the various types of grants that a district could receive, we see that a lump sum or flat grant to a district increases the income of the district and has only an income effect on its purchase of education. On the other hand, a matching grant has both an income effect and a price effect on the district's pur chase of education. The fact that the district receives additional 30 income causes the income effect to occur. By enabling districts to purchase more units of education with a given amount of local funds than would be possible without the matching grant, the relative price of education is lower for the district. This drop in the price of education relative to other goods and services is called the "price effect" (Wilensky, p. 106).
Therefore, if a state changed from a system which did not reward local millage with additional state aid to a modified matching grant system for a given amount of aid, we would expect higher local expenditures since the matching grants provide both an income effect and a price effect. This was precisely the expectation in Ohio when the State Foundation Program was changed to the Equal Yield Formula.
The crucial implicit assumption was that the price elasticities of demand for education for the low to medium wealth districts were fairly high or at least not low. In other words, it was assumed that districts with less than a $42 local property tax yield per pupil per mill would increase their millage (and thus purchase more education) between 20 and 30 mills because the matching funds would lower the price of education to these districts. High wealth districts (those with local property tax yield per pupil per mill greater than $42) received no matching funds for increased millage between 20 and 30 m ills.
The response options a school d is tric t would face are depicted graphically in Figure 1 (this section based on Barro, 1974, pp. 15-75). 31
FIGURE 1
+ J
Local Tax Rate
Figure 1. Response options a district at point P faces when a district power equalizing formula is implemented and the opportunity line shifts from XX to 1 1 .
In this figure, the'district's opportunity line is shown as
a trade off between expenditures per pupil and localtax rate. Higher
expenditures come at the burden of. higher tax rates.
Assume line XX is the original opportunity line of the dis trict. The slope of this line indicates the additional yield per pupil for incremental increases in the local tax rate. The steeper the slope, the higher the yield per pupil per mill the line represents. Line 1 1 represents the new opportunity line resulting from implementation of a new guaranteed yield type program. The new schedule 1 1 is higher and the slope is steeper due to state matching funds for each mill levied. Point P is the original position of the district. Points
A, C, B and D represent possible positions for the district after im plementation of the new system. There is no clear cut response that a district will take. Point A shows the case of no expenditure increase and a tax decrease. Point C shows the district taking part of its in crease in state aid in the form of an increase in expenditures and the rest is allocated to a tax rate reduction. Point B depicts the case in which the tax rate remains constant but all the increase in state aid is spent on education. Point D demonstrates the case in which education's relative price is lower so the d is tric t is subject to both an income effect and a price effect. This case shows increased expen ditures and an increased tax rate. Point D, therefore, represents the position expected in Hypothesis III for low to medium wealth districts with millage between 20 mills and 30 mills.
The question to be answered is did the d istric ts move to point D and if so, was this movement due to the incentive effects of the Equal Yield Formula.
Significance of Problem. Much has been said and written about the
Equal Yield Formula and what i t has and has not done. Some school superintendents have complained that the Equal Yield Formula is unfair and has not had the effects it was intended to have. Other superin tendents have said that it has helped their districts. Its consti tutionality is being chalTended and defended in the courts. A few studies (to be reviewed in Chapter 2) have been conducted to assess various aspects of the Equal Yield Formula and its effects during the 33 first years of its implementation at less than 100 percent funding.
School d istric ts continue to have severe financial problems.
The present study is an attempt to assess whether the Equal
Yield Formula has had certain intended effects on reducing per pupil expenditure variations among districts, increasing fiscal neutrality of the system, and providing incentives for more local school tax effort. This study looks at these effects for two years prior to the
Equal Yield Formula and for each of four years of its implementation, the last year being one in which the formula was fully funded.
The survey results will pull together the opinions of the school superintendents regarding the incentive effects of the Equal
Yield Formula on local effort.
In light of the groundswell of revision taking place in school finance systems in Ohio and throughout the country, there is a need for research information in the decision-making process of policymakers. The results of research such as the present study should provide information to other researchers on which to build, and to policymakers to use in the formulation of future school finance systems. It is hoped that these results will shed some light on the school finance problems we are facing today.
Limitations of the Study. This study looks at a set of measures of expenditure variation and fiscal neutrality for two years prior to implementation of the Equal Yield Formula and for four years after implementation of the Equal Yield Formula. It attempts to detect changes in these measures to determine i f the expectations of the 34
Equal Yield Formula were realized after four years. Certain inade quacies of individual measures were recognized and various alternative measures were used so as to not base conclusions solely on one measure.
However, a limitation of this type of uncontrolled research is the confounding of the effects of other factors with the effects of the treatment variable. In paralleling this study to experimental research, the implementation of the Equal Yield Formula could be con sidered the "treatment."
While the intent of the Equal Yield Formula is to affect the distributions of both basic aid per pupil and total operating expendi tures per pupil, other factors may also affect both of these distri butions. Therefore, if the measures of variation indicate a movement toward less variation in expenditures or basic aid, we cannot be sure that this movement was "caused" by the Equal Yield Formula. On the other hand, if this expected movement did not occur, we can say that the Equal Yield Formula did not cause what i t was expected to.
The analysis of the incentive effects was an attempt to determine if the Equal Yield Formula had caused increases in local effort by asking this question directly. This expected increase in local effort by low valuation-low millage districts was intended to reduce variation in expenditures per pupil. Thus, the inclusion of the survey is an attempt to determine if the Equal Yield Formula had any effects on local effort and expenditure variation.
A similar limitation is also applicable to the fiscal neutrality analysis. If the Gini coefficients and regression analyses indicate movement toward fiscal neutrality, we cannot say for sure that the Equal Yield Formula caused this movement. However, if no movement is indicated, we can conclude that the Equal Yield Formula did not have this expected effect.
It should also be noted that the results of this study are of limited generalizability to other states with power equalizing-type formulas due to the intricacies of the Ohio formula and Ohio tax law.
These factors unique to Ohio should be taken into account in making inferences from the results of this study. CHAPTER II
LITERATURE RELATED TO THE PRESENT STUDY:
A REVIEW
As stated earlier, the purpose of the present study is to investigate the effects of the Equal Yield Formula on school finance in Ohio using several criteria. This chapter presents a review of a representative sample of studies related to the present study. This review will serve as a context in which this study should be viewed.
An extensive number of studies have been made dealing with issues that are directly or indirectly related to the present study, and no attempt will be made to review all of these. First, the literature was divided into several categories based on how the studies relate to the present study and then a sample of those studies from each category was chosen as representative of each category. •
The first group of studies deals with assessing the varia tions in school expenditures and with determining what factors are related to such variations. These studies relate to the present study's analysis of expenditure variation and regression analyses.
With school finance reform taking center stage in many state legislatures, the need to predict the possible effects of various al ternative reforms provided the impetus for school finance theorists and economists to build upon the earlier determinants studies to develop sophisticated theoretical models for prediction and analysis
36 37 of school finance reform proposals. Some of the major studies of this type constitute the second category. Some of the expectations and the hypotheses of the present study are based upon the theory and results of these studies.
The next category of studies includes those whose overall purposes correspond to the purpose of the present study. These studies deal with evaluating state aid formulas using various criteria of equity.
The final group of studies are actually a sub-group of the evaluation studies, but have been separated out since they are studies of Ohio's school finance system. The present study could be looked at as an expansion or continuation of this series of evaluation studies in Ohio.
It is hoped that this review will provide the reader with relevant information and a frame of reference in which the present study may be viewed.
Expenditure Determinants Studies
A number of studies have been done to assess variations in school expenditures and to determine what factors are responsible for these variations. The major purpose of these studies was to determine which factors affect school spending. While this is not a purpose of the present research, a few of these studies will be reviewed because the regressions run in these studies are similar to the regressions I will use to measure fiscal neutrality. The relationship between ex penditures per pupil and various fiscal capacity measures is of 38
interest here. Also, the significance of some of the independent
variables serves as a partial rationale for inclusion of similar
measures in the regressions of the present study.
Renshaw. - Using state level cross-section data for all 48 states in
the early 1950s, Edward Renshaw conducted a study (1960) to determine
the effects of state aid on public educational expenditures. Using two
dependent variables - current expenditures per pupil and total school
revenue per pupil, he ran a series of separate regressions. The nine
independent variables entered two, three or four at a time included
percentage of total operating funds that were from the state, percent
age of nonwhite population, per capita personal income 1949, per
capita personal income 1954, estimate of state-mandated local revenue
per pupil, total state aid per pupil, population density, income per
school-age population, and percentage of population that is school
age. In all regressions involving any of the measures of income, a
positive and statistically significant result was found for these measures.
Shapiro. - In 1962, Sherman Shapiro reported the results of an edu cational expenditure determinants study using cross-sectional data from all 48 states for the years 1920, 1930, 1940, and 1950. Data were divided into two sub-groups consisting of southern states and non southern states, and separate regressions were run for each year and for each sub-group. The two dependent variables used were current expenditures per pupil in public schools and current expenditures per pupil in public and private schools. After some preliminary runs, 39
Shapiro reduced his original list of nine independent variables to per capita income, pupils as a percentage of school-age population, per centage of labor force not in agriculture, high school enrollment as a percentage of total enrollment, and a regional "dummy" variable that was used only in the runs for all 48 states. 2 The multiple R values ranged from .40 to .67 for the non southern sub-group runs, .50 to .80 for the southern sub-group runs and from .65 to .90 for the runs for all 48 states.
The capacity measure in this study was per capita income.
With one exception, for all years per capita income was statistically significant and positively related to the dependent variable expendi ture measure for the non-southern sub-group. However, for the southern sub-group, per capita income had mixed signs and was not statistically significant.
James, et al. - In 1963, the results of a study designed to explore differences in current educational expenditures per pupil was reported by James, Thomas, and Dyck as part of a series of school finance studies at Stanford University. Data for school year 1958-59 were collected with a questionnaire from 772 school districts in ten states and combined with 1960 census data. Regressions were run and reported for the sample as a whole and for nine of the ten states separately.
The independent variables included measures of equalized property value per pupil, median family income, percentage owner occupied housing, median school years completed by persons over twenty-five, unemployment rate, percent non-white population, percentage rural pop ulation, percentage elementary school children attending private 40
schools, percentage of total revenue from the state, and a fiscal
independence dummy variable. The dependent variable was current
education expenditures per pupil.
The capacity measure, equalized property value per pupil, was found positively related to the dependent variable for all nine states, but only statistically significant in five of the nine. The coefficient for income was found to be positive for eight states and negative for one. Four of the positive and the one negative coef ficients were not significant.
Also, contrary to the conclusion of Johns and Salmon, no support was given to the hypothesis that the variation in expendi tures among districts declines as the level of state aid increases.
Bowman. — Part of the analysis of this 1973 doctoral dissertation by
John H. Bowman involved regression analysis using data for West
Virginia school d is tric ts for school year 1969-70. One of the major questions to be explored in this study was the nature of the relation ship between local taxes per pupil for education and local wealth or capacity. Although the regression analysis techniques used by Bowman resemble earlier determinents studies discussed above, this study dif fers from those previously cited in that local property taxes per pupil and not total expenditures per pupil is the dependent variable, and several of the independent variables were not used in the previous studies. Seven basic independent variables were non-local weighted assessed value per family, 1970 population as a percent of 1960, assessed value of local property per family, non-"marginal11 inter governmental school aid per family, percent families with income of 41
$15,000 or over, percent of families with children under eighteen, and
"marginal" intergovernmental school aid per pupil. Several different regressions were run with some variable substitutions. Both fiscal capacity measures--!ocal property value per pupil and percent families with income of $15,000 or over were found to be positively and sig nificantly related to local public school taxes per pupil. The 2 multiple R for the basic regression was .83.
Yang and Chaudhari. — This study reported on in 1976 by Thomas
Wei-Chi Yang and Ramesh Chaudhari looks at the relationship between selected socioeconomic variables and local tax effort. Data were collected for seventeen selected socioeconomic variables including measures of income, educational attainment, occupational status, . race, assessed property valuation, growth of property value and enrollment, population density, owner occupancy of housing and the ratio of local revenue to total revenue of school d istric ts in
Illin o is. Data from the 1970 census and school year 1974 were employed. Separate analyses were run for the three types of dis tricts—430 elementary, 127 high school, and 381 unit districts.
D istricts were ranked in ascending order by the level of their local tax rate and then were divided evenly into four groups— low, low medium, high medium, and high tax rate groups. Multiple discriminant analysis was employed to provide a geometric model of the similarities and differences among the tax rate groups in a reduced measurement space. As in the Bowman study, local tax effort rather than total expenditure per pupil was the "dependent" variable. 42
However, unlike the Bowman study which looked at tax expenditures per pupil, this study used tax rates as the tax effo rt measure. The use of multiple discriminant analysis took the earlier regression models a step further by attempting to delineate the relationships of the independent variables with more specific levels of the dependent variable than did the regression analysis studies.
The results confirmed that differences in fiscal capacity and educational aspirations contributed to differences in local tax effort.
However, contrary to the Bowman results, the relationship was negative for property wealth in that the low-tax-effort group had higher asses sed valuation per pupil than did the other three groups. However, the high-tax-effort group was characterized by high average income among other variables. While the high-medium-tax effo rt group was very similar to the high group, the low-medium-tax effort group was unlike any of the groups. It was characterized by average assessed valuation sim ilar to the high group, lower education attainment and a higher concentration of low income families.
Cohn. — As part of an excellent 1974 book title d Economics of State
Aid to Education, Elchanan Cohn presented an empirical model designed to analyze the effects of state aid on educational expenditures (see
Chapter 5). Several equations were developed and analysis was con ducted using ordinary leasts squares considering each equation independent of the others, and then with two-stage-1east-squares considering the set of equations as an interdependent system. Cohn used 1967-68 school year data for each state (except Hawaii). One of 43
the equations involved current expenditures per pupil as the dependent
variable and school size, nonpublic enrollment, par value of bond
issues per pupil, state equalization score, percent of total enroll ment that was Negro, percent of population that was urban, state aid
per pupil, student-teacher ratio, and personal income per pupil as the
independent variables.
For the ordinary-1east-squares analysis, personal income per pupil was found to be positively and significantly related to expendi tures per pupil. However, for the two-stage-1east-squares analysis, the coefficient for personal income per pupil was -.006 and not significant. The author speculated that this inconsistency could have been due to bias in the estimates produced by the ordinary-1east- squares approach.
Two limitations of this model cited by the author are that i t is too aggregative and i t lacks an adequate measure of how state aid is distributed. To overcome these difficulties, Cohn used a similar type model to perform analysis on sixty-seven Pennsylvania counties for the 1970 school year. Thus, the state aid formula was specified since only one state was involved. Two of the equations specifying this model were demand and supply equations for local educational expenditures. Both contained percentage of families with annual income of $10,000 or more as one of the independent variables.’
Using the two-stage-1east-squares technique, the income measure was shown to be positively and significantly associated with local expen ditures for the supply equation and positively but not significantly associated with local expenditures in the demand equation. 44
Directly related to the present research is the author's
comment (p. 91) that:
A comparison of the interstate and intrastate (county) models suggests - though i t does not prove - that utilization of the percentage equalizing procedure is likely to result in greater local incentives to in crease educational spending than is the case for states using the foundation approach. This conclusion is based on our results indicating a substitution of approximately $.65 for each dollar of state aid when the average of the forty-nine states is considered, as opposed to a substitution of only $.12 per dollar of state aid in Pennsylvania (a state using the per centage equalizing formula).
The expenditure determinants studies reviewed above had as
their major purpose the identification of variables or sets of vari
ables that had effects on either total or local spending on education
and the identification of the nature of these effects. The review
focused in on the relationship between capacity measures (income or
property wealth) and expenditures and showed that the majority of the
studies found positive and significant relationships between these measures.
The studies were reviewed in this way because it is this
relationship between expenditures and capacity that will be used as a criterion for fiscal neutrality in the present study.
Theoretical Models for Prediction and Analysis
With the impetus of various court cases challenging state
financing systems, many states have attempted to revise their plans.
In order to predict the consequences of various alternative revisions with respect to local district response, several statistical models have been developed. The studies involving these statistical pre diction models were based on a much more theoretical plane than the expenditures determinants studies previously discussed. While the major thrust of these studies was to develop a theoretical model of local fiscal response to various state aid plans, several contained empirical applications. Their relevance to the present study lies in the fact that the theory behind these models (especially the price and income effects of consumer maximization theory) and the results of their applications are related to the hypotheses stated in the present study. The following is a brief review of some of the more prominent studies of this type.
Stern. — The explanation of his model and the results and conclusions of his Ph.D. dissertation were reported by David Stern in a 1973 journal article. In this study, Stern attempted to do what some of the previous determinants studies had attempted to do in developing prediction models of local district response to state aid programs, but he used u tility maximization theory as a basis. Stern developed a model based on the key assumption that local school authorities will show some consistency in evaluating different combinations of local tax rates and total expenditures between which the board must make trade-off decisions. Under this and other assumptions, Stern inputed a preference function to local school boards and estimated its para meters using cross-sectional data on 1968-69 school expenditures in
157 Massachusetts school d is tric ts . The preference function took the following general form:
max. f (g(L); t(L,V); z) where L = local "reimbursable expenditure" per pupil
g(L) = the school board's perception of total current resources available per pupil including grants
t = current mi 11 age rate = L/v
V = equalized assessed value per pupil
Z = a vector of demographic variables
Thus, a school board's decision to raise taxes to spend more on education was assumed to depend on the board's perception of total current resources available per pupil including grants, the current mi 11 age rate, and other d is tric t demographic variables such as income, education, etc.
A behavioral equation was derived and maximized to arrive at the fully specified functional form. Using a nonlinear regression 2 program, an R of .76 was obtained applying the model to the
Massachusetts data. At this time, Massachusetts was operating under a combination percentage equalizing-block grant formula.
Adjustments were made in the model and simulations were run under a power equalizing formula. D istricts were ranked into quartiles by socioeconomic classification and again by property wealth. By comparing actual expenditures and tax rate data with the regression results and the district power equalizing simulation results among the various quartile combinations, Stern concluded that "power equalizing would apparently reduce the disparities due to property value but not the differences associated with socioeconomic status"
(p. 95). This conclusion is similar to the objection raised against 47
the Equal Yield Formula in Ohio that even though property valuation per
pupil is equated within limits, disparities in voter willingness to
pass levies still contributes to disparities in educational expendi tures .
Feldstein. ■— A 1975 a rtic le in the American Economic Review by Martin
Feldstein presents the development of a system which he claims neutralizes the effect of local wealth differences without sacrificing the opportunity for local choice in determining educational expendi tures. Two major conclusions drawn by Feldstein are:
1. ...the state can achieve any desired degree of wealth neutrality by using matching grants that cause the net price of educational services to vary with wealth in the appropriate way
(p. 87), and
2. based on elasticities estimated in a regression model
...the widely advocated district power equalizing form of percentage equalization grant, which has a price-wealth elasticity of 1, would not be wealth neutralizing but is more likely to result instead in an inverse relation between local wealth and local educational spending (p. 88).
Feldstein proposes a general constant elasticity equation of the form
K
in E,= B0 ♦ B, in W, ♦ B2 in P . ♦ j § B in ♦ e , (1) 48 where = expenditure per pupil in distirct i
W. = property wealth per pupil in district i
P. = net price that district i pays per dollar of educational expenditure
X. = other demographic variables
The price term P.. reflects the extent and form of grants from the state. The author states (p. 95) that the above equation
...im plies that the state government can neutralize the effect of intercommunity wealth differences by using a matching grant formula in which the rate of matching varies' inversely with local wealth.
This means that if m. is the amount of additional state aid a district receives when it spends an additional dollar on education, the net price P. = 1-m.j and, for example, if were high because were high, the state could make m^ low which would make P.. higher and cause E.. to be lower.
Feldstein then defines his wealth neutrality measure in terms of the behavioral parameters in the equation above and combines some terms to end up with
°^1 Bw + Bp Ypw (2)
where B = wealth ela stic ity combined with demographic variable elasticities with respect to wealth
Bp = partial price elasticity
Ypw = constant elasticity of prices with respect to wealth 49
Thus, Felstein is saying that wealth neutrality depends on how expenditures change with corresponding changes in wealth and other demographic variables that change with wealth (Bw) and on how expendi tures change with corresponding changes in the price of education which changes with the changes in wealth (Bp Y ).
To reach perfect wealth neutrality in which there is no relationship between wealth and expenditures, © The author then concludes that this equation implies that district power equalizing can achieve wealth neutrality only in the special case in which elasticity of wealth equals negative price elasticity. This is because a basic district power equalizing formula implies that price is proportionate to wealth and thus Y = 1. Therefore, substituting in equation 3, Feldstein shows that r 1 ■ - v V Using cross-sectional data for 1965 and 1970 for a sample of 105 Massachusetts towns that contained 72 percent of the state popu lation, Feldstein estimated the elasticities. Regression analyses were performed using equation (1) adjusted for the effects of wealth and using various combinations of 1965 and 1970 data. The results indi cated that price elasticities varied from -.716 to -1.016 while wealth ela stic itie s varied between .168 to .228. The price e la stic itie s were greater in absolute value than were the wealth elasticities which lead 50 Feldstein to conclude that district power equalizing would have the peculiar effect of causing spending in districts of low wealth to be much higher than in districts with high wealth. Since Y = 1 under a pw district power equalizing plan, the estimates of B and B imply that p w o Furthermore, the elasticity estimates implied that wealth neutrality could be achieved with a price elasticity with respect to wealth of between .35 and .42 (as opposed to YpW = 1 using a straig h t forward DPE). It is interesting to note that Feldstein used practically the same data as Stern used in the study cited above, but drew exactly opposite conclusions. However, Feldstein's results have been questioned due to questions voiced concerning the empirical techniques employed. These questions include: how many d istric ts have a zero matching rate which affects price elasticity estimates; do relatively low R s indicate important factors left out of the model; why have the effects of a decrease in state block grants which would accompany the imple mentation of a district power equalizing formula not been taken into account by Feldstein (Firestine & Carvellas, 1975, pp. 35-36). Barro. — In 1972, Stephen Barro wrote one of a series of Rand studies as the results of a theoretical inquiry into the factors that determine the fiscal behavior of local school d is tric ts . In this study, Barro developed models for predicting the consequences of grant-in-aid programs on local d is tric t response. These models were based on the 51 economic theory of constrained maximizing behavior in which each school district is viewed as a consumer faced with certain trade-off possi bilities between the level of support for schools and educational tax burden. One of the basic assumptions of the model was that districts would behave rationally in selecting the combination of expenditures and taxes that would maximize their utilities. The general utility function was expressed as U = U [e, b (t,y ), z] (4) where e = real educational expenditures b = tax burden t = tax collected per household y = household income z = taste variables The model further assumed that d istric ts attempt to maximize utility subject to the budget constraint t = « where f = lump sum federal grants g = state lump sum grants °<= local share of expenditures under a matching grant p = relative price of education a = enrollment ratio (students/population) As you will note, ®< corresponds to Feldstein's concept of "price." That is, price is the net price a district must pay out of 52 local funds for a dollar increase in educational expenditures. Barro calculated the marginal rate of trade-off or the rate at which the district is willing to raise taxes per household to increase edu cational expenditures to be ocpa. Barro then derived the general form of the expenditure equation e = e [y, © with plus and minus signs showing the expected direction of the effect of each independent variable on per pupil expenditures. Barro stated some of the major implications of his model as follows: 1. The effect of lump sum aid is always positive and less than one. It is therefore partly stimulative and partly substitutive. 2. The effect of a dollar increase in aid will be greater than a dollar increase in income (contrary to most consumer demand models). 3. The e lasticity of expenditures with respect to changes in p, a, or o< is the same. 4. Matching aid may stimulate higher local taxes than would have been levied in their absence. The model is extended to examine the following effects: (1) variation in non-school taxes, (2) variations in the composition of the local property tax base, (3) different equalization models, (4) categorical versus non-categorical, and (5) enrollment growth. 53 It was Barro's intent to develop a model that could be used by policymakers in making changes in school aid formulas. He pointed out that the equivalence of the effects of p, a, and means that changes in the composite price effect ( ©*pa) can be used to estimate the potential effect of matching grants in states which previously did not have matching grants. Firestine and Carvellas (1975, p. 52) point out that when Barro attempted to apply this model to actual data, he had to make several adjustments in the model which lessened the rigor of the original system. These changes caused Barro's empirical work to be similar to the earlier determinants studies. However, Barro's theoretical framework made explicit the expected effects of the given - variables while the e arlier work was done on a more ad hoc basis. It is Barro's model that served as a benchmark to aid in the selection of variables and to cross-check the validity of estimates made in the study by Firestine and Carvellas that follows. Firestine and Carvellas. — The results of a very ambitious study on estimating local district response to power equalized forms of state aid were reported in 1975 by Robert E. Firestine and John N. Carvellas. In this study, the authors develop a simple technique for estimating the impacts of new district power equalizing formulas in response to what they feel is a gap between rigorous theoretical models and their practical applications. The authors make use of some theory and make use of the estimates developed in previous theoretical models to select independent variables for their determinants equation. 54 This study thus combines what the authors feel are the most useful aspects of the straightforward determinants studies and the rigorous models based on utility maximization theory. Firestine and Carvellas present a simple regression "determinants" model utilizing ordinary-least-squares regression to determine school d is tric t fiscal and socioeconomic variables related to total current expenditures. The estimating equations took the following form: R = f (Fa , P, Gb, Z) where R = total district current education revenue per pupil F = financial ability (either income, property value or a combination of the two) P = price to the school d is tric t of increasing revenues one- dollar per pupil Gb = block grants for current purposes received by district Z = taste variables Barro's concept of "price" was used here. That is, price is the amount of tax dollars per capita that must be raised locally to increase total current revenues one dollar per student. Barro's combined price effects term o Using data from a sample of 176 Michigan school d is tric ts for 1972-73, a number of equations were estimated. These equations re flected the pre-district power equalizing situation. Then the values of selected variables and formula parameters were modified to replicate the implementation of the district power equalizing formula 2 in Michigan. The R for many of these equations ranged from .73 to .77. Various levels of total expenditures, local expenditures, and local tax rates were estimated under four situations. These were: (1) using actual figures for 1972-73 under the previous formula, (2) simulation of district power equalizing in which the DPE ratio or average state-local revenues per pupil for each mill of equalized tax rate = 38, (3) simulation of district power equalizing with a DPE ratio = 45, and (4) estimates from actual 1973-74- tax rates. Thus, changes in various variables and the relationships between variables from the pre-DPE year to the f ir s t year of the DPE were compared. Also, simulations were compared with estimates based on actual 1973-74 tax rates as a check for accuracy. Among the many findings of this analysis were the following: 1. The model appeared to predict fairly accurately as com pared to the first DPE year data. . 2. The correlation between property valuation and total expenditures per pupil dropped after the change to the d is tric t power equalizing formula. 3. The correlations between total expenditures per pupil and both median family income and percent of families with greater than $15,000 annual income were constant through the change. 56 4. Results indicated a slight increase in total expenditures per pupil in higher property value districts and a relatively larger increase for the same variable in the lower property value districts a fte r the change. 5. Increasing the per pupil base from $38,000 to $45,000 resulted in a slightly greater pattern of equalization accompanied by general decrease in the level of local tax rates. 6. The highest local tax rates are paid by districts with either high property wealth or high income levels and the lower tax rates are paid by lower property wealth or lower income districts. The model developed in this study and its results are particularly relevant to the present study because the Michigan formula and its phase-in are similar to the Ohio situation. The studies reviewed thus far were either simple determi nants models or more complicated statistical models based on various aspects of economic theory. Their puposes were to analyze the relationships between school spending and other school, demographic, and socioeconomic variables and also to predict fiscal responses of school d istric ts to changes in state aid formulas. We now turn to another part of the literature which includes studies in which the effects of state aid formulas are evaluated on various criteria. Evaluations of State Aid Formulas Perhaps the earliest study of this type was done in 1905 when Elwood Cubberley observed the considerable disparities in fiscal capacity and tax effo rt among local school d is tric ts . It was this observation that lead Cubberley to the opinion that it was the state's responsibility in apportioning state school funds to make sure each district could provide at least an acceptable minimum level of in struction (1905, p. 17). Prior to the 1970s, many of the studies that attempted to evaluate the variation and equity of school spending used samples of districts from various states aggregated to the state level. Perhaps the most significant of these studies was the work done for the National Educational Finance Project by Johns and Salmon. In more recent years, with the impetus of school finance litig atio n and the movement for more equity in state aid distributions, many studies have been conducted to evaluate state aid formulas within the individual states. Using d is tric t level data, these studies look at the distribution of state aid and total expenditures per pupil and at the relationships between expenditures and other district variables. It is within this group of studies that the present study should be classified. First, the Johns and Salmon study will be discussed. Then a representative sample of the within-state studies will be reviewed. Johns and Salmon. — One of the major studies evaluating equalization of state school finance systems was conducted by Johns and Salmon in 58 1971. Using a typology developed for the National Educational Finance Project (NEFP), these investigators studied the school finance plans of all fifty states using 1968-69 school year data. State and local funds were classified into five levels of equalization. These levels can be summarized as follows: Level 0: State funds are allocated so as to leave districts with the same or greater differences in financial capacity as they were before receiving state aid. Level 1: State funds are allocated on a flat grant basis without taking into account necessary variations in unit costs or local taxpaying ability. Level 2: State funds are allocated as flat grants which take into consideration necessary unit cost variations but not variations in local taxpaying ability. Level 3: State funds allocated in the form of uniform flat grants without taking into consideration necessary unit cost variations but take into consideration variations in local taxpaying ability. Level 4: State funds are allocated so as to take into account both variations in unit costs and variations in local taxpaying ability. The state, local, and state and local funds combined, and the percentage of funds for each state were calculated for each level of equalization. Then an equalization score based on the percentage of total state and local funds in each level was calculated for each state. The results indicated that each state was achieving some degree of equalization with the percentage of funds in equalization levels ranging from 100 percent in Level 4 for Hawaii to 72.2 percent in Level 0 for Connecticut (for Ohio the percentages of total state and local funds were Level 0 = 33.7 percent, Level 1 = 0 percent, Level 2 = 1 percent, Level 3 = 0 percent, and Level 4 = 65.3 percent). 59 The equalization scores ranged from 8.4 for Hawaii to 2.295 for Connecticut with a mean for all states of 5.131 (Ohio ranked seventeenth among the states on the equalization scores with a score of 5.882). The correlation between the percentage of funds provided by the state and the equalization score was +.646. The authors maintained that this supports the notion that when the state assumes primary responsibility for funding schools, greater financial equity is usually attained. Perhaps the most important conclusion that is derived by Johns and Salmon is that the extent to which financial resources for education are equalized depends more on the content of the plan and its level of funding than it does on the type of plan used. Caesar, et a l. — A 1974 report by Gene Caesar analyzes the first year results of school finance reform in Michigan. Beginning in 1973-74, Michigan changed from a Strayer-Haig type formula to a d is tric t power equalizing type formula in an attempt to-narrow the revenue gap between property wealthy districts and property poor districts. Also accompanying this change was the implementation of "circuit-breaker" tax relief which along with the incentive effects of the new state aid formula was aimed at narrowing the tax effort gap. The basic elements of the formula provided for a three-year phase-in with the state guaranteeing each local district an amount of combined state membership aid and local revenue equal to $39 per pupil per mill for the first 22 mills levied for operating purposes in 1973-74, $39 per pupil per mill for the f ir s t 25 mills in 1974-75, and 60 $40 per pupil per mill with no limitations for 1975-76. Transition guarantees were also implemented so d istric ts would not experience losses in total dollars. Caesar assessed the results of the first year implementation of these reforms by closely examining changes in school expenditure patterns, individual district millage rates, and effective property tax rates. Caesar found that the disparities in basic revenue per pupil among Michigan's 600 school d istric ts had been narrowed by looking at the frequency distribution of school d istric ts by basic revenue per pupil and noting that while the median basic revenue per pupil for the bottom 30 percent had increased 23.6 percent, the comparable figure for the top 30 percent had increased by only 9.4 percent. Similar analyses was done within these subdivisions of the distribution with similar results being obtained. In terms of total expenditures, Caesar found that the state's share increased from 46 percent to 51 percent. The passage rate for school millage issues was found to have increased to nearly 80 percent. The author attributes this increase to the incentive effects of the equal yield formula since 72 percent of the millage increases came in d is tric ts below the reimbursement ceiling. Caesar analyzed changes in millage rates in terms of d is tric t expenditures and district property valuations. He concluded that many districts were able to reduce their millage because of increased state aid. These reductions were for the most part in d istric ts with pre viously high tax rates and low property tax bases. On the other hand, 61 it was discovered that most of the millage increases came in districts with low mi 11 ages and expenditures. Caesar concluded that the first year results indicated some measure of success in bringing greater equity in school funding to both students and taxpayers. With respect to the future, Caesar stated (p. 32): ...And although there are many unknowns in the years ahead, there is still every reason to believe the favorable trends in school-finance and property-tax policy which have already started will continue. It is interesting to note that in a 1977 article, Caesar expressed concern that after four years of the Michigan Equal Yield Plan, the tax effort gap and the revenue gap had both widened to greater disparities. Looking at distributions of tax rates and expenditures and comparing the 85th percentile to the 15th percentile (both including and excluding Detroit), Caesar made the following observation: The 1976-77 figures not only show the largest revenue increases occuring in the highest-spending districts; they also show, for the first time in four years, the biggest millage hikes being imposed upon the very tax payers who already bear the heaviest burden (p. 9). Caesar points out that while disparities due to variations in tax base were narrowed, the disparities due to property tax effort grew worse. He goes on to suggest some revisions in the formula to allevi ate these tax effort-caused disparities. This situation is very similar to the one which Ohio now faces and which will be discussed later. 62 Vincent and Adams. - This 1978 report by Phillip E. Vincent and E. Kathleen Adams summarizes the results of a study of the fiscal responses of school d is tric ts to reforms in the states of Colorado and Minnesota. Using linear and log-linear regression equations, the authors analyzed the relationships between various socioeconomic, school, and state formula variables and both total and local expenditures per pupil. Also, two measures of equity, the coefficient of variation and the correlation between expenditures and property values were calculated for before and after reform years. After the 1971 reform, Minnesota continued to use a basic foundation type formula, but the foundation level guaranteed was raised 50 percent from fiscal year 1971 to fiscal year 1972 with no required local tax rate. In 1973-74, incentives were added for low spending districts. Levy limits were imposed in conjunction with property tax relief. In 1973, Colorado's modified foundation program was changed to a modified guaranteed tax base formula with a limited range of applicability and with a minimum guarantee of nine dollars per pupil per m ill. In the separate year-cross-sectional analyses, most indepen dent variables had their expected signs. However, for state equali zation in Colorado in 1973, total spending was estimated to decrease $1.38 for each $1 of additional aid per pupil. The authors express concern over this result and suggest that the results were confounded because expenditures were positively correlated with assessed valuation and equalization aid was negatively correlated with expenditures. 63 Using a regression analysis technique in which the dependent variable was a set of differences in expenditure levels from 1976 to 1972 and the independent variables were sets of differences between 1972 and 1976 values for the explanatory factors, the authors esti mated that districts would spend an additional $.49 out of every additional dollar in state equalization aid. The same differencing analysis between years indicated that there may be some substitution over time of state categorical aid for local expenditures that was not indicated by the cross-sectional analysis. The change-over-time analysis for federal aid indicated that federal aid stimulated local spending in Colorado and was substituted for local spending in Minnesota. The authors provided no explanation for this difference. In assessing movement toward equity and fiscal neutrality, Vincent and Adams analyzed changes in the coefficient of variation and the correlation between total expenditures and property wealth. In Colorado, the coefficient of variation dropped from .40 in 1973 to .35 in 1975 after two years under the new formula. The correlation between total expenditures and assessed value of property dropped from .89 to .81. For Minnesota, the coefficient of variation was .16 in 1972 and was .16 in 1976. The correlation of expenditures per pupil with assessed value per pupil went from .15 to .07. This study shows that both states experienced some success in moving toward equity and fiscal neutrality. 64 Sherman and Long. — This 1978 study by Joel D. Sherman and David C. Long was conducted with two main objectives—to evaluate Louisiana's present school finance structure with two equity criteria and to evaluate changes on both of these dimensions of equity over the la st few years. Using data for the 64 parish and 2 municipal school d istric ts from 1970-71 to 1975-76, the authors constructed frequency d is tr i butions of the district based on total state and local revenue per pupil. They then assessed the disparities in total state and local revenue per pupil using the range from lowest to highest, the range from the 95th percentile to the 5th percentile and the federal range statistic (the range of the 95th to the 5th percentile divided by the 5th percentile). It was demonstrated by all three measures that disparities had increased from 1970-71 to 1975-76. To analyze the relationship between district wealth and school d is tric t revenues to measure fiscal neutrality, Sherman and Long developed three measures of wealth. This had to be done because, unlike most states, in Louisiana about 30 percent of the local school district revenues come from property taxes while about 57 percent come from local sales taxes. The three wealth measures used were equalized assessed valuation per pupil, yield of a one percent sales tax per pupil and per capita income. The assessment of the equity of Louisiana's modified minimum foundation program indicated that the range in each of the wealth measures was quite large for property ($40,255) and sales tax ($312 for 1 percent) and fairly large for income ($1094 per capita). 65 Sherman and Long used 1975-76 data to divide d istric ts into quintiles based on property wealth per pupil and various statistics were calculated for each of the five groups of d is tric ts . These statistics included assessed valuation per pupil, property tax rate, property tax revenues, sales tax revenues, property plus sales tax revenues, and local revenues per pupil. A similar analysis was per formed using sales tax bases per pupil to rank the districts into quintiles. In both analyses, state aid was found to be equalizing, but this effect was substantially offset by local revenues. Correlations between selected school finance variables and equalized property valuation per pupil were calculated. The corre lation between equalized property valuation per pupil and state and local revenues per pupil was +.33 and was significant at the .003 level. The authors conclude that the Louisiana system has failed to achieve equity and is moving farther away from equity than it was in previous years. They then cite several reasons for this and go on to make some recommendations for revising the system. Hickrod, et al. — This study reported by G. Alan Hickrod, Ben C. Hubbard, Thomas W. C. Yang and Tharin Rasanond was the third in a series of evaluations of the 1973 reforms in the Illinois state aid formula (1976). The f ir s t study provided a historical background of the Illin o is school finance program, described the evaluative c rite ria , and applied these criteria to the first year of operation (Hickrod, Hubbard & Yang, n.d.). The second study was basically a revision of 66 the first. The third study is an extension of the first two and evaluates the new state aid formula after three years of operation. Prior to 1973, state aid was provided to the three types of school d is tric ts (elementary, high school, and elementary and high school combined) using a Strayer-Haig type foundation plan. In 1973, not only was the foundation plan revised but school districts were also given the option to receive funds through a district power equalizing type formula. The district power equalizing (DPE) schedule operates only up to ceilings of 30 mills for unit districts, 19.5 mills for elementary districts and 10.5 mills for secondary districts. Flat grants are also available for property rich districts that do not qualify under the DPE schedule and no recapture provision is appli cable. These reforms were phased in over a four-year period. Hickrod et al. selected three c rite ria to evaluate the school finance system. These were permissable variance, fiscal neutrality, and reward for effort. Permissable variance in total expenditures per pupil was measured using the coefficient of variation (the standard deviation divided by the mean times 100). A second measure, the McLoone Index, measured the percentage, of students found below the median expenditure level and the amount of dollars needed to bring all students to the median expenditure in the state. Looking at the coefficients of variation for the pre-reform year of 1972-73 through 1975-76, the authors found that variance dropped after the f ir s t year for all types of d is tric ts and continued to drop for the unit and high school 67 districts. For elementary districts variance increased after the second and third years but stayed below the pre-reform level. The McLoone Index analysis indicated the reform was successful in bringing students up above the median expenditure level for unit and high school d istric ts. The elementary d istric ts showed no improvement. Fiscal neutrality was measured in two major ways. Gini coefficients of Lorenze curves were derived by plotting a cumulative distribution of wealth per pupil against a cumulative distribution of state and local expenditures. Under an expanded interpretation of fiscal neutrality pupils were weighted according to need. Also, a regression with expenditures as the dependent variable and wealth as the independent variable was run.' The Gini analysis using property valuation per weighted pupil indicated that progress was made toward fiscal neutrality in the first year but tapered off slightly in the second and third years for the elementary d istric ts. For high school d is tric ts , more progress was made in the first and third years than in the second. The unit d istric ts showed more progress in the second and third years than in the firs t. Using income per weighted student as a wealth measure, the elementary d is tric ts were shown to have progressed toward fiscal neutrality especially in the first and third years. For the unit districts, there was movement away from fiscal neutrality in the first two years, and then in the third year, movement toward fiscal neutrality. 68 The regression of expenditures against property valuations and against income per weighted pupil showed all three types of dis tric ts moving toward fiscal neutrality. The next part of this study involved discriminant analysis of the relationships between various socioeconomic, demographic, and school variables and passage rates of school tax levies. Passage rates were compared for periods before and after the reform with results indicating that the passage rates and frequencies of tax issues were greater after the reform. It should be noted that the discriminant models explained only between 16 and 34 percent of the variance in percentage of favorable votes. The authors indicate that the passage rate results could indicate but do not prove that the reward for effort effect of the DPE formula had some effect. This study is particularly relevant since some of the c rite ria used to evaluate Ohio's school finance formula in the present study are the same as those used in this Illinois study. Yang. — The purpose of this 1975 Ph.D. dissertation study by Thomas W. C. Yang was to investigate the degree to which equity was achieved as a resu lt of school finance reform in the states of Illin o is, Michigan and Kansas.- The author defines equity as a two-fold notion that (1) the same amount of school revenue should be available to each educational need unit, and (2) that school revenue per educational need unit should not be a function of district wealth. These are the familiar permissible variance and fiscal neutrality criteria used in the Hickrod et a l ., study previously reviewed. As in the Hickrod 69 study, the coefficient of variation and the McLoone Index were used to measure the variation in expenditure per unit of need. For the fiscal neutrality analysis the Gini coefficent- Lorenze curve analysis and "parallel regression analysis" were employed. Gini coefficients were derived from Lorenze curves arrived at by plotting a cumulative percentage of aggregate school revenue of districts sorted in ascending order by wealth against the cumulative proportion of the districts educational need units. Since the effects of state aid were of primary importance in this study; components of total expenditures were added successively on the vertical axis. The effect of the changing Gini index was identified and attributed to state aid and other components of total expenditures per need unit. In the "parallel regression analysis" Yang first regressed d is tric t revenue on wealth per need unit. Then he added state aid to local revenue and regressed the sum on wealth per need unit. The changes in the e la stic itie s of wealth with respect to the different dependent variables were attributed to the added portion of the dependent variable. Using data from the pre-reform year 1972-73 and from post reform years 1973-74 and 1974-75 for all districts in each state, comparisons were made of the various measures of equity on a "before and after" basis. The coefficient of variation showed improvement in the effect of state aid on equity in all three states. The McLoone Index indicated "leveling up" had occurred except for elementary districts in Illin o is, all d istric ts in Michigan and small d is tric ts in Kansas. 70 The Gini index analysis indicated movement toward fiscal neutrality in all three states. The wealth elasticity analyses also indicated movement toward fiscal neutrality in all three states. The author concluded that in each state the reform had moved the state toward equity and fiscal neutrality but cautioned the reader that this conclusion is based on only two years experience under the new plans. Carroll. - This study, reported in 1979 by Stephen J. Carroll of Rand Corporation in three volumes, analyzed the effects of recent school finance reforms in California, Florida, Kansas, Michigan and New Mexico. Its two main objectives were (1) to describe the pre- and post-reform school finance outcomes in the individual states and to compare the differential outcomes of the reforms, and (2) to estimate the effects of fiscal and other factors on school districts' budgetary behavior. Data were obtained from state sources on districts' revenues, expenditures, tax rates, wealth and numbers o.f pupils for school years 1972-1973 through 1976-1977. Community characteristics were obtained from the 1970 Census, fifth count corresponding to d is tric t boundaries as of 1974. These characteristics were household income, urbanity, race, and poverty level. To facilitate interstate comparisons, district property values were equalized using sales- assessment ratios. 71 Prior to reform all but one of the five states distributed funds under some form of foundation plan. New Mexico had a flat grant program. After reform, California and Florida retained their foun dation programs with several modifications, while Kansas and Michigan adopted Guaranteed Tax Base plans. New Mexico assumed full financial responsibility with a combination of state funds and a state mandated local levy. In the descriptive part of the analysis revenues per pupil, instructional expenditures per pupil and adjusted tax rates for each state were compared before and after reform. Also, the coefficient of variation was compared on a before and after basis. In general, results from these analyses indicated the following: 1. All states experienced significant growth in revenues and,with the exception of Michigan and Kansas' small and medium districts, experienced a drop in tax rates, 2. Instructional expenditures per pupil have grown in every year in every state except California, but their rate of growth is lower than the rate of growth of revenues. 3. Reform efforts have brought about somewhat more^equal distributions of per-pupil revenues and more equal distribution of adjusted tax rates. 4. Disparities in per-pupil instructional expenditures have been reduced very little. 72 To determine if reform had any effect on fiscal neutrality, Carroll looked at elasticity coefficients between revenues (or expenditures) and fiscal capacity and between revenues (or expendi tures) and fiscal effort. Fiscal capacity was measured alternatively by the property tax base per pupil and household income per pupil and fiscal effort was measured by the property tax rate. General conclusions drawn from comparing e la stic ity coefficients were the following: 1. Reform has generally reduced the degree to which revenues depend on wealth and instructional expenditures per pupil were generally as closely related to property tax basis after reform as they were before. 2. In some cases income neutrality was improved and in other cases it was not. 3. Revenues and instructional expenditures are more closely related to tax effort in some cases, but less closely in others. To assess changes in budgetary behavior of school d istric ts econometric regression models were applied to Michigan data. Expendi tures per pupil was the dependent variable in these regressions. Various combinations of types of grants (federal or state, categorical, unrestricted block, or matching), the prices of revenues per pupil (matching rates), and d is tric t population characteristices were the independent variables. 73 Looking at the various elasticities, the author concluded that there was no evidence that state matching or block grants did much to stimulate school district expenditures. However, federal and state categorical aid had a much stronger effect. The study concluded that two factors explain why finance reforms accomplished relatively little in equalizing spending out comes and opportunities. First, reform attempted to accomplish objectives that were conflicting such as trying to equalize revenues and avoiding political hazards of reducing spending of high-spending districts. Secondly, each state tended to make add-ons and adjustments to its basic school finance plan. Ohio Evaluation Studies This la st group of studies to be reviewed is perhaps the most relevant to the present study because these studies are fairly recent evaluation studies of the Ohio school finance system. With the exception of the first, they all deal with evaluating the effects of the Equal Yield Formula as does the present study. Ohio Legislative Service Commission. — This 1972 study conducted by research staff of the Ohio Legislative Service Commission was reported to the Ohio General Assembly to keep them apprised of developments re lating to the possible impact that Serrano v. Priest and other court decisions applying to state school finance systems might have on Ohio public school finance. A detailed review of court decisions relating the equal pro tection provisions of the U.S. and state constitutions to public school finance is first presented. The legal standards formulated by the courts in declaring state aid systems unconstitutional are explained. These standards are then shown to be related to the provisions in the Ohio Constitution for equal protection and provision of a thorough and efficient system of public schools. Using 1970-71 data at the school d is tric t level, the authors analyzed disparities in taxable valuation, per pupil expenditures and available revenues per pupil by looking at the range and the five lowest and five highest d is tric ts ranked on the various measures. The authors summarized this analysis as follows: 75 Analysis of financial data from Ohio school d is tric ts reveals significant disparities in taxable wealth per pupil, current expenses per pupil, teacher-pupil ratios, and wealth available per pupil from state and local revenue sources combined. They also show that wealthy districts with low rates can often raise more revenue per pupil than poorer districts with higher tax rates (p. 20). Comparing these taxable wealth and expenditure per pupil disparities with those in other states, the authors conclude that: ...the disparities in per pupil taxable wealth and per pupil expenditures in Ohio are comparable to those in states where courts have ruled that school finance systems do not meet the Equal Protection Test. ... Financial data of Ohio school districts suggest that Ohio school financing is vulnerable to attacks based on the Equal Protection Test (p. 28). The study goes on to discuss various alternative revisions that could be made in the Ohio school finance system to meet the equal protection test. These included full state funding, district power equalizing and district reorganization. Harrison. — The results of this study are contained in a 1976 staff report to the Education Review Committee of the Ohio General Assembly written by William A. Harrison, Jr. The purpose of this report was to review the results achieved after one year under the new Equal Yield Formula. Using 1975-76 data a t the d is tric t level, Harrison looked at the distributions and levels of such variables as total state and local revenue per pupil, dollars per pupil per mill, disadvantaged pupil impact aid, passage rates of school levies, basic aid, cate gorical aid, and enrollment. Some of the findings of this study included the following: 76 1. Total state and local revenues per pupil increased in nearly all school d istric ts from 1974-75 to 1975-76 2. While many districts with low to average property wealth per pupil showed increases due primarily to increases in basic and categorical state aid, the largest increases were due to local reve nue growth 3. Local revenue plus basic state aid per pupil increased in all but 12 districts 4. Total state and local support per pupil ranged from $732 in Huntington Local to $3,577 in Bratenahl Local 5. Total state and local support increased from $945 per pupil in 1974-75 to $1,071 per pupil in 1975-76 6. In 1975-76 the number of dollars per pupil per mill in combined local and basic state aid ranged from $283 in Cuyahoga Heights to $21.82 in Windham Exempted Village. Although the level of yieldper pupil per mill increased in most districts from 1974-75 to 1975-76, there was l i t t l e change in the distribution of fiscal capacity among districts. Harrison concludes that the first year data indicate that movement toward the objectives of directing state support to the state's most urgent educational needs, linking educational opportunity to tax effort rather than property-wealth,providing incentives to exert more local tax effort, and increasing equity in tax burdens was questionable. Gensemer. — This 1977 paper was prepared by Bruce L. Gensemer as one of several background papers for the Citizens' Council for Ohio 77 Schools April, 1977 Conference on School Finance. The first section provides an analysis of school finance trends from 1970 to 1975. This period was described as one of large increases in state aid to edu cation due mainly to the enactment of a state income tax. However, Gensemer pointed out that this period was also characterized by sharp increases in the costs of education so as to almost completely offset the effects of the growth in state aid to education. The 1975-1977 period was described as one of marked in crease in the rate of growth of state aid coupled with more intense inflationary pressures on school costs. In order to analyze the distributional impact of the newly implemented Equal Yield Formula, school d istric ts were grouped on several different dimensions. Looking at quintiles of districts ranked on equalized property value per pupil, Gensemer found that the difference in average state basic aid between the highest quintile and the lowest quintile had grown smaller after two years under the Equal Yield Formula. Looking at district quintiles ranked by current operating mi 11 age and at passage rates of school operating levies from 1974-75 to 1976-77, Gensemer concluded that "the formula succeeded in re warding those districts already levying high millage but failed to. encourage other districts to raise their millage rates." Finally the author grouped d is tric ts as large city, s a te llite city (suburban), independent small city , and rural d istric ts. The distribution of basic state aid per pupil among these four types of districts was found to be fairly constant with one exception. The satellite city districts gained at the expense of the other three types of districts. Gensemer's final conclusion was that a reform in the formul to distribute state aid can accomplish very little unless it is ac companied by substantial increases in state funding. It should be noted that the conclusions reached by Gensemer parallel those reached by Harrison in the previously reviewed study. Bowers. — This 1977 study conducted by G. Robert Bowers in conjunc tion with School Management Institute assesses the extent to which fiscal year 1976 school expenditures in Ohio were wealth neutral. A mathematical formula for determining the percentage of total local and state education funds which were wealth neutral was employed. This formula was expressed verbally as follows: Local Local and State Other Percent Wealth Neutral = Leeway + Equalization + State Revenues ______Funds ______Funds X 100 Total Local and State Funds A five-district model state consisting of New Boston Local, Canton Local, Hamilton City, Lexington Local and Miller City-New Cleveland Local school districts was formed to calculate the percent wealth neutral for the 17 percent, 26 percent, 45 percent, 75 percent and 100 percent phase-in levels of the Equal Yield Formula. Using fiscal year 1976 data in the percent wealth neutral formula, the percentages shown in Table 3 of the five-district state total edu cation funds that were wealth neutral were calculated for each phase- in percentage. 79 TABLE 3 Percentage of Funds Wealth Neutral By Phase-In Level For Five D istrict State Model Percentage of Funds That Are Phase-in Percentage Wealth Neutral______ 17 83.45 26 85.14 45 88.99 75 94.08 100 98.54 Then Bowers calculated the percent wealth neutral funds for Ohio using all 617 school d istric ts for the 17 percent phase-in, the 17 percent phase-in recalculated for state budget cuts, and for 100 percent phase-in. These calculations yielded the results shown in Table 4. TABLE 4 Percentage of Funds Wealth Neutral For Selected Phase-in Levels: All 617 D istricts Percentage of Funds That Are Phase-in Percentage Wealth Neutral______ 17 76.04 17% less reductions 74.95 100 . 96.46 Thus, looking at state and local funds in the aggregate, Bowers concluded that the Equal Yield Formula was about 75 percent wealth neutral in fiscal year 1976 and would have been 96 percent wealth neutral i f i t had been fully funded. 80 Sherman. — This 1978 study by Joel D. Sherman of the Lawyers' Com mittee for Civil Rights Under Law was conducted with several purposes. These included (1) describing and analyzing the manner in which school revenues are collected and distributed in Ohio, (2) evaluating the Equal Yield Formula i f i t were fully funded, (3) assessing the impact of the school finance structure on educational services in Ohio, and (4) analyzing the impact of the school finance structure on Ohio's cities. Using 1975-76 data, Sherman looked at the distributions of property wealth, tax effort, state aid and median family income. D istricts were divided into deciles based on equalized valuation per pupil and the levels of other variables such as number of FTE, state revenue, local revenue, and state and local revenue. Some of the major findings of this analysis include the following: 1. The ranges of taxable property value per pupil between the lowest and highest district and between the 90th percentile and the 10th percentile were substantial 2. The ratio of operating expenditures per pupil in the highest spending district to the lowest spending district was about 5 to 1 3. There is a strong positive relationship between d is tric t property wealth and educational revenues per pupil 4. Even i f the Equal Yield Formula were fully funded, significant wealth related revenue disparities among school d istricts would remain 81 5. The revenue disparities due to differences in property wealth had substantial impact on educational services as indicated by pupil-teacher ratios, pupil service personnel, library staff, and materials and equipment. High revenue districts had higher levels of educational services 6. While the needs of the large city students are much greater than those in surrounding districts, expenditures per pupil are only slightly higher. These conclusions are similar to those reached by Harrison and Gensemer in the studies previously reviewed. Project 842. — This extensive study was conducted by a set school finance experts and economists serving as consultants for the Ohio Department of Education (School Management In stitu te, nd). Its pur pose was to evaluate the effectiveness of Ohio's Equal Yield Formula using nine equity questions derived from a set of equity questions formulated by the U.S. Department of Health, Education and Welfare. These questions can be summarized as follows: Does the finance plan 1. ensure that educational expenditures reflect the wealth of the state as a whole? 2. take into account needs of individual pupils? 3. take into account variations in cost of providing education among districts? 4. include all necessary and required programs — supportive service and facility costs? 5. encourage citizen interest and participation in the educational decision-making process? 6. encourage efficiency in size and organization of school d is tric ts and the delivery of educational programs to promote equality in educational opportunity? 7. provide for evaluating the achievement of its objectives? 8. Is the finance plan financed by a tax system which is equitable? 9. Does the state determine a level of educational pro grams, services and fa c ilitie s to assure equal educational opportunity for all pupils? Using data for fiscal year-s 1973 through 1977, several types of analysis coupled with expert judgment were used to answer these equity questions. Of particular relevance to the present study is the analysis for question 1 above performed and reported by Bruce Gensemer (see Gensemer, 1978). The purpose of this study was to assess the extent of fiscal neutrality of Ohio's school finance system and to assess the equity of the fiscal capacity measure used. To examine the fiscal neutrality question-, the distribution of current operating expenditures among a stratified random sample of 146 school d is tric ts was analyzed. Two alternative measures of fiscal capacity (equalized property tax value and a composite of property wealth and personal income) and both unweighted FTE ADM and FTE ADM weighted by incidence of poverty in the d is tric t were used in this analysis. To look at the relationship between fiscal capacity and school expenditures, four techniques were used. These were: (1) regression analysis, (2) Lorenze curves and Gini indexes, (3) quintile analysis, and (4) examination of the distribution of reve nues by four classifications or types of school d istric ts ; i . e . , central city, independent small city, satellite city, and rural districts. The various measures of fiscal neutrality were calculated for each year from 1974-75 through 1977-78 to determine trends of movement. Gensemer concluded that regardless of the fiscal capacity measures used, Ohio moved away from the goal of fiscal neutrality over this period of time. The author cited the small state share of total state and local revenue, the correlation between high income and the passage of incentive millage in wealthy districts, save harmless guarantees, rollback aid to wealthy districts and the exclusion of an income factor in measuring fiscal capacity as the causes of not attaining fiscal neutrality. Gensemer also pointed out that uneven assessment practices and the effects of H.B. 920 have contributed to this problem. The following recommendations were then made to move Ohio toward fiscal neutrality: 1. Increase the state share of total state and local revenues to 70 percent 2. Eliminate rollback aid and direct these funds to basic aid 3. Allocate basic aid using a revised measure of fiscal capacity 84 4. Eliminate the guarantees 5. Include need and cost factors in the basic aid distri bution formula. It should be noted that some of the same analysis techniques and data used in this study by Gensemer will be employed in the present study. In summary, this brief review of literature related to the present study has shown the following: 1. Expenditure determinants study results vary with respect to the sign and significance of variables hypothesized to influence school expenditures but fiscal capacity (either income or property wealth) appears to be positively related to school expenditures 2. S tatistical models based on prior research and economic theory have provided us with prediction and analysis tools useful in school finance reform 3. Various evaluation studies of state school finance systems have indicated varying levels of success for reforms intended to increase equity and equal educational opportunity. Equal yield type formulas appear to have moved other states toward partial fiscal neutrality in several cases 4. All but one of the Ohio evaluation studies indicated that revenue disparities among districts and an inadequate level of funding from the state level were in the past and continue to be un solved problems which plague Ohio's school finance system With this as a context from which to work, le t us now turn to a discussion of the research problem and data analysis procedures of the present study. CHAPTER III METHOD This study sought to determine if certain expectations have been realized with regard to the effects of the Equal Yield Formula. These expectations have been formulated into three hypotheses which were tested using a variety of techniques. Hypotheses I and II were tested by looking at various measures of variation and fiscal neutrality for the two most recent years prior to implementation of the Equal Yield Formula and then for the first four years under the Equal Yield Formula. The unit of analysis of these analyses is the school district of which there are at present 615 in Ohio. Hypothesis III was tested by analyzing the results of a questionnaire sent to 454 school district superintendents regarding their opinions concerning the influence of the Equal Yield Formula on their districts' operating levies. The analysis for Hypotheses I and II involved financial and enrollment data for school years 1973-74 through 1978-79 with 1978-79 expenditures being preliminary estimates. These data were obtained from either hard copy or data tape files of the Ohio Department of Education, Division of School Finance and Division of Computer Services. Some data from the 1970 U.S. Census was employed and was obtained from the data file s of Dr. Bruce Gensemer at the Academy for Contemporary Problems, Columbus, Ohio. 85 86 In some cases data transformations had to be made to make certain variables comparable across all six years. A description of these transformations and a listing of all variables used are found in Appendix A. The remainder of this chapter will include a restatement of each hypothesis and a detailed discussion of the analysis techniques that were used to te st each hypothesis. Variation Analysis Hypothesis I: The variation in operating expenditures per pupil among school d istric ts will be found to have been reduced under the Equal Yield Formula. In order to test this hypothesis, distributions of total current operating expenditures per pupil for the school years 1973-74 through 1978-79 were examined. The f i r s t two years represent pre-Equal Yield Formula (EYF) years. There are various criteria that can be used to assess the equity of a distribution. For this hypothesis we are concerned with those criteria that measure variation. Not all changes in a distribution can be detected by a single measure. Thus, several measures of variation were employed to assess movement toward less variation from the pre-EYF period to the EYF period. Each of these measures will be discussed in turn. Restricted Range. This measure presents the difference between the 5th and 95th percentiles of the distribution. Although this measure is in sensitive to a large number of movements toward equity such as equal additions of revenue to each d is tric t or transfers of revenue from d istric ts above the median to d is tric ts below the median, i t is 87 sensitive to equal percentage increases of revenue to each d is tric t. That is i f each d is tric t were to receive an equal percentage increase in its revenues, the restricted range measure would increase indicating larger variation. A decrease in the restricted range will indicate movement toward less variation. See table 5. Federal Range Ratio. This measure was proposed in recent federal regu lations (see Federal Register March 22, 1977, Part li). It is expres sed as the following ratio: FRR = 95th percentile - 5th percentile 5th percentile As shown in table 5, unlike the restricted range the federal range ratio is sensitive to equal additions of revenue to each dis trict, but is insensitive to equal percentage increases in revenue to each district. A decrease in this ratio will be assumed to indicate a movement toward a more equitable distribution of revenues. Permissible Variance. This measure was developed by Eugene McLoone and will be called the McLoone Index (1974, p. 97). It is given by a ratio with the numerator being the actual spending of d istric ts below the median expenditure level, and the denominator being what the total spending of all d istric ts below the median would be i f each were raised up to the median expenditure level. It may be expressed as follows: i=l McLoone Index = ------J M ^P f 1=1 where J = number of d istric ts below the median level of per pupil expenditures X. = total operating expenditure per pupil in district i P. = number of students in district i M = median level of per pupil expenditures Since this index is sensitive only to transfers across the median, transfers that take place below the median will not be detected as improvements by this measure. The closer this index is to one, the more equitable the distribution will be considered to be. Coefficient of Variation. This measure, used in several studies re viewed in Chapter II (see Hickrod et a l., 1976; Vincent & Adams, 1978; Yang, 1975), is the standard deviation of the distribution divided by the mean of the distribution and is given by the formula: N ^ P . , . ( u - x .)2 Coefficient of Variation = 1-~* ______x 100 N 89 where U = mean total operating expenditure per pupil of entire distribution N = number of districts in the state P.. = number of pupils in district i X.. = total operating expenditure per pupil in district i As indicated in table 5, the coefficient of variation is sensitive to a number of transfers and changes. However, it places the same weight on small transfers from one district to another regardless of where in the distribution these transfers occur. Some would argue that transfers at the lower end of the distribution should increase equality more than transfers at the upper end. Progress toward equity will be inferred if the coefficient of variation decreases. Standard Deviation of Logarithms. A measure that does weight changes a t the lower end of the distribution more heavily than at the upper end is the standard deviation of logarithms. The formula for this measure is the following: N $ > . (log U - log X^2 i=l Standard Deviation Logarithms3 ------N ^ pi i=l 90 where the natural logarithm is employed and U = mean total operating expenditure per pupil of entire distribution N = number of districts in the state P. = number of pupils in district i = total operating expenditures per pupil in district i As indicated in table 5, this measure is sensitive to the same changes and uses the same reference point (the mean) as does the coefficient of variation with the additional property that changes at different levels are recorded differently. Once again, lower values of this measure will be associated with movement toward equity. Table 5 below summarizes what units are taken into account, what reference points are used for comparison by each measure and the types of changes or transfers that affect the values of each measure. These factors should be kept in mind when evaluating the results of the variation analysis. TABLE 5 Characteristics of Selected Variation Measures Variation Measures Standard Federal Coefficient D eviation R e s tric te d Range Permissible o f o f Gini Selected Criteria Range R atio V ariance V a ria tio n Logarithm s Coefficient 1. All units taken into account? No No No Yes Yes Yes 2. Improvement for transfers on one side of the mean? No No No Yes Yes Yes 3. Improvement for transfers on one side of the median? No No No Yes Yes Yes 4. Improvement for transfers that cross mean? No No No Yes Yes Yes 5. Improvement for transfers that cross median? No No Yes Yes Yes Yes 6. Sensitive to equal additions? No Yes Yes Yes Yes Yes 7. Sensitive to equal percentage increase? Yes No No No No No 8. Changes at different levels recorded identically? No No No Yes No No 9. Mean for comparison? No No No Yes Yes No 10. Median for comparison? No No Yes No No No 11. All levels for comparison?* No No No No No Yes SOURCE: Adapted from Robert Berne, "Equity and Public Education: Conceptual Issues of Measurement", Working Paper No. 4, Public Policy Research Institute, New York University, New York, October, 1977, p. 21. 92 The variation analysis discussed above was done using total current operating expenditures per pupil and then was repeated using total current operating expenditures per pupil less basic aid. A comparison of the variation measures between these two analyses will indicate the extent to which basic aid is contributing to the vari ation. The distributions used in these analyses included all districts in the state. D istricts were then divided into four groups by d is tric t type using a classification system developed by Gensemer (1978, p. 64). These groups are (1) d istric ts associated with large central c itie s , (2) urban area d istric ts associated with cities near and dominated by large central cities, (3) districts associated with small independent cities, and (4) rural districts. A more detailed description of these groups is found in Appendix B. The variation analyses using all of the measures used above was conducted for each d is tric t group. Looking at the variation-measures within each of these more homogeneous groupings serves to at least parti.cially take into account such factors as dis trict size, sparsity and'cost factors since these should be similar for d istric ts within these groups. Taking into account such "need" and "cost" factors makes the measures of variations more accurate measures of equity. Fiscal Neutrality Analysis Hypothesis II: The Ohio system of funding public education will be found to be more fiscally neutral after implementation of the Equal Yield Formula than it was before implementation of the Equal Yield Formula. 93 As described in Chapter I, this study examined both ex post and ex ante fiscal neutrality. Movement toward ex post fiscal neutrality was assessed using two basic forms of analysis—regression analysis and Gini coefficient analysis. Ex ante fiscal neutrality was assessed using a combination of Gini coefficients. Simple Regression. F irst, a set of simple linear regressions was run in which expenditures per pupil and per weighted pupil were regressed on various measures of fiscal capacity per pupil and per weighted pupil. For the weighted pupil analysis, each pupil receiving Aid to Dependent Children (ACD) received a weight of 1.5 while all other pupils received a 1 weighting. This weighting attempts to account for the greater need for educational resources assumed to be present for poor (ADC) students. While there is general agreement among educators that this additional need of poorer students is real, there is no con sensus as to exactly how much more resources are required to serve disadvantaged pupils (for example see Busselle, 1973; Leppert and Routh, 1978; McLure, 1977; Patton, 1978). The 1.5 weighting for ADC pupils was used by Gensemer in Project 842 and is an arbitrary choice (1978, p. 7). The regression equations took the general linear form: 94 y = m x + b where y = expenditure per pupil measure m = regression coefficient x = fiscal capacity measure b = a constant term A separate regression was run for each of the six years and the relationship between the expenditure measure and the fiscal 2 capacity measure was compared for each year by looking at the R for 2 each regression. The R gives the percentage of the variance in the expenditure measure that can be associated with the fiscal capacity 2 measure. If the R decreases from year to year, the relationship between expenditures and fiscal capacity will be assumed to have les sened and this will imply that movement toward fiscal neutrality will have taken place. Table 6 shows the various combinations of dependent and independent variables that were used in these regressions. The use of weighted pupils and various fiscal capacity measures served as a check to the sensitivity of this analysis to the use of different measures. 95 TABLE 6 Combinations of Dependent and Independent Variables Used 1n Simple Regressions Dependent Var1able Independent Variable Measure of Expenditure Measures of Fiscal Capacity Total Operating Expenditure Per Pupil Adjusted Equalized Property Value Per Pupil (AEPVPP) Total Operating Expenditure Per Pupil Mean Family Income 1969 (MFI) Total Operating Expenditure Per Pupil Adjusted Gross Income Per 1976 Tax Return (AGI) Total Operating Expenditure Per Pupil AEPVPP weighted by MFI Total Operating Expenditure Per Pupil AEPVPP weighted by AGI Total Operating Expenditure Per Weighted Pupil Adj. Equalized Prop. Value Per Weighted Pupil (AEPVWP) Total Operating Expenditure Per Weighted Pupil MFI Total Operating Expenditure Per Weighted Pupil AGI Total Operating Expenditure Per Weighted Pupil AEPVWP weighted by MFI Total Operating Expenditure Per Weighted Pupil AEPVWP weighted by AGI Multiple Regression. While simple regressions indicate the relation ship between the fiscal capacity and expenditures, there are other variables whose effects on educational expenditures may be confounded in the fiscal capacity measure. In order to sort out these factors and get a clearer view of the separate effect of fiscal capacity on expenditures, multiple regression analysis was employed. The regression equations took the general linear form y = m x + m, f, + m0 f 0 + ------m„ f + b o i i £ z n n where y = expenditure measure m. = regression coefficients of independent variables x = fiscal capacity measure f. = demographic and socioeconomic factors thought to affect expenditures b = constant term 96 Various combinations of f.'s were included in the regressions with each of the fiscal capacity measures. The f.'s included the fol lowing variables listed by the type of effect on expenditures. The fiscal capacity measures used in these regressions involved property wealth alone and property wealth weighted by MFI or AGI just as in the simple regressions. TABLE 7 Independent Variables Included in Multiple Regressions with Capacity Measures Income Adjusted Gross Income Per 1976 Tax Return (AGI) Mean Family Income 1969 (MFI) Percent of Students from Families Below Poverty Level Cost Average Weekly Wage Percent of Students that are on ADC Percent of Students from Families Below Poverty Level Demand for Education Percent of Students Classified as Minority Percent of Students Attending Nonpublic Schools Percent of Population less than 18 Years Old Percent of Population that are College Graduates As with the simple regressions the variable of major interest in this analysis is the fiscal capacity measure. The other variables wereincluded in the equations to "control" for other factors that might affect expenditures on education. Therefore, attention will be focused on the regression coefficients of the fiscal capacity measures which will be compared over the six years. For any particular 97 regression it will be inferred that fiscal neutrality is being ap proached if the regression coefficient for fiscal capacity diminishes over time. This will mean that the relationship between expenditures and fiscal capacity will be weaker. It was this expenditure-fiscal capacity relationship which was the focus of many of the expenditure determinants studies reviewed in Chapter II and this regression analysis technique was used in numerous studies to assess fiscal neutrality (see for example Gensemer, 1978; Hickrod et a l ., 1976; Vincent & Adams, 1978; Yang, 1975). Now we will turn to the second major approach to assessing ex post fiscal neutrality. Lorenze Curves and Gini Coefficients. Another way to look at the re lationship between expenditures and wealth is to look at the distribution of expenditures spent on students ranked by the wealth of their respective school d is tric ts . This relationship can be expressed by using a measure originally used mostly in economics and more recently employed in school finance studies (see Barkin, 1967; Gensemer, 1978; Grubb & Michelson, 1974; Hickrod et a l ., 1976; Wilensky, 1970; Yang, 1975). This bivariate measure is the Lorenze curve and its associated Gini coefficient. To employ this analysis districts were sorted in ascending order on some measure of wealth per pupil. The cumulative proportions of pupils in the districts were plotted on the horizontal axis of a coordinate system. The cumulative proportions of some measure of expenditure accounted for by these districts was plotted on the 98 vertical axis. Then the Lorenze curve was plotted on the coordi nate system. If expenditures per pupil are the same for each d is tric t the curve plotted would be a straight 45° line. This line represents "perfect equality." (See Figure 2 below.) 100% CO ✓ A O 100% % of Students Ranked by Wealth Figure 2. Lorenze Curve and 45° Line Under these circumstances points on the 45° line indicate, for example, that the poorest 10 percent of the students receive 10 percent of the expenditures, the poorest 20 percent receive 20 percent of the expenditures and so on. If expenditures per pupil are less in the poorer districts than in the richer districts (which is usually the case), the Lorenze curve will fall below the 45° line as shown in Figure 2. The points on this curve could indicate for example that the poorest 10 percent of the students received only 6 percent of the expenditures and that the richest 10 percent received more than 10 percent of the expenditures. 99 It could be argued that the 45° line does not represent "perfect equality" because of differential needs of economically dis advantaged students who require additional resources for their edu cation. Thus, it would be argued that the poorest 10 percent of the students should receive more than 10 percent of the expenditures. In order to account for these differential needs and to make use of the Lorenze curve analysis, weighted pupils were used on the horizontal axis. As with the other analysis students on ADC were weighted 1.5 while non-ADC students received weights of 1.0. Using either weighted or unweighted pupils, fiscal neutrality will be measured by how far the Lorenze curve falls below the 45° line. The closer the Lorenze curve is to the 45° line the more fiscally neutral the system will be assumed to be. An index called the Gini index can be derived from the Lorenze curve and used to express the degree of fiscal neutrality. Referring to figure 2, the Gini coefficient can be given by the fol lowing: Area A G = ------Area (A + B) Thus, as the Lorenze curve moves closer to the 45° line, the Gini coefficient approaches zero. In this study in order to avoid dealing with decimals, all Gini coefficients were multiplied by 1,000. Actual calculations of the Gini coefficients were done using the following formula (Hickrod et a l ., p. 80): 100 n G = ^ (xi_i Yi “ xi Yi-i) x MOO 1=2 where: X. = cumulative proportion of pupils for the ith d is tric t Y. = cumulative proportion of expenditures for the ith d is tric t n = number of districts in the state To assess the movement toward ex post fiscal neutrality Gini coefficients were calculated for each of the six years and the values of the coefficients will be compared. If the values decline over the years, movement toward fiscal neutrality will be assumed to have oc curred. Five separate analyses were done using the following wealth per pupil measures: Adjusted Equalized Property Value Per Pupil (AEPVPP) Mean Family Income 1969 (MFI) Adjusted Gross Income Per 1976 State Income Tax Return (AGI) AEPVPP weighted by MFI AEPVPP weighted by AGI These five were then repeated using the five wealth measures per weighted pupils and the cumulative proportion of weighted pupils on the horizontal axis. Using the various wealth measures provided infor mation on the sensitivity of this analysis technique. In all cases the expenditure measure was Total Current Operating Expenditures. We now turn to an analysis of ex ante fiscal neutrality using Lorenze curves and Gini coefficients. However, before this can be done some manipulations based on the work of Yang and Chaudhari (Hickrod et a l., n.d., p. 47) must be done with the basic model. 101 The general case is pictured below in Figure 3 with curve OTC being the Lorenze curve using proportions of Total Current Operating Expenditures on the vertical axis. 100% Q. +J I— 100% % of Students Ranked by Wealth Figure 3. Lorenze curve showing disequalization due to variation in Total Current Operating Expenditures per pupil The disequalizing effect of the way Total Current Operating Expenditures are distributed among the districts is indicated by the degree to which Lorenze curve OTC sags below the 45° line. The ac companying Gini coefficient for this total expenditures effect is expressed as Gtg . The local expenditures effect can be shown graphically by putting local expenditures, indicated by assessed valuation (AV) times local operating tax rate (TR), on the vertical axis as shown in Figure 4. 0 % Students Ranked by Wealth Figure 4. Lorenze curve showing disequality due to local expenditures, that is, combined effect of AV and TR. Curve OBC indicates the extent of disequalization caused by local expenditures. The accompanying Gini is G ^. The state aid effect could be found by looking at the difference between the total expenditures effect and the local expenditures effect. If we assume that the Gini coefficients are additive, we can find the Gini for the state aid effect by subtracting the local expenditure Gini from the total expenditure Gini. That is Going one step further, the local expenditure effect is at tributable to variations in assessed valuation and to the variations in tax rates. The disequalization due to assessed value is depicted by curve OAC in Figure 5. 103 100% •x 0 % Students Ranked by Wealth Figure 5. Lorenze curve showing disequality due to variation in AV only. Note that %AV has been placed on the vertical axis. The ac companying Gini is G_,. Once again, assuming additivity of the Gini a V coefficients, the Gini coefficient for the tax rate effect (Gtr) can be found as follows: Gtr = Gle " Gav This brings us to the assessment of ex ante fiscal neutrality. As you will recall, under this ex ante notion, fiscal neutrality would be achieved if variation in expenditures due to fiscal capacity were eliminated and variation due to tax rates were permitted. In order to assess ex ante fiscal neutrality we need a measure that looks at variations due to state aid and fiscal capacity only, ignoring the effects of tax rates. This will be termed the capacity effect and its corresponding Gini coefficient (Gc) will be given by 104 in which the effect of tax rates has been subtracted out. In order to calculate the Gc's, Lorenze curves and accompa nying local expenditure Ginis and tax rate Ginis were derived. For the local expenditure Ginis, the Local Revenue variable was the expenditure measure. For the assessed value Ginis, Adjusted Equalized Property Value was the "expenditure measure." Each of these Lorenze curves and Gini coefficients were caluclated.using all of the various wealth measures used in the ex post fiscal neutrality analysis. This, of course, included both weighted pupil and unweighted pupil analysis. This multiple analysis approach served as a sensitivity check on the analysis technique being employed. The Gc's will be compared across the six years with a decrease in their values indicating movement toward ex ante fiscal neutral ity.- We now turn our attention to the third hypothesis to be tested in this study. Incentive for Effort Analysis Hypothesis III. School d is tric ts with mi 11 age between 20 and 30 m ills, and with yields per pupil per mill less than $42 will be found to have been stimulated to spend more local dollars under the Equal Yield For mula than before the Equal Yield Formula. In order to test this hypothesis several approaches were con sidered. Analyzing passage rates of operating levies and analyzing increases in mi 11 age among local districts to see if the incentive for 105 effort effect had caused these increases were investigated. However, due to the myriad of factors that can effect voter behavior and cannot be controlled for in this type of analysis, it was concluded that no valid inference stating that the incentive for effort effect caused or did not cause certain behavior could be made (for support of this idea see Alexander & Bass, 1974, p. 16; Alexander & Jordan, 1976, p. 347; Barro, 1974, p. 70; Kay, 1973, p. 322). Also, the decisions made by school d is tric ts regarding school operating levies are in reality an aggregation of individual voter decisions. It is unlikely that the average voter would be aware of the matching grant properties of the Equal Yield Formula and make a voting decision based on income effects and price effects. Therefore, the analysis for this hypothesis will be couched in terms of the re sponses of the decision-making unit which is likely to be aware of the effects of the state aid formula. This decision-making unit consists of the superintendent and board of education of the school d is tric t. It is at this level that the decision to put an operating levy on the ballot or the size of an operating levy could be a response to the matching grant effects of the Equal Yield Formula. In order to determine what effects the Equal Yield Formula has had on the proposal of operating levies, a questionnaire was developed and sent to superintendents of all districts that could be subject to the incentive effect of the Equal Yield Formula. (See Appendix C for a copy of the questionnaire and the memorandum sent with it .) This survey group consisted of all school d istric ts that had equalized current operating millage between 20 and 30 mills and local per pupil per mill yields of less than $42 for any of the school years 106 1975-76 to 1978-79. There were 454 d istric ts in this group with a total of 376 superintendents returning the completed questionnaire. This is an 83 percent response rate. Questionnaire recipients were asked to respond to five short questions for each of the four years. The f ir s t question simply asked if a new operating levy had been put on the ballot. The second question asked if a levy had been put on the ballot, was it at least partially in response to the incentive effect of the Equal Yield Formula. This question gets directly at the idea in Hypothesis III and its results were used to te st the hypothesis. Question 3 asked if there had been a levy on the ballot, was the amount of millage less than what the district would have had to propose had i t not been for the additional state aid under the Equal Yield Formula. Taking this idea a step further, question 4 asked if the district was able to put off proposing a levy because of additional state aid received. Both of these questions are actually getting at the idea that if the income and price elasticities of demand for edu cation are very low, additional state aid, whatever its form, will be substituted for local effort. The fifth question was a "catch all" question asking if the Equal Yield Formula had any effect of any kind on the proposing of operating levies. Finally, space was provided for comments. These comments were reviewed and summarized i f relevant and appropriate. The number and percent answering "yes" and "no" to each question was calculated for the total group of respondents. After looking at frequency distributions on millage and on local yield per pupil per m ill, d istric ts were grouped into categories based on these two variables. This was done to see if results for each question differ for the various categories. It was suspected that if the incentive effect was operating, it would be strongest for those districts with low tax yields and low millage, since it was these districts which would gain the most state aid from additional millage. CHAPTER IV RESULTS Chapter III presented a detailed description of the data analysis techniques used in this study. This chapter presents and interprets the results of each of these analyses. Variation Measures Hypothesis I stated the expectation that the variation in total operating expenditures per pupil among school d istric ts before the implementation of the Equal Yield Formula would be reduced with the implementation of the Equal Yield Formula. In order to assess i f there was a change in expenditure variation, five different measures of variation were calculated for the two years prior to and the f ir s t four years under the Equal Yield Formula. These measures are presented in Table 8. The results for each of these measures will be discussed in turn. 108 TABLE 8 Variation Measures for Total Operating Expenditures Per Pupil for All Districts, 1974 to 1979 Pre-Equal Yield Years Unc er Equal Yield Years Variation Measures 1974 1975 1976 1977 197.8 1979 Restricted Range 569 580 662 751 839 925 Federal Range Ratio .8314 .7468 .8260 .8522 .8783 .8504 McLoone Index .9309 .9277 .9076 .9122 .9069 .8902 Coefficient of Variation 5.23 4.67 4.42 4.38 4.92 4.08 Standard Deviation of Logarithms .1879 .1830 .1813 .1814 .1902 .1762 110 Restricted Range. While total operating expenditures increased from the pre-EYF years to the EYF years so did the restricted range. The restricted range (95th percentle - 5th percentile) increased steadily from $569 in 1974 to $925 in 1979. Thus, if the restricted range is used as a measure of equity of the districution, these figures indicate that the Ohio school finance system has moved away from equality rather than toward equality. Federal Range Ratio. As i t can be seen in Table 8, the value of the federal range ratio increased during the period under the EYF. This result is similar to the result obtained using the restricted range, but this is not surprising since the federal range ratio is the restricted range divided by the fifth percentile. As with the restricted range, the federal range ratio indicates a movement away from equity. McLoone Index. This index, sometimes called the permissible variance index, is based on the dollars needed to raise all districts spending below the median level of total operating expenditures per pupil up to the median. The closer the McLoone Index is to one, the more equitable the distribution will be considered to be. From Table 8 one can see that with the exception of 1977 the McLoone Index has moved steadily away from one since 1974. Therefore, this measure also indicates movement away from equity. I ll Coefficient of Variation. This measure of variation is actually the standard deviation of the distribution divided by the mean of the distribution. A decrease in this would be considered an indication of a decrease in variation or a movement toward equity. The results in Table 8 indicate that the coefficient declined very slightly from 1975 through 1977, went back up to 4.92 in 1978 and then dropped again to 4.08 in 1979. These fluctuations are so slight that i t will be inferred that in general there was no change in variation or equity as measured by the coefficient of variation although slight movement toward less variation was indicated for 1979. Standard Deviation of Logarithms. This measure of variation is very similar to the coefficient of variation except that this measure weights changes at the lower end of the distribution more heavily than at the upper end. A decline in this measure will be inferred to indicate movement toward less variation. The results for this measure followed the same pattern as the coefficient of variation. The standard deviation of logarithms declined slightly through 1977, increased to its largest value (.1902) and then dropped to its lowest value (.1762) in 1979. This would appear to, indicate.a slight movement toward less variation or more equity in 1979 when the EYF was fully funded for the first time. Therefore, with three measures indicating movement toward' increased variation, and with two measures showing l i t t l e change with slight movement toward less variation in 1979, i t can be concluded that the expectation for less variation in total operating expendi tures per pupil under the EYF has not been realized. Effect of Basic Aid on Variation Measures. In order to determine if the distribution of basic aid from the state improved the equity of the distribution of total operating expenditures per pupil more under the EYF than before the EYF, Table 9 was constructed. The various measures were calculated on the distribution of total operating expenditures per pupil minus basic aid per pupil. These variation measures were then compared to those calculated using total operating expenditures per pupil to determine the percentage of change in the measures that was caused by the inclusion of basic aid in the distribution. These are the percentages displayed in Table 9. TABLE 9 Percentage Change in Variation Measures Caused by Distribution of Basic State Aid Being Added to Other State Aid Plus Local Revenue for all D istricts, 1974 to 1979 Pre-Equal Yield Years Under Equal Yield Years Variation Measures 1974 1975 1976 1977 1978 1979 1 ro Restricted Range -26.6 -28.0 -20.8 -20.7 -20.0 CTJ Federal Range Ratio -68.2 -70.8 -64.7 -64.0 -62.6 -63.7 McLoone Index +19.8 +18.6 +14.0 +15.8 +14.2 +12.5 Coefficient of Variation -74.0 -75.9 -72.4 -71.6 -70.4 -74.3 Standard Deviation of Logarithms -50.2 -51.1 -47.7 -47.3 -46.0 -49.0 114 The lower percentages for the first three Equal Yield years indicate that basic aid was less effective in improving the distribu tion of revenue in that period than it was in the pre-Equal Yield years. Only the coefficient of variation and the standard deviation of logarithms for 1979 showed improvement comparable to the pre-Equal Yield years. Variation Measures for District Types. For this part of the analysis school d is tric ts were divided into four groups: large central city d is tric ts , suburban d is tric ts surrounding the large c itie s, d istric ts associated with independent small cities and rural districts. Then the five variation measures were calculated within each group. It was f e lt that using these more homogeneous groupings would make the vari ation measures faire r measures of equity since the distributions would be among districts in more similar situations than the distribution for all districts in the state. The results shown in Table 10 indicate no clear pattern of movement toward less variation under the EYF for any of the four d is tric t groups. For each group the results are mixed with most measures indicating either no decrease or more variation and a few indicating slight decreases in variation for selected years. Therefore, even when d istric ts are analyzed in more homogeneous groups, it must be concluded that variation has not been reduced. This supports the conclusion drawn from the all-district analysis. TABLE 10 Measures of Variation Among School D istrict Types of Total Operating Expenditures Per Pupil, 1974 to 1979 • Variation Pre-Equal Yield Years Under Equal Yield Years Measures "1974" ' 1975 1976 ' ' 1977 1978 1979 Large Central City D istricts Restricted Range 363 405 294 358 370 422 Fed. Range Ratio .4505 .4472 .2793 .3243 .2961 .2914 McLoone Index .9769 .9643 .9281 .9069 .9549 .9535 Coef. of Variation 1.35 3.88 5.98 11.57 21.94 45.63 S.D. of Logarithms .1073 .1735 .2127 .2839 .3768 .5095 Suburban D istricts Restricted Range 338 385 385 419 465 476 Fed. Range Ratio .4895 .4866 .4609 .4509 .4677 .4161 McLoone Index .9302 .9200 .9143 .9260 .9209 .9260 Coef. of Variation 1.31 3.36 6.68 13.96 24.18 49.39 S.D. of Logarithms .1121 .1612 . .2211 .3080 .3916 .5245 Districts Associated with Small Independent Cities Restricted Range , 805 889 983 1099 1388 1307 Fed. Range Ratio 1.0996 1.0713 1.1381 1.1493 1.3265 1.0595 McLoone Index .8690 .8678 .8647 .8687 .8472 .8994 Coef. of Variation 5.03 5.71 7.88 13.88 28.23 49.03 S.D. of Logarithms .2154 .1960 .2213 .2845 .3852 .4955 Rural D istricts Restricted Range 297 310 405 453 490 537 Fed. Range Ratio .4381 .4031 .5147 .5277 .5233 .5023 McLoone Index .9399 .9340 .9317 .9291 .9236 .9074 Coef. of Variation 3.56 4.66 6.71 12.88 20.68 45.96 S.D. of Logarithms .1286 .1581 .2017 .2794 .3564 .5035 Fiscal Neutrality Analysis Results Hypothesis II was based on the expectation that fiscal neutrality of the Ohio school finance system would be increased under the EYF. Both ex post fiscal neutrality (actual level of education support must not be correlated with wealth) and ex ante fiscal neutrality (the ability to fund schools should not depend on wealth) were explored. Simple regressions, multiple regressions, Gini coefficients and Lorenze curves were used to assess the degree of movement toward fiscal neutrality. A discussion of the results of each of these analyses will follow. Simple Regressions. This analysis consisted of running a series of simple regressions of both total operating expenditures per pupil (TOPEPP) and total operating expenditures per weighted pupil (TOPEWP) on various measures of d is tric t wealth (see Table 6). These 2 regressions were run for all six years and the R was calculated for 2 each regression. A decrease in R was assumed to indicate that the relationship between the expenditure measure and the wealth measure had decreased and wealth neturality was being approached. Table 11 2 indicates the R values for three representative regressions. Similar results were found for the other regressions using weighted expenditure and weighted wealth measures. These results are shown in Tables 19 and 20 of Appendix C. 117 TABLE 11 2 R Values for Representative Simple Regressions, 1974 to 1979 Pre-Equal Yield Years Under Equal Yield Years Regression3 1974 1975 1976 1977 1978 1979 TOPEPP on AEPVPP .6383 .6312 .5783 .5461 .6330 .5947 TOPEPP on MFI .1738 .1936 .2266 .2236 .2143 .2046 TOPEPP on AGI .1565 .1581 .1897 .1774 .1423 .1286 TOPEPP = Total Operating Expenditures Per Pupil, AEPVPP = Adjusted Equalized Property Value Per Pupil, MFI = Mean Family Income 1969 and AGI = Adjusted Gross Income 1976 State Tax Return. The results indicate a drop in the dependency of TOPEPP on AEPVPP in the f ir s t two years under the EYF with the dependency going back up to the pre-EYF level in 1978. The final 1979 value for R^ was about 94 percent of the 1975 value. Thus, the gains in equity in 1976 and 1977 were largely lost by 1978 and 1979. The results involving MFI as the wealth measure show an increase in the strength of the relationship between TOPEPP and MFI for the equal yield years. This represents slight movement away from fiscal neutrality. Although AGI is fairly highly correlated with MFI (.85) the regression of TOPEPP on AGI yielded a slightly different pattern 2 from the MFI regressions. The R increased in 1976 and 1977 as was 2 the case for the MFI regressions. The R then dropped to below the pre-EYF level with the 1979 value being only 81 percent of the 1975 value. Thus, with AGI as the wealth measure, the simple regressions 118 indicated movement away from fiscal neutrality in the first two years under the EYF and some movement toward fiscal neutrality in second two years under the EYF. In absolute terms, the percent of variance in the expendi tures measures explained by AEPVPP and other property wealth measures p (R ) was approximately three times as great as the percent of variance explained by the MFI and AGI measures. Therefore, the results obtained regressing TOPEPP on AEPVPP are probably more valid than the regressions using MFI or AGI. The simple regression analyses indicate little or no progress toward ex post fiscal neutrality. Since there are factors other than wealth which could affect expenditures on education, multiple regressions were run to control for these other factors in assessing ex post fiscal neutrality. The results of the multiple regression analyses follow. Multiple Regression Results. For this part of the analysis of movement toward ex post fiscal neutrality a series of stepwise multiple regressions were run in which the dependent variable was total operating expenditures per pupil or per weighted pupil, and the independent variables included some wealth measure and various combi nations of the variables listed in Table 7 of the previous chapter. It was found that the best fitting equations were those with the independent variables being any one of the property wealth (weighted or unweighted) measures plus mean family income 1969 (MFI), percent 119 of population less than 18 years old (PRCTLT18), and percent of o district enrollment that is minority (PCTMIN). The R for these regressions ranged from .67 to .76. Once i t was determined that MFI, PRCTLT18, and PCTMIN would be included in the regressions with the property wealth measures, a set of regressions was run for each year with either total operating expenditures per pupil (TOPEPP) or total operating expenditures per weighted pupil (TOPEWP) as the dependent variable. The focus of this analysis was on the regression coefficient of the property wealth measure. It was assumed that if the regression coefficient of the property wealth measure decreased, the relationship between the expenditure measure and the property wealth measure would be reduced and thus movement toward ex post fiscal neutrality would be inferred. Table 12 lis ts six regressions that were run for each year. TABLE 12 Regression Coefficients of Property Wealth Measures of Multiple Regression with TOPEPP or TOPEWP as the Dependent Variable, 1974 to 1979 Regression 1974 1975 1976 1977 1978 1979 TOPEWP = AEPVWP + others3 .010170 .010425 .008812 .008913 .010577 .009960 TOPEWP = AEPWPM + others .010745 .011066 .009128 .009203 .010988 .010325 TOPEWP = AEPWPA + others .011034 .011358 .009263 .009331 .011197 .010513 TOPEPP = AEPVPP + others .009977 .010518 .008912 .008850 .010536 .009864 TOPEPP - AEPPM + others .010504 .011140 .009236 .009127 ,010939 .010219 TOPEPP = AEPPA + others .010773 .011426 .009377 .009251 .011147 .010406 aIn each regression others = MFI + PRCTLT18 + PCTMIN 121 The regression coefficients in Table 12 reveal similar patterns for all six regressions. During 1976, the f ir s t year under the EYF, the regression coefficients decreased between fifteen and eighteen percent of the 1975 values. This lower level of the coef ficients remained in 1977 but went back up to pre-EYF levels in 1978. Finally, in 1979 the coefficients were about six to nine percent of the 1975 values lower than the 1975 values. Therefore, regardless of which regression is observed, the gains in movement toward ex post fiscal neutrality in the f ir s t two years under the EYF appear to have been eliminated in the third year under the EYF and only slightly regained in the fourth year. It is obvious from these results that there is no clear trend toward ex post fiscal neutrality under the EYF. Gini Coefficents and Lorenze Curves. Lorenze curves and their associated Gini coefficients were also employed to assess movement toward ex post fiscal neutrality. Lorenze curves were plotted with cumulative percentages of Total Operating Expenditures (TOPEX) on the vertical axis and cumulative percentages of students of districts ranked by various wealth measures on the horizontal axis. This pro cedure was then repeated using cumulative percentages of weighted students on the horizontal axis. Corresponding Gini coefficients were computed for each Lorenze curve for each year. It was assumed that if the Gini coef ficients decreased over the years, movement toward ex post fiscal neutrality would be inferred. The Gini coefficients for all Lorenze curves for all six years are given in Table 13. TABLE 13 Gini Coefficients for Lorenze Curves Using TOPEX on Vertical Axis and Weighted and Unweighted Students Ranked by Various Wealth Measures on Vertical Axis, 1974 to 1979 Wealth Measure 1974 1975 1976 1977 1978 1979 Unweighted Students AGI 7.961 7.767 12.444 14.279 15.489 15.683 MFI 27.205 27.125 31.463 33.087 37.257 36.341 AEPVPP 79.783 74.654 72.474 69.923 70.145 53.630 AEPPPA 80.539 74.894 73.079 70.554 70.555 54.139 AEPPM 80.715 75.036 72.654 70.089 69.672 53.397 Weighted Students AGI 27.090 27.299 32.634 34.073 35.055 36.190 MFI 42.365 42.471 47.351 48.576 52.639 52.329 AEPVWP 69.321 64.871 65.990 64.723 65.324 47.251 AEPWPA 70.928 65.994 67.001 65.404 65.275 47.478 AEPWPM 70.878 65.626 66.366 64.548 65.136 46.673 123 The results in Table 13 indicate that if we look at the change in the distribution of expenditures to either students or weighted students in d istric ts ranked by the two income measures of wealth (AGI or MFI) there appears to be movement away from ex post fiscal neutrality. The Gini coefficients increased in the EYF years for all four sets of Lorenze curves. On the other hand, the results for the remaining Gini coefficients in Table 13 which were derived using various property wealth measures (weighted and unweighted) indicate an opposite trend. For each case the Gini coefficient declined about 5 points from 1974 to 1975 and then either remained about the same or declined slightly over the first three years under the EYF and then dropped in 1979 to a level eighteen to twenty points below the 1975 level. This indi cates a clear movement toward ex post fiscal neutrality in 1979. Representative Lorenze curves for 1975 and 1979 are shown- in Figure 6. As you will note the 1979 Lorenze curve is closer to the 45° line than the 1975 curve indicating movement toward a more equitable distribution. The different results for income measures and property wealth measures can be explained by the fact that the distribution of basic state aid is based in part on district property wealth with income not being taken into account. Certainly the full funding of the EYF in 1979 contributed to the movement toward fiscal neutrality in 1979. Cumulative Percentage of Total Operating Expenditures 100 xedtrs o tdns n sti s rne b Ajse Eulzd rpry au Pr Pupil(AEPVPP) Per Value Property Equalized Adjusted by ranked ts tric is d in students to Expenditures Figure 6. lorenze curve showing disequalization due to d istrib u tio n of Total Operating Operating Total of n tio u istrib d to due disequalization showing curve lorenze 6. Figure uuaie ecnae f tdns n srcs ae b AEPVPP by Rated istricts D in Students of Percentage Cumulative 1979 1975 124 125 Ex Ante Fiscal Neutrality Analysis. This analysis was an attempt to look at the relationship between the distribution of capacity for ex penditures and the distribution of students and weighted students in districts ranked by various income or property wealth measures. Gini coefficients indicating the disequalizing effects of the distribution of adjusted equalized property value were calculated (G 1. These av Gini coefficients were then subtracted from Gini coefficients indi cating the disequalizing effect of the distribution of local revenue (G-|e) to arrive at a tax rate Gini coefficient (Gtr). The ability or capacity Gini (Gc) was then calculated by subtracting the tax rate Gini coefficients from the Gini coefficients calculated to indicate the disequalizing effects of the distribution of total operating expenditures (Gte). In other words the disequalizing effect due to tax rate differences among school d istric ts (Gt r ) was taken out of the measure of disequality of the distribution of -total operating expenditures (Gt0) so that only the effects of those factors that reflect the capacity of the d is tric t to spend on education remained in the disequality measure. The resulting Gini coefficients (Gc) represent the disequalizing effects of capacity to spend on education which include local property valuation and state and federal aid, but exclude the disequalizing effects of varying tax rates or willingness to spend on education. The capacity Gini coefficients are presented in Table 14 for each wealth measure for all six years. TABLE 14 Capacity Gini Coefficients Calculated by Subtracting Tax Rate Gini from Total Expenditure Gini for Various Wealth Measures, 1974 to 1979 Wealth Measure 1974 1975 1976 1977 1978 1979 Unweighted Students AGI 3.280 .294 1.137 5.059 7.722 8.662 MFI -8.542 -13.326 -6.845 -1.910 .774 6.218 AEPVPP 64.382 61.120 63.722 61.731 59.423 48.606 AEPPPA 61.766 59.020 62.249 61.014 • 59.182 48.703 AEPPM 63.292 60.512 63.514 61.898 59.806 48.951 Weighted Students AGI 22.854 20.278 21.786 25.195 27.777 29.611 MFI 6.721 1.980 9.033 13.480 16.107 22.231 AEPVWP 60.967 57.608 59.589 62.298 61.018 47.963 AEPWPA 59.782 56.600 58.834 61.891 61.280 47.996 AEPWPM 60.070 56.977 59.461 62.408 61.086 48.097 127 The results for the capacity Ginis shown in Table 14 display the same pattern as the Gini coefficients for total expenditures in Table 13. The Gini coefficients for the income wealth measures in creased through the EYF years indicating movement away from ex ante fiscal neutrality. Gini coefficients for the property wealth measures showed only slight fluctuations from 1975 to 1978 and then dropped about eleven to thirteen points in 1979. Thus, there appears to be movement toward ex ante fiscal neutrality in 1979 if the property wealth capacity Gini analysis described above is used. Once again, this is probably due to the fact that state basic aid is distributed in an inverse relation to local property wealth capacity and the full funding level was able to lessen the positive relationship between total expenditures per pupil and property wealth. Incentive for Effort Analysis This analysis sought to determine what effects, if any, the operation of the EYF had on the proposal of new operating levies of local school d is tric ts . It was hypothesized th at school d istric ts would attempt to increase local spending due to the incentive effects of the EYF. On the other.hand, this.study also sought to determine if perhaps the EYF might allow districts to make less local effort if the EYF allocated to them increased funds from the state. The five-item questionnaire included in Appendic C was sent to 454 school district superintendents and an 83 percent response was obtained. After looking at frequency distributions the responding districts were broken down into categories based on various combinations 128 of equalized millage and local yield per pupil per mill. This was done so that results could be analyzed for districts at various levels of property wealth and tax effort in addition to the analysis for all d istric ts. These results are displayed in Tables 15, 16, 17 and 18, one table for each of the f i r s t four years of the EYF. For each cell in the tables the figures represent the percentage of respondents to each question in that cell that answered "yes" to the five respective questions. The corresponding number answering "yes" is in parenthesis next to each percentage. Results for all five questions are included in each cell with the percentage for question one being the figure at the top of the column in the cell and with percentages for subsequent questions listed below. Thus, for example, in Table 15 in cell A, 31.5 percent of the d istric ts in cell A that answered question one answered "yes" to question one, 51.9 percent that answered question two answered "yes" to question two, 36.5 percent that answered question three answered "yes" to question three, 39.7 percent that answered question four answered "yes" to question four, and 33.7 percent that answered question five answered "yes" to question five. Although there would be no incentive effect for d istric ts with equalized millage above 30 mills, these districts were included in the tables because some may have had millage below the 30 mill level during the year and then raised their millage above the 30 mill level- during the same year. There were only a few d istric ts that had above 30 mills in any of the four years so these percentages are based on very small numbers. Therefore, the results for d istric ts above 30 mills were included for completeness and will be largely ignored in discussing individual cell results. 129 Question one asked for each of the years if the district had had a new operating levy on the ballot. The results for question one in cell L of each table indicate that the percent answering question one that proposed new operating levies was 31.6 percent in 1976, then 45.6 percent in 1977, then 53.7 percent in 1978, and only 29.5 percent in 1979. The extent to which the proposal or non-proposal of these levies was affected by the EYF was explored in the remaining questions. Question two asked if the new operating levies referred to in question one were proposed at least partially in response to the incentives of the EYF. For 1976, the results for question two in cell L indicate that 45.8 percent of all responding districts proposed levies at least partially in response to the incentive effect of the EYF. The results for this question were only slightly higher for lower millage districts (cell I) than for higher millage districts (cell J). Also, lower property wealth districts (cell D) were 7.1 percentage points higher than higher property wealth districts (cell H) as would be expected. However, 75 percent of the districts in cell F responded "yes" to question two. This cell could reasonably be expected to indicate the lowest percentage rather than the highest. However, it should be noted that this 75 percent represents only three districts. Question three asked if the levies referred to in question one were less than they would have been had i t not been for additional aid of the EYF. Cell L indicates that 33.3 percent of those answering question three said that their levy was not as high as it would have been had it not been for additional state aid under the EYF. Lower property wealth districts in cell D responded 40.9 percent "yes" while 130 the higher property wealth d istric ts responded only 12.5 percent "yes." Also, d is tric ts in cell J responded substantially higher than cell I for question three. Question four sought to determine if EYF state aid had been substituted for local effort by asking if no levy had been proposed was this due to additional state aid received under the EYF. For question four, 36.2 percent of all respondents said that they did not put a levy on the ballot because of additional state aid from the EYF. As with question three percentages were higher for the higher millage and lower wealth districts. Question five asked if the implementation of the EYF had any effect on whether the district did or did not put a levy on the bal lot. For question five 31 percent of all respondents indicated that the EYF had some effect on whether they did or did not have a levy on the ballot. Lower property wealth districts in cell D responded with a much higher percentage (34.7) than did the higher property wealth districts (15.5). The results for 1977, 1978 and 1979 were very similar to those for 1976. It can be seen in Tables 16, 17 and 18 that the per cent answering "yes" to question two was 55.7 in 1977, then 56.6 in 1978 and 50 in 1979. Thus, overall, about half of the districts responding to question two said that the EYF had given them at least some incentive to put new operating levies on the ballot. As was true in 1976 the lower property wealth districts appear to have been more affected by the incentive effect than were the higher property wealth d is tric ts . The percent answering "yes" for the lower property 131 wealth districts ranged from 68.4 percent in 1977 to 60.5 percent in 1979. The higher millage districts in cell J had slightly higher per centages than the lower millage districts in cell I for all three years. This is opposite the 1976 results. For question three the percent answering "yes" increased from 40.9 in 1976 to 43.7 in 1977, to 45.6 in 1978 and then dropped to 37.5 in 1979. The overall percent answering "yes" to question four was about 43 percent for 1977 through 1979 as compared to 36.2 percent in 1976. Thus, the overall results for questions two, three and four indi cated that while the incentive effect affected a higher percentage of the districts proposing levies in the last three years of the EYF than it did in 1976, a higher percentage also substituted some state funds for local effort. The overall percent answering "yes" to question five was slightly higher than the 1976 value of 31 percent, ranging from 38.1 in 1977 to 33.9 in 1979. Over all four years percents answering "yes" to question five were consistently higher for the lower property wealth districts than for the higher property wealth districts. In summary, the results for all four years yielded these general conclusions. 1. About .half of the districts that put new operating levies on the ballot and answered question two indicated that they put these levies on the ballot at least partially in response to the incentive effect of the EYF. This percentage was higher for lower property wealth districts than for higher property wealth districts. 2. About 40 to 45 percent of the districts responding to questions three and four indicated some degree of substitution of state aid for local effort. These percentages were higher for the lower property wealth districts than for the higher property wealth d is tric ts . 3. About one-third of the districts answering question five indicated that the implementation of the EYF had some effect on whether they did or did not put a levy on the ballot. This percentage was about 45 percent for lower property wealth districts. TABLE 15 Percent and Number Answering Questions That Answered "Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1976 LYPPPM EQUMIL 20-26.99 27-30 30.01-33.99 All 31.5 (52) 23.6 (13) 15.0 ( 3) 28.3 (68) 51.9 (27) 35.1 ( 5) 33.3 ( 1) 47.8 (33) < 25 36.5 (19) 58.3 ( 7) 50.0 ( 1) 40.9 (27) 39.7 (29) 54.2 (13) 33.3 ( 5) 42.0 (47) 33.7 (55) 38.5 (20) 33.3 ( 7) 34.7 (82) A B C D 46.8 (22) 33.3 ( 4) 50.0 ( 1) 44.3 (27) 36.4 ( 8) 75.0 ( 3) 0 ( 0) 40.7 (11) 25 - 41.99 10.0 ( 2) 33.3 ( 1) 0 ( 0) 12.5 ( 3) 11.8 ( 2) 12.5 ( 1) 0 ( 0) 11.5 ( 3) 14.0 ( 6) 23.1 ( 3) 0 ( 0) 15.5 ( 9) E F G H 34.9 (74) 25.4 (17) 18.2 ( 4) 31.6 (95) 47.3 (35) 44.4 ( 8) 25.0 ( 1) 45.8 (44) All 29.2 (21) 53.3 ( 8) 33.3 ( 1) 33.3 (30) 34.4 (31) 43.8 (14) 31.3 ( 5) 36.2 (50) 29.6 (61) 35.4 (23) 30.4 ( 7) 31.0 (91) I J K L aFigures in parenthesis indicate number of districts answering "yes" to respective question. TABLE 16 Percent and Number Answering Questions That Answered "Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1977a LYPPPM EQUMIL 20-26.99 27-30 30.01-33.99 All 43.5 ( 73) 18.5 ( 5) 100.0 ( 2) 40.6 ( 80) 68.1 ( 49) 80.0 ( 4) 50.0 ( 1) 68.4 ( 54 <25 55.6 ( 40) 20.0 ( 1 ) 100.0 ( 2) 54.4 ( 43) 50.0 ( 33) 52.9 ( 9) 0 ( 0) 50.6 ( 42) 47.2 ( 75) 56.0 (14) 0 ( o) 47.3 ( 89) A B C D 54.7 ( 47) 52.0 (13) 100 ( 1) 54.5 ( 61) 37.5 ( 18) 50.0 ( 6) 0 ( 0) 39.3 ( 24) 25 - 41.99 25.6 ( 11) 41.7 ( 5) 0 ( 0) 28.6 ( 16) 24.1 ( 7) 45.5 ( 5) 0 ( o) 30.0 ( 12) 24.1 ( 21) 20.0 ( 5) 0 ( 0) 22.8 ( 26) E F G H 47.2 (120) 34.6 (18) 100 ( 3) 45.6 (141) 55.8 ( 67) 58.8 (10) 33.3 ( 1) 55.7 ( 78) All 44.3 ( 51) 35.3 ( 6) 67.0 ( 2) 43.7 ( 59) 42.1 ( 40) 50.0 (14) Q ( o) 43.9 ( 54) 39.0 ( 96) 38.0 (19) 0 ( 0) 38.1 (115) I J K L aFigures in parenthesis indicate number of districts answering "yes" to respective question. TABLE 17 Percent and Number Answering Questions That Answered "Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1978a LYPPPM EQUMIL 20-26.99 27-30 30.01-33.99 All 58.7 ( 91) 27.8 (10) 33.3 (2) 52.3 (103) 65.5 ( 57) 66.7 ( 6) 0 (0) 64.3 ( 63) <25 56.8 ( 50) 55.6 ( 5) 100.0 (1) 57.1 ( 56) 43.5 ( 20) 72.7 (16) 33.3 (1) 52.1 ( 37) 43.0 ( 65) 57.6 (19) 50.0 (3) 45.8 ( 87) A B C D 57.8 ( 63) 58.3 (14) 25.0 (1) 55.7 ( 78) 44.3 ( 27) 61.5 ( 8) 0 (0) 46.7 ( 35) 25 - 41.99 28.8 ( 17) 38.5 ( 5) 0 (0) 30.1 ( 22) 25.8 ( 8) 16.7 ( 1) 50.0 (1) 25.6 ( 10) 24.8 ( 28) 18.2 ( 4) 20.0 (1) 23.6 ( 33) E F G H 57.7 (154) 40.0 (24) 30.0 (3) 53.7 (181) 56.8 ( 84) 63.6 (14) 0 (0) 56.6 ( 98) All 45.6 ( 67) 45.5 (10) 50.0 (1) 45.6 ( 78) 36.4 ( 28) 60.1 (17) 40.0 (2) 42.7 ( 47) 35.2 ( 93) 41.8 (23) 36.4 (4) 36.4 (120) I J K L aFigures in parenthesis indicate number of districts answering "yes" to respective question. TABLE 18 Percent and Number Answering Questions That Answered “Yes" for Questions 1-5 for Various Combinations of Equalized Millage (EQUMIL) and Local Yield Per Pupil Per Mill (LYPPPM), 1979a , LYPPPM EQUMIL 20-26.99 27-30 30-01-33.99 All 29.5 (31) 10.6 ( 5) 0 (0) 22.4 ( 36) 59.4 (19) 66.7 ( 4) 0 (o) 60.5 ( 23) <25 41.4 (12) 60.0 ( 3) 0 (0) 44.1 ( 15) 51.1 (24) 64.5 (20) 50.0 (3) 56.0 ( 47) 39.0 (41) 51.1 (23) 81.8 (9) 45.3 ( 73) A B C D 44.5 (49) 16.1 ( 5) 23.5 (4) 36.7 ( 58) 45.8 (22) 40.0 ( 2) 0 (0) 42.9 ( 24) 25 - 41.99 32.6 (15) 40.0 ( 2) 33.3 (1) 33.3 ( 18) 20.5 ( 9) 42.1 ( 8) 25.0 (2) 26.8 ( 19) 23.3 (24) 24.1 ( 7) 10.0 (2) 21.7 ( 33) E F G H 37.2 (80) 12.8 (10) 15.4 (4) 29.5 ( 94) 51.3 (41) 54.5 ( 6) • 0 (0) 50.0 ( 47) All 36.0 (27) 50.0 ( 5) 33.3 (1) .37.5 ( 33) 36.3 (33} 56.0 (28) 35.7 (5) 42.6 ( 66) 31.3 (65) 40.5 (30) 35.5 (11) 33.9 (106) I J K L aFigures in parenthesis indicate number of districts answering "yes" to respective question. 137 Open-Ended Comments from Questionnaire Of the 376 questionnaires returned, 106 contained written comments in the space provided for open-ended comments. Some of the comments were brief explanations of why the district put a levy on the ballot. Among these reasons were making up for inflation, to keep from closing and to make up for a decrease in equalized millage caused by reappraisal. This last reason, combined with the effects of H.B. 920 was commented on more than twenty-five times. Since this last reason seemed to be the major concern of the respondents who wrote comments, a brief digression on the effects of reappraisal and its relationship to the incentive effect will be in cluded here. As explained in Chapter I, if a district has a yield per pupil per mill less than $42 and has equalized current operating millage between 20 and 30 mills, the state will provide the difference between the $42 yield and the district's yield per pupil per mill for each mill above 20 mills but less than 30 mills. If the county in which the district is located undergoes a sexennial reappraisal or triennial update its actual assessed property value will rise. However, H.B. 920 mandates that only increases due to new construction growth in tangible personal property or due to the applications of inside millage to the increased valuation are allowed. Revenues from application of outside millage to the increased valuations are given back to the taxpayers in the form of a property tax credit. Thus, districts cannot take advantage of the inflation of the value of their tax bases. This type of property tax growth restriction is unique to Ohio. 138 Furthermore, the state basic aid calculation is based on equalized mills which are found by dividing local revenues by assessed valuation. After reappraisal the local revenue figure used is net of the tax credits but the assessed valuation used is the actual new valuation. Thus, the denominator of this fraction increases much more than the numerator, causing equalized mills to drop. Also, the state basic aid calculation is based on the assessed valuation per pupil using the new higher assessed valuation. Thus, the district is credited with more valuation than it actually has available to it. Therefore, the district loses basic state aid due to in creased valuation and decreased equalized mills. This is even further complicated by a degree of uncertainty that districts face in pro jected state aid. With property values being updated every three years and reappraised every six years, a steady growth in state aid may not be realized by districts and this makes it difficult to plan future budgets. This situation was reflected in such comments as "920 killed us;" "the equal yield formula is a joke;" "reappraisal was a disaster;" "920 was a catastrophe for our district;" "the equal yield formula is not equal and does not yield;" and "why vote in millage - the county auditor will just roll you back." - Some commented that this situation provided incentive to increase millage to make up the losses. Others seemed discouraged to increase millage because it didn't seem to help them that much due to the losses in state aid. Seven of the questionnaires contained the comment that the d is tric t was on a guarantee and the workings of the equal yield formula 139 had little effect on them. This lack of incentive was applicable for these districts and probably for most of the guarantee districts. However, a district could go off its guarantee to gain additional funds through increased millage. Of the 616 districts in school year 1978-79 there were 253 on some type of guarantee. Thus, the major effect of these guarantees is to keep the equalizing effects of the EYF from operating to the fullest extent possible. In summary, while there were a few favorable comments con cerning the effects of the EYF, the majority were complaints concerning the effects of how reappraisal is handled in the EYF in conjunction with H.B. 920. CHAPTER V SUMMARY AND CONCLUSIONS Summary The major objective of this study was to assess the effects that implementation of the EYF has had on school finance in Ohio using a selected set of criteria. It was hoped that following its imple mentation the EYF would have three basic effects on the distribution of school operating revenues in Ohio. These expected effects included (1) the proposal of local operating levies by school d is tric ts (especially low tax-low property wealth districts) in response to the matching funds of the EYF, (2) the reduction in variation among districts in operating expenditures per pupil, and (3) a decrease in the dependence of d is tric t total operating expenditures per pupil on local wealth (movement toward fiscal neutrality). The specific research problem of the study was to determine, after four years under the EYF, if these expectations had been met. These expectations were formulated into hypotheses which were tested using a variety of analysis techniques employing actual data for school years 1973-74 through 1977-78, preliminary estimates for school year 1978-79, and some data from the 1970 U.S. Census. To determine whether variation in operating expenditures had been reduced a set of variation measures was calculated for the distribution of total operating expenditures per pupil for each of the six school years from 1973-74 through 1978-79. The variation 140 141 measures for the f ir s t two years (pre-EYF) were compared to the measures calculated for the last four years (under the EYF). The six variation measures included the restricted range, the federal range ratio, the McLoone Index, the coefficient of variation and the standard deviation of logarithms. This variation analysis was also performed for various types of school d is tric ts . Movement toward fiscal neutrality was assessed using two basic forms of analysis—regression analysis and Gini coefficient analysis. Also, two types of fiscal neutrality were assessed—ex post, in which the actual level of educational support must not be correl ated with wealth, and ex ante, which required that the ability of a district to support schools should not depend on local wealth. To assess ex post fiscal neutrality simple and multiple regressions and Gini coefficients were calculated. In the simple regressions total operating expenditures per pupil or per weighted pupil were regressed on various measures of fiscal capacity such as mean family income, property wealth per pupil and property wealth weighted by income. Separate regressions were run for each year and 2 the R values for the pre-EYF years were compared with those for the EYF years. Lower values were assumed to indicate movement toward ex post fiscal neutrality. . A stepwise regression procedure was then used to determine the best fitting multiple regression equation. For each year total operating expenditures per pupil was regressed against various forms of the fiscal capacity measure along with mean family income 1969, percent population less than eighteen years old, and percent of 142 enrollment that was minority. In this analysis the regression coef ficien t of the fiscal capacity measure used was the indicator of the strength of the relationship between fiscal capacity or wealth and expenditures. A decline in the regression coefficient was assumed to indicate movement toward ex post fiscal neutrality. The last assessment of ex post fiscal neutrality employed Lorenze curves and Gini coefficients. For each year Lorenze curves were plotted with total operating expenditures on the vertical axis and the cumulative proportion of students in d istric ts ranked by various measures of fiscal capacity on the horizontal axis. Movement toward ex post fiscal neutrality was to be inferred if the Gini coefficient was reduced during the EYF years. To assess ex ante fiscal neutrality only expenditure ability excluding expenditure preference was at issue. In order to take the effect of expenditure preference out of the Gini coefficient, a Gini coefficient was calculated for the disequalizing effect of the district tax rate. Then, this Gini coefficient was subtracted from the Gini coefficient calculated for the effect of total operating expenditures to arrive at a Gini coefficient representing the capacity or ability effect. These capacity Gini coefficients were calculated for all six years with movement toward ex ante fiscal neutrality being assumed if the coefficient declined. To determine if the EYF had had any effect in providing incentive to school districts to propose operating levies, a question naire was sent to the 454 school d istric ts who, based on their equalized millage rates and property valuation per pupil,could be subject to this 143 incentive effect. This questionnaire (see Appendix C) asked five questions as to whether they had a levy on the ballot, had the EYF provided any incentive for this levy, had the district substituted state aid for local effort and whether the EYF had affected their levy proposing strategy in any way. Open-ended comments were also requested. Eighty-three percent of the questionnaires were returned and analyzed on an overall basis and then responses were cross tabulated on higher and lower property wealth by higher and lower millage rates. The percent answering "yes" to each question was the basis of the analysis. Conclusions The various analyses described above yielded the major findings described below. 1. The restricted range, federal range ratio and the McLoone Index indicated no movement toward less variation in total operating expenditures per pupil among all school districts under the EYF. The standard deviation of logarithms and the coefficient of variation indicated only slight movement toward less variation in 1979. Further more no clear pattern-of movement toward less variation was found using this variation analyses on the distribution for the four district types. 2. The distribution of basic aid was less effective in reducing variation in total operating expenditures per pupil during the Equal Yield Years than i t was in the two years immediately 144 preceding the EYF. Therefore, the evidence suggests that the EYF has not significantly reduced the variation in total operating expendi tures per pupil among Ohio's school d istric ts or within various groupings of school d istric ts. 3. Simple regression analysis indicated the relationship between total operating expenditures per pupil and property wealth per pupil had lessened during the first two years of the EYF but increased to pre-EYF levels in 1978 and was only slightly lower than pre-EYF years in 1979. The results for simple regressions involving income measures as the independent variable were mixed. The strength of the relationship between total operating expenditures per pupil and mean family income increased for the EYF years while the strength of the relationship between total operating expenditures per pupil and adjusted gross income per 1976 state income tax return increased during the first two years of the EYF but then decreased during the last two years. In general the simple regression analyses showed no clear pattern of movement toward ex post fiscal neutrality although one part of the analysis did indicate some movement toward ex post fiscal neutrality in 1979. 4. The regression coefficients of the property wealth measures of all six multiple regressions showed similar patterns over the years. -During the first two years under the EYF the regression coefficients were lower than in the pre-EYF years but went back up to pre-EYF levels in 1978 and then dropped slightly in 1979. Thus, the multiple regression analyses indicated gains in movement toward ex post fiscal neutrality in the first two years under the EYF that were lost in the third year and only slightly regained in 1979. 145 5. Gini coefficients involving income wealth measures in creased during the EYF years indicating movement away from ex post fiscal neutrality. Gini coefficients involving property wealth measures either remained the same or declined slightly during the first three years under the EYF and then declined substantially in 1979. This indicates a clear movement toward ex post fiscal neutrality in 1979. 6. The capacity Gini coefficients calculated to assess ex ante fiscal neutrality yielded results similar to the ex post fiscal neutrality analysis. Coefficients involving income wealth meausres increased during the EYF years indicating movement away from ex ante fiscal neutrality. Gini coefficients for the property wealth measures showed only slight fluctuations from 1975 to 1978 and then dropped substantially in 1979. Thus, the property wealth measure Gini coef ficients indicated movement toward ex ante fiscal neutrality only in 1979. 7. The survey results indicated that about half of the d is tric ts that put new operating levies on the ballot and answered question two indicated that they put these levies on the ballot at least partially in response to the incentive effect of the EYF. This percentage was higher for lower property wealth districts than for higher property wealth districts. 8. About 40 to 45 percent of the districts responding to questions three and four indicated some degree of substitution of state aid for local effort. These percentages were higher for lower property wealth districts than for the higher property wealth d istric ts. 146 9. About one-third of the districts answering question five indicated that the implementation of the EYF had some effect on whether they did or did not put a levy on the ballot. This percentage was about 45 percent for lower property wealth districts. 10. Many of the 106 open-ended comments on the question naire indicated that inflation, the threat of closing, and making up for a decrease in equalized millage caused by reappraisal were reasons for putting levies on the ballot. Several districts expressed their concern over the effects of H.B. 920 and how property tax reappraisal is treated in the EYF. Many districts expressed dissatisfaction with the way the EYF operates. Discussion The extent to which variation in total operating expenditures per pupil was reduced was dependent upon the distribution of state aid and the response of local school d is tric ts to the incentive effect of the EYF. More specifically, the higher the amount of basic aid (aid distributed through the EYF in an inverse relationship to property wealth) the more the variation in total operating expenditures per pupil could be expected to decrease ceteris paribus. Also, the more low expenditure districts that increased local tax efforts, the more the variation in total operating expenditures per pupil could be expected to decrease ceteris paribus. The major findings stated in 1 and 2 above suggest that with the exception of two measures indicating some slight reduction in variation in 1979 the distribution of basic aid and the incentive 147 effect of the EYF have failed to bring about a reduction in variation in total operating expenditures per pupil. The findings in 9 that only 33 percent of all districts and 45 percent of low property wealth dis tricts said that the EYF had any influence on whether or not they put a levy on the ballot lends credence to the idea that the incentive effect had a limited effect on reducing expenditure variation. The fact that the EYF was phased in over a four-year period probably had something to do with the lack of reduction in expenditure variation. Gensemer (1978) pointed out that rollback aid is included in total expenditure figures and since more of this rollback aid is allo cated to high property tax districts, this tends to widen the expenditure gap between rich and poor districts. Combined with local revenues and some categorical aid, rollback aid negates the equalizing effect of basic aid. Also save-harmless guarantees have hampered the equalizing effects of the workings of the EYF. Apparently the income and price elasticities of demand for education are low for many school districts, thus explaining why the effect of the matching grants of the EYF have been minimal. This is similar to the conclusion drawn by Stern (1973) in his Massachusetts study. If it can be assumed that expenditures per pupil are closely related to quality of educational opportunity, then large variations in total operating expenditures per pupil imply a wide variation in educational opportunity among school districts in Ohio. 148 Replacement of the rollback aid with a "circuit breaker" property tax relief system, increases in basic aid and a phasing-out of save-harmless guarantees would all contribute to reducing expendi ture variation among school d istric ts. However, the findings in number 7 above indicated that only half of those that put a levy on the ballot said that the EYF had any influence on their decision. Findings in number 10 indicated also that there were many other reasons why districts did or did not put levies on the ballot. Again, in terms of economic theory, the price and in come elasticities of demand for education are influenced by a number of factors other than the state aid formula. This idea is supported by James et al (1961) in their finding that the degree to which a community supports education depends more on the shared aspirations of its citizens than on any other single factor and therefore sometimes state aid stimulates local expenditures and sometimes it is substituted for local expendi tures. The findings in number 8 also indicate why variation in total operating expenditures was not reduced. As was shown in many of the determinants studies reviewed in Chapter II, many districts do sub stitute additional state aid for local effort. In this time of Proposition 13 psychology and the taxpayers' revolt it is not surprising that many districts find it advantageous to put off tax levies or propose small levies. Movement toward fiscal neutrality was assessed using a variety of criteria and several different measures of fiscal capacity. 149 This was done in recognition of the fact that many types of analyses are sensitive to the choice of criteria and the specification of variables used. The mixed results of the assessment of movement toward ex post fiscal neutrality depending on the criteria and fiscal capacity measure used are examples of this sensitivity. The simple regression results outlined in 3 above indicated different results between income wealth measures and property wealt- measures. Even between the two income wealth measures results varied. Residual analysis of these regressions indicated that perhaps a signi ficant linear term had been le ft out of the models. This could partially explain the difference in the results based on the two income measures. In light of the fact that the correlation between either of the income wealth measures and any of the property wealth measures is low (.24 or less for any measure for all six years), i t is not surprising that the relationship between either of these two wealth measures and a third variable (total operating expenditures per pupil) is not the same. To correct for the lack of fit in the simple regressions, multiple regressions were run with variables in addition to the property wealth variables as the independent variables. The movement toward ex post fiscal neutrality was uneven with 1978 showing a loss in gains achieved in the first two years under the EYF. Perhaps a more detailed analysis of the distributions of all the components of total operating expenditures would reveal reasons for this 1978 loss. The coefficient analysis indicated movement away from ex post fiscal neutrality for the income wealth measures and movement toward 150 ex post fiscal neutrality (especially in 1979) for the property wealth measures. This indicates that under the EYF the relationship between total operating expenditures per pupil and income has increased while the dependence of total operating expenditures per pupil on property wealth has lessened. These results are in line with the conclusion drawn by Stern (1973) that district power equalizing type systems could eliminate disparities due to property wealth distribution but not due to socioeconomic status of d is tric ts . Also, Caesar (1977) concluded that a fte r four years under the equal yield formula in Michigan, dis parities due to variation in tax base were narrowed but disparities due to property tax effort grew worse. The- strong movement toward ex post fiscal neutrality came in the fourth year of the EYF which was the first year of full funding. This supports the contention of Hickrod et al (1976) that the full impact of similar reforms in Illinois on tax referenda should not be expected to be felt until the third or fourth year of reform. Although the fiscal neutrality analysis of Gensemer (1978) did not include 1979 his results can be compared with the results of the present study for 1974 through 1978. Gensemer's regression results did not show the movement toward fiscal neutrality that the present study showed during 1976 and 1977. Also, the Gini coefficient analysis of the present study did not indicate the movement away from fiscal neutrality indicated by Gensemer for 1978. The reasons for these differences are unclear. The dif ference in the regression analyses could stem from the fact that 151 Gensemer used a sample while the present study employed the entire pop ulation. Also, the regression equations used different independent variables. In the regressions and the Gini analysis the specification of the fiscal capacity measure differed slightly as did the specifi cation of total operating expenditures. If these are the reasons for the differences, this is evidence of the sensitivity of this type of analysis. The ex ante fiscal neutrality analysis was an attempt to apply a less demanding notion of fiscal neutrality to the Ohio school finance system. Although the results indicated the same trend as the ex post Gini coefficients it should be kept in mind that the deriva tion of the capacity Gini coefficients was based on the assumption that the Gini coefficients are additive. A similar procedure presented by Ramesh Chaudhari in Hickrod e t al (1975) was also based on the assumption of additivity of Gini coefficients. However, Chaudhari has stated more recently that this assumption is questionable. Intuitively this assumption appears to be sound but more study needs to be done in this area. One could reasonably expect that if the disequalizing effects of tax rates were taken out of the Gini coefficients any movement toward fiscal neutrality detected using the total expenditure Ginis would appear even more pronounced. This was not the case. The magni tude of the movement was about the same for both cases. Recommendations for Further Research The results of this study and other studies indicate a need for further research described below. The criteria used in this study indicate whether certain changes have taken place but we are unable to infer wh^ these changes did or did not occur. The sharp movement away from fiscal neutrality in 1978 cited earlier is a case in point. Much more detailed investi gations, looking at the effects of each of the various components of total operating expenditures need to be conducted to determine why certain changes do or do not occur. A detailed analysis of the effects that the various "save harmless" guarantees have on the dis tribution of state aid to school districts would also be helpful in answering this question. Analyses of the types used in this study should be performed for future years to determine if the movements toward equity made in 1979 will continue under the fully funded formula. This is especially needed in light of the revisions made in the EYF which were effective for school year 1979-80. More in-depth studies of why districts do or do not put levies on the ballot are needed to help us understand the interaction of these effects with state aid incentive effects. Case studies of individual districts rather than statewide aggregated studies would be useful for sorting out the factors that determine school levy proposal strategy and passage. Also, the results of this study suggest that the intended effects of the incentive millage provision were of limited success. This leads to the question of whether this type of incentive 153 millage provision would be more or less effective in a state in which voter approval is not required to impose school operating levies on the taxpayers. A study comparing effects of the matching grants on operating expenditures in states requiring voter approval of local tax levies with states not requiring voter approval would be interesting. Most school finance studies have been conducted by looking at the school finance system as a whole or by looking at certain facets of the system. However, the school finance system does not operate in isolation but as a part of the overall economy. Therefore, the devel opment of a sophisticated econometric model of the Ohio economy with the school finance system as a part of this model could provide a more re a lis tic picture of what has happened in the past and what is likely to happen in the future in the school finance system. As the results of the present study indicate, only a modicum of success has been attained under the EYF with respect to the expectations for such a formula. Variation in total operating expendi tures per pupil has been reduced very little, the school finance system is far from being fiscally neutral, and the incentive effect was only partially successful. The underlying reason for this lack of success is not that a lack of research information exists but rather that the political considerations of the decision-making process of the legislative and executive branches of Ohio state government have out weighed this information. Grubb and Michelson (1974, p. 79) indicated that this problem is not unique to Ohio by the following statement: The fate of percentage equalizing plans in Massachusettes, Rhode Island, New York, Iowa and Utah casts some doubt on the ability of district power equalizing to come through the legislative process unscathed. Like other formula manipulations, its intent seems doomed to partial failure. Walter F. Garms (1974, p. 94) also expressed this idea in discussing the change of several states from a foundation type formula to a guaranteed yield type formula: Given past performance, it seems likely that the newly adopted programs will be any more successful than the foundation programs in avoiding the evils of underfunding and of catering to special interests that have caused disenchantment with the foundation programs. The important point here is that regardless of what research is recommended here or performed in the future, its impact will be minimal unless narrow political interests are put aside. APPENDIX A Definition of Variables 155 156 DEFINITION OF VARIABLES Pupil Count Variables Kindergarten Average Daily Membership (AVGKDM) Joint Vocational School Average Daily Membership (JVSADM) Total Average Daily Membership (TOTADM) as reported at bottom of Form SF-12 of the Ohio Department of Education (ODE) Full Time Equivalent Average Daily Membership (FTEADM) = TOTADM - 1/2 AVGKDM - 3/4 JVSADM Pupils Receiving Aid to Dependent Children (ADC) - for school years 1974 and 1975: ADC count used in Disadvantaged Pupil Program Fund (DPPF) - for school years 1976-1979: ADC count used in Disadvantaged Pupil Impact Aid Program (DPIA) Weighted FTEADM (WTFTE) = FTEADM + .5 (ADC) Each ADC pupil receives a weight of 1.5 Percent ADC = ADC * FTEADM Socioeconomic Variables Percent of TOTADM that are Classified as Minority (PCTMIN) The figures for school year 1978 were used for school year 1979. Percentage of Families Classified as Poor (PRCTPOOR) This variable represents the percentage of the families in the district that were below the poverty level according to the 1970 U.S. Census. Percent of Population Aged 18 Years or Less (PRCTLT18) This variable is the percent of the district population that is 18 years or less and is from the 1970 U.S. Census. Percent of Adults Who Completed 4 or More Years of College (PRCTCOLL) Taken from 1970 U.S. Census data for each d is tric t. Percent of Occupied Housing Units that are Owner Occupied (PRCTOWN) Taken from 1970 U.S. Census data for each d is tric t. 157 Average Weekly Wage (AWW) This variable is the average weekly wage for the calendar year prior to the s ta rt of the school year for all employees covered by workmen's compensation and as reported by the Ohio Bureau of Employ ment Services. Percent of Students Attending Nonpublic Schools (PCTNPA) = Nonpublic ADM f (Nonpublic ADM + FTEADM) Property Value Variables Assessed Value of Real Property (AVRLP) - for 1974 and 1975 AVRLP = (Residential + Commercial + Industrial + Agricultural + Mineral Real Property) + Public Utility Real Property - for 1976 through 1979, AVRLP = Total Valuation - Tangible Personal Property - (Non-Real Public Utility Property) Assessed Value of Non-Real Property (AVNRLP) = Total Valuation - AVRLP Public U tility Real Property (PUREAL) - for 1974 through 1977 PUREAL is calculated by multiplying the Total Public Utility Property for each of the respective years by the percentage of Total Public Utility Property for 1978 that was real property - for 1978 and 1979 actual Public U tility Real Property was used - for 1974 for Kyger Creek, Hannan Trace, North Gallia and Southwestern locals in Gallia County PUREAL = Total. Public Utility Valuation x .1 Adjusted Equalized Property Value (ADJEPV) = AVRLP + (.35 t Sales Ratio) + AVNRLP where Sales Ratio is the sales assessment ratio calculated for the county in which the central office of the school district is located. This variable is a measure of property value if real property were assessed at 35 percent of true market value. Equalized Value of Real Property (EVRLPR) = Equalized Valuation - Tangible Personal Property - Public Utility Tangible Property - this will be calculated for school years 1976-1979 only and Equalized Valuation will be line 4 of the Form SF-12 158 aAdjusted Equalized Property Valuation Per Pupil (AEPVPP) = ADJEPV t FTEADM aAdjusted Equalized Property Valuation Per Pupil Weighted by MFI (AEPPM) = ( AEPVPP 1 AEPVPP , / MFI \ MFI \AEPVPP + MFI/ \AEPVPP + MFIJ where MFI = Mean Family Income for school d istric ts from 1970 U.S. Census aAdjusted Equalized Property Valuation Per Pupil Weighted by AGI (AEPPPA) = / AEPVPP \ AEPVPP , ( AGI \ AGI . 1 AEPVPP + AGI/ VAE'PVPP + AGI/ where AGI = Adjusted Gross Income per 1976 state income tax return in school d is tric t Expenditures or Aid Variables Total Current Operating Millage (MILLTT) = Inside Millage + Outside Millage + UVS Millage all for operating purposes - for school years 1976-1979 MILLTT was "equalized" to adjust for variation in assessment ratios. This formed a new variable called Equalized Current Operating Millage (EQUMIL) which was used in SF-12 calculations. Rollback Aid (RBAID) = .10 (EVRLPR X EQUMIL X .001) for 1976-79 or .10 (VALRLP X MILLTT X .001) for 1974 and 1975 - This variable represents the 10 percent real property tax credit granted to all taxing districts by the state. Credits are given to taxpayers and the state reimburses the taxing d istric ts. The reimbursement going to school d istric ts is classified as state rollback aid to education. Local Revenue (LOCREV) = (Total Assessed Value x MILLTT x .001) - RBAID for 1974 and 1975 or = (EVRLPR X EQUMIL X .001) - RBAID for 1976-1979 - Local Revenue is a measure of total local, operating expendi tures. These variables were also calculated on a per weighted pupil basis by dividing by weighted FTEADM instead of FTEADM. That is: Adjusted Equalized Property Valuation Per Weighted Pupil = AEPVWP Adjusted Equalized Property Valuation Per Weighted Pupil Weighted by MFI = AEPWPM Adjusted Equalized Property Valuation Per Weighted Pupil Weighted by AGI = AEPWPA 159 Disadvantaged Pupil Program Fund Allocation (DPPF) - for school year 1974 the maximum allocation amount was used - for school years 1975-1979 actual DPPF expenditures were used Total State Support Adjusted (TSUPAD) = Total State Support + RBAID + DPPF for 1974 or = Total State Support + RBAID + DPPF + $60 x TOTADM for 1975 or = Total State Support + RBAID + DPPF + County Board Deduction for 1976-1979 where (1) Total State Support represents total state aid as reported on Form SF-12 (2) The $60 represents $20 and $40 flat grants provided to each d is tric t for each pupil (3) County Board Deduction is funding going to local districts to pay county district offices for administrative and supervisory services Total Operating Expenditure (TOPEX ) = TSUPAD + LOCREV This represents all operating expenditures excluding federal aid which is not used in this study. Total Operating Expenditure Per Pupil (TOPEPP) = TOPEX * FTEADM Total Operating Expenditures Per Weighted Pupil (TOPEWP) = TOPEX * WTFTE Basic Aid (BA) = TSUPAD - DPPF - Total Categorical Allowance - Other Special Education Allowance - Extended Service Allowance - 1976 County Board Deductions for 1974 and 1975 or = TSUPAD - DPPF - Total Categorical Allowance for 1976-1979 The County Board Deduction was subtracted from the f i r s t case to make i t comparable to the 1976-1979 Basic Aid which excludes the County Board Deduction from TSUPAD. Basic Aid Per Pupil (BAPP) = BA * FTEADM APPENDIX B District Type Definitions 160 161 For the variation analysis districts were divided into four district type groups. These four strata are defined below. 1. Large Central D istricts Districts associated with cities having a 1970 population of at least 49,000. These are termed "central cities" of Standard Metropolitan S tatistical Areas by the U.S. Census Bureau. There are 16 school d istric ts in this stratum. 2. Satellite City Districts Urban area districts associated with satellite cities, i.e. c itie s which are near and dominated by a larger city. These cities include bedroom suburbs, industrial enclaves, and balanced cities in the vicinity of larger central cities. There are 176 school d is tric ts in this'stratum . 3. Independent Small City Districts Districts associated with independent cities having between 5,000 and 42,000 population in 1970. These cities are employment centers, surrounded by rural areas. There are 74 school d is tric ts in this stratum. 4. Rural D istricts Districts without any city of over 5,000 population in 1970. These d is tric ts are rural, though some of them cover such a large area that their pupil enrollments are larger than those of. some urban d is tric ts . There are 351 school d istric ts in this stratum. APPENDIX C Regression Tables and Survey Instrument 162 TABLE 19 Results of Simple Regressions of Total Operating Expenditures Per Pupil (TOPEPP) on Various Wealth Measures, 1974 to 1979a 1974 1975 1976 1977 1978 1979 TOPEPP on AEPVPP Constant 633.89 700.98 777.75 837.95 874.43 1051.51 Regression Coefficient .011031 .011907 .010349 .010445 .011966 .011117 R2 .6383 .6312 .5783 .5461 .6330 .5947 TOPEPP on MFI Constant 365.35 424.71 422.49 451.23 491.96 640.23 Regression Coefficient .049617 .054516 .061086 .068457 .075902 .080078 R2 .1738 .1936 .2266 .2236 .2143 .2046 TOPEPP on AGI Constant 369.87 455.72 449.96 499.47 606.36 779.55 Regression Coefficient .036477 .038169 .043302 .047247 .047917 .049187 r2 .1565 .1581 . 1897 .1774 .1423 .1286 TOPEPP on AEPPM Constant 653.87 725.69 807.69 875.38 919.01 1097.65 Regression Coefficient .011646 .012642 .010741 .010763 .012400 .011482 r2 .6212 .6152 .5521 .5210 .6153 .5842 TOPEPP on AEPPPA Constant 637.37 709.86 797.92 868.75 911.99 1093.69 Regression Coefficient .011969 .012994 .010942 .010944 .012654 .011696 R2 .6221 .6164 .5475 .5173 .6183 .5899 definitions of all variable labels are found in Appendix A. TABLE 20 Results of Simple Regressions of Total Operating Expenditures Per Weighted Pupil (TOPEWP) on Various Wealth Measures, 1974 to 1979a 1974 1975 1976 1977 1978 1979 TOPEWP on AEPVWP Constant 614.40 683.90 758.19 808.92 843.59 1012.13 Regression Coefficient .011280 .011915 .010310 .010599 .012081 .011287 R2 .6421 .6254 .5769 .5519 .6463 .6133 TOPEWP on MFI Constant 314.16 368.27 356.90 374.21 402.15 534.83 Regression Coefficient .052683 ,057753 .064896 .072798 .081034 .086088 R2 .2022 .2314 .2653 .2617 .2500 .2410 TOPEWP on AGI Constant 319.33 394.49 384.19 421.90 516,25 673.46 Regression Coefficient .038704 .040921 .046142 .050508 .051746 .053696 r2 ,1819 .1936 .2235 .2099 .1698 .1562 TOPEWP on AEPWPM Constant 632.80 707.13 787.17 845.53 887.37 1057.94 Regression Coefficient .011960 .012687 .010701 .010940 .012528 .011664 r2 .6260 .6083 .5492 .5252 .6272 .6016 TOPEWP on AEPWPA Constant 615.02 690,39 776.92 838.04 879.67 1053.37 Regression Coefficient .012314 .013061 .010905 .011135 .012785 .011881 R2 .6278 .6096 .5444 ,5216 .6301 .6073 definitions for each variable label are found in Appendix A. TO: City, Local and ExemptedVillage School D istrict Superintendents of Districts with Equalized Current Operating Millage of 30 Mills or Less and with a Yield Per Pupil Per Mill less than 542 Per Pupil Per Mill for any of the School Years from 1975-1976 through 1973-1979 PROM: Jim Payton SU3JECT: Request for Assistance in Dissertation Research DATE: February 7, 1979 I am a PhD candidate in the Faculty of Educational Foundations and Research at The Ohio State University. I am oresently doing my dissertation research on the effects that the implementation of the Equal Yield Formula has had on Ohio school finance. One of the effects I'm studying deals with the idea of "incentive for effort." It was hypothesized that under the Equal Yield Formula, those school d istric ts with current operating millage of 30 mills or less and with yields per pupil oer mill of less than 342, would have incentive to increase their millage since additional local dollars would be matched with state dollars up to the 30 mill limit. Part of the analysis for this hypothesis will involve the responses to the attached questionnaire. 3oth 8ASA and 0S3A have expressed an interest in reviewing the results of this survey and other parts of this research for possible inclusion in their publi cations. Results will be provided to them. I would appreciate i t very much if you would take a few minutes to give me your opinions and some information by completing the survey and returning it to me in the enclosed self-addressed, postage-paid envelope. Your valuable opinions will be an important part of my dissertation research. If you were not employed by the school district for some of the earlier years, the opinions and information could be obtained from members of your staff who were. As you will note, your district IRN number has been written on the questionnaire. This has been done so that your responses can be grouped according to d istric t equalized millage and yield per pupil per mill which I already have coded by IRN number. Your responses will be kept strictly confidential and results will be reported only by d is tric t groups based on ranges of millage and yield per pupil per mi 11. Thank you for your cooperation in this matter. Enclosure SCHOOL DISTRICT RESPONSE QUESTIONNAIRE IRN It has been suggested that under the Equal Yield Formula, those school d is tric ts with current operating mlllage of 30 mills or less and with yields per pupil per mill of less than $42, would have incentive to increase their millage. This incentive effect was hypothesized since additional dollars would be matched with state dollars up to the 30 mill lim it, and additional mills would yield more than they would have under the former school foundation plan. With this idea in mind, please answer the following questions by placing a check mark ( ( /’) in the appropriate boxes for each year. 1975 -76 1976 -77 1977 -78 1978 -79 Yes No Yes No Yes No Yes No 1) Did your district have a new operating tax levy on U n n o n n n the ballot this year? o 2) If you answered "yes" to question 1, was this levy n O n n n n u o proposed at least partially In response to the in centive effect of the equal yield formula described above? 3) If you answered "yes" to question 1, was this levy a o LJ n O o n LJ less than i t would have been because additional state aid allowed you to ask for less millage than you would have had to ask for under the former foundation plan? 4) If you answered "no" to question 1, was the reason u n n LJ n n o a you did not have a levy on the b allot because your district received additional state aid under the equal yield formula? 5) In your opinion, has the implementation of the o n n a o u u a equal yield formula had an effect on whether your district did or did not put an operating levy on the ballot (answer for each year)? Comments: (use reverse side if necessary) BIBLIOGRAPHY Aiken, J. S. & Clune, W. H. Economic and Legal Justification of Fiscal Neutrality. Education and Urban Society, February 1973. Alexander, A. J. & Bass, G. V. Schools Taxes and Voter Behavior: An Analysis of School D istrict Property Tax Elections (R-1465-FF). Santa Monica: Rand Corporation, April 1974. Alexander, K. & Jordan, K. F. Equitable State School Financing. Educational Need in the Public Economy. Gainsvilie, Florida: The University of Florida Presses, 1976. Augenblick, J. G. School Finance at a Third Glance (a wall chart). Denver: Education Finance Center, Education Commission of the States, 1978. Barkin, David. The Equalizing Impact of State Aid to Education. Washington University, Institute for Urban and Regional Studies, 1967. Barlow, R. Factors Affecting American State Expenditures. Yale Economic Essays Fall 1967, T_. Barro, S. Theoretical Models of School District Expenditure Determination and the Impact of Grants-in-Aid (Rand Corporation Paper, R-867-FF). Santa Monica: Rand Corporation, February 1972. Barro, S. M. Alternative Post-Serrano Systems and Their Expenditure Implications. In John Pincus (Ed.), School Finance in Transition. Cambridge, Mass.: Ballinger Publishing Company, 1974. Benson, Charles S., Goldfinger, Paul M., Hoachlander, E. Gareth, & Pers, Jessica S. Planning for Educational Reform: Financial and Social Alternatives^ New York: Dodd Mead, Inc., 1974. Benson, C., Guthrie, J. et a l , The Fleischmann Report. New York: Viking Press, 197-3. Berke, J. S. Answers to Inequity: An Analysis of the New School Finance. Berkeley, California: McCutchan Publishing Corp., 1974. Berne, R. Equity and Public Education: Conceptual Issues of Measurement (Working Paper No. 4). New York: Graduate School of Public Administration, New York University, October 1977. Bowers, G. R. Wealth N eutrality, Percentage Formula the State of Ohio, Fiscal 1975-76. Columbus: Ohio Department of Education, April 15, 1977, Mimeographed. 167 168 Bowman, J. H. Cost and Benefit Spill outs as Factors Affecting Local Taxation for Public Schools'! Ph.D. dissertation, The Ohio State University, 1973. Browing, R. S. & Lang, D. C. School Finance Reform and the Courts After Rodriguez. In John Pincus (Ed.), School Finance in Transition Cambridge, Mass.: Ballinger Publishing Company, 1974. Busselle, T. N. The Texas Weighted Pupil Study (0EC-0-73-2700). July 1973. Caesar, G., McKarr, R. N. & Phelps, J. New Equity in Michigan School Finance: The Story of the Bursley~A"ct. rev, ed. Lansing: Senate Committee on Education, State of Michigan, June 1974. Caesar, G. Michigan Schools and the Crim-O'Neill Formula. Michigan School Board Journal, April 1977. Carroll, S. The Search for Equity in School Finance: Summary and Conclusions (Rand/R-2420-NIE). Santa Monica, California: Rand Corporation, March 1979. Cohen, D. C. School Finance and Social Policy: Serrano and Its Progress. In J. Pincus (Ed.), School Finance in Transition. Cambridge, Mass.: Ballinger Press, 1974. Cohn, E. Economics of State Aid to Education. Lexington, Mass.: Lexington Books, 1974. Complaint and Action for Declaratory Judgment filed in Hamilton County Common Pleas Court by Board of Education of the City of Cincinnati etc. against Martin W. Essex et al. on April 5, 1976. Coons, J. E., Clune, W. H. Ill & Sugarman, S. D. Private Wealth and Public Education. Cambridge, Mass.: Belknap Press of Harvard University, 1970. Cotton, J. & O'brien, I. . Unconditional Versus Matching Grants: A Further Comment. National Tax Journal, March 1968, 21. Cowle, I. M. School Aid in New York State. New York: Teachers College Press, 1968. Cubberly, E. P. School Funds and Their Apportionment. Ph.D. dissertation, Teachers College, 1905. Education Review Committee Staff. Report of the Education Review Committee. Columbus, Ohio: December 15, 1974. Feldstein, M. S. Wealth Neutrality and Local Choice in Public Education. American Economic Review, March 1975, 65, 75-89. 169 Firestine, R. E. & Carvellas, J. N. Estimating Local District Fiscal Response to Power-Equalized Forms of State Educational Grants-in-Aid. Syracuse: Education Finance and Governance Center, Syracuse University Research Corporation, 1975. Fowlkes, J. G. & Watson, G. E. School Finance and Local Planning. Chicago: Midwest Administration Center, University of Chicago, 1957. Garms, W. I. The Financial Dimensions of Recent School Finance Reforms. Planning and Changing, Summer 1974, pp. 93-97. Gensemer, B. Ohio School Finance: 1970 to Present. School Finance in Ohio: Background Papers for Conference on School Finance. Cleveland: Citizens' Council for Ohio Schools, April 1977. Gensemer, B. Fiscal Neutrality in Ohio School Finance. Westerville, Ohio: School Management In stitu te, April 11, 1978, Mimeographed. Gramlich, E. The Effects of Federal Grants on State and Local Expenditures: A Review of the Economic Literature. Papers and Proceedings, National Tax Association, 1969. Grubb, N. W. & Michelson, S. States and Schools. Lexington, Mass.: Lexington Books, 1974. Hamilton, D. & Cohen, H. Policy Making by Plebiscite: School Referenda. Lexington Mass.: D. C. Heath & Company, 1974. Harrison, William A., Jr. Am. Sub. S.B. 170: The F irst Year of Implementation, Staff Report to Education Review Committee, Ohio General Assembly. Columbus: July 21, 1976, Typewritten. Henderson, J. Local Government Expenditures: A Social Welfare Analysis. Review of Economics and S ta tis tic s , June 1969, 50. Hickrod, G. A., Hubbard, B. C. & Yang, T. W. C. The 1973 Reform of the Illin o is General Purpose Educational Grant-in^Aid: A Description and an Evaluation. Normal, 111.: Dept, of Educational Administration, Illinois State University, n.d. Hickrod, G. A., Hubbard, B. C., Yang, T. W. C. & Rasanond, T-. The 1973 Reform of the Illin o is General Purpose Grant-in-Aid: An Evaluation After Three YeaTsT Normal, 111,: Dept, of Educational Administration, Illinois State University, 1976. Hickrod, G. A., Yang, T. W. C., Hubbard, B. C. & Chaudhari, R. Measurable Objectives for School Finance Reform: A Further Evaluation of the Illinois School Finance Reforms of 1973. Normal, 111.: Illinois State University, 1975. Hirsch, W. The Economics of State and Local Government. New York: McGraw-Hill, 1970. 170 James, H. T., Thomas, J. A., & Dyck, H. J. School Revenue Systems in Five States. Palo Alto, California: Stanford University School of Education, 1961. Johns, R. L. The Development of State Support for the Public Schools. In R. L. Johns, K. Alexander & D. H. Stollar (Eds.), Status and Impact of Educational Finance Programs (Vol. 4). Gainesville, Fla.: National Educational Finance Project, 1971. Johns, R. L. & Salmon, R. G. The Financial Equalization of Public School Support Programs in the United States for the School Year 1968-69. In R. L. Johns, K. Alexander & D. H. Stollar (Eds.), Status and Impact of Educational Finance Programs (Vol. 4). Gainesville, Fla.: National Educational Finance Project, 1971. Johns, R. L. Evaluating State Plans for Financing the Public Schools. Proceedings of the 15th National Conference on School Finance, 1972, Washington, D.C. Jones, T. H. Review of Existing State School Finance Programs (Vol. 1). Washington, D.C., President1s Commission on School Finance. Kay, H. B. A Study of the Relationship Between Selected Socio economic Variables and Local Tax Effort to Support Education in Kentucky. Ph.D. dissertation, University of Florida, 1973. Leppert, J. & Routh, D. Providing for Special Education in Missouri. Denver: The Education Commission of the States, January 1978. Levin, B. & Cohen, M. A. Levels of State Aid Related to State Restrictions on Local School Decision Making (Urban In stitu teP ap er No. 727-1). Washington, D.C.: Urban In stitu te, 1973. Levin, M. Equal Educational Opportunity and the Distribution of Educational Expenditures. Education and Urban Society, February 1973. McLoone, E. P. Profiles in School Support 1969-70. Washington, D.C.: National Center for Educational S ta tistic s, U. S.-Government Printing Office, 1974. McLure, W. P. Measures of Educational Need. In Perspectives on Illinois School Finance: Illinois State Aid EqualTFation Study. Springfield, 111.: Illinois Office of Education, November 1977. McMahon, Walter W. An Economic Analysis of Major Determinants of Expenditures on Public Education. The Review of Economics and S ta tis tic s , August 1970. 171 Michel son, Stephen, What is a "Just" System for Financing Schools? Law and Contemporary Problems, Winter-Spring 1974. Morrison, H. C. School Revenue. Chicago: University of Chicago Press, 1930. Ohio Department of Education. Cost Per Pupil School Year 1974-1975. Columbus, Ohio: Division of Computer Services and S tatistical Reports, 1975. Ohio Department of Education. The Ohio Law for State Support of Public Schools. Columbus: Ohio Dept, of Education, 1976. Ohio General Assembly, Legislative Service Commission. Serrano v. Priest, Equal Protection of the Laws, and Ohio Public School Finance (Staff Research Report No. 106). Columbus, Ohio: 1972. Ohio Legislative Service Commission. Serrano v. P riest, Equal Protection of the Laws and Ohio Public School Finance, Report~No. 106. Columbus, Ohio: n.p. 1972. Patton, D. C. Cost Factors of Educational Programs in Indiana (Indiana School Finance Study Final Report). Ball State University, May 1978. Piele, P. K. & Hall, J. S. Budgets, Bonds and Ballots. Lexington Books, 1973. Rasamond, Tharin. Socio-Economic Variables and Illin o is Education Tax Referenda, 1971-1976. Unpublished doctoral dissertation, Illinois State University, 1976. Renshaw, E. F. A Note on the Expenditure Effect of State Aid to Education. Journal of Political Economy, April 1960, 68, 170-74. Robinson v. Cahill, 61 N.J. 473, 303 A. 2d 273 (1973), 67 N.J. 333, 339 A. 2d 193 (1975). School Management Institute. School Finance Study: An Analysis of the Equity of Ohio's School Finance Plan (Project 842). W esterville, Ohio, School Management In stitu te, n.d. Schoettle, F. P. Judicial Requirements for School Finance Reform and Property Tax Redesign: The Rapidly Evolving Case Law. National Tax Journal, 1972, 25, 455-72. Serrano v. P riest, 5 Cal. 3d 584, 487 p. 2d 1241, 96 Cal. Rptr. 601 (1972). Shapiro, S. Some Socioeconomic Determinants of Expenditures for Education: Southern and Other States Compared. Comparative Education Review, October 1962, 6, 160-66. 172 Sherman, J. D. An Evaluation of the Current Structure of School Finance in Ohio. Washington, D.C.: Lawyers' Committee for Civil Rights Under Law, March 1978. Sherman, J. D. & Long, D. C. School Finance in Louisiana. Washington, D.C.: Lawyers' Committee for Civil Rights Under Law, August 1978. Stern, D. The Effects of Alternative State Aid Formulas on the Distribution of Public School Expenditures in Massachusetts. Review of Economics and S ta tis tic s , February 1973, 6(3, 91-97. Thurow, L. The Theory of Grants-in-Aid. National Tax Journal, December 1966 19. Thurow, L. Unconditional Versus Matching Grants. National Tax Journal, September 1967, 20. Vincent, P. E. & Adams, K. Fiscal Responses of School D istric ts: A Study of Two States - Colorado and Minnesota (Report No. F78-3). Denver: Education Finance Center, Education Commission of the States, 1978. Wilde, J. Grants-in-Aid: The Analytics of Design and Response. National Tax Journal, March 1971, 24. Wilensky, G. State Aid and Educational Opportunity. Beverly Hills: Sage Publication, 1970. Yang, T. W. C. Measurement of School Revenue Equity in the States of Illinois, Michigan and Kansas. Ph.D. dissertation, Illinois State University, 1975. Yang, T. W. C. & Chaudhari, R. A Study of the Relationship Between Selected Socioeconomic Variables~and Local Tax Effort to Support Public Schools in Illinois. Norma1, 111.: Illinois State University Center for the Study of Educational Finance, 1976.