School Finance: Options for Improving Measures of Effort and Equity in
Title I (Letter Report, 08/30/96, GAO/HEHS-96-142).
Pursuant to a congressional request, GAO reviewed the measures of equity
and effort included in the Title I Education Finance Incentive Program,
focusing on: (1) several options for improving these measures; (2) the
characteristics of states with higher levels of equity and effort; and
(3) alternative ways of using these measures in allocating title I
funds.
GAO found that: (1) the Education Finance Incentive Program's definition
of effort does not include a comprehensive measure of states' funding
capacity, penalizes states with a high proportion of school-age
children, and provides no incentive for low-effort states to increase
their effort in funding elementary and secondary education; (2) the
title I definition of equity excludes a comprehensive measure of the
differences in students' needs among school districts and the
differences in purchasing power among districts; (3) adjustments to the
effort factor should include using states' total taxable resources in
establishing the states' education funding capacity and eliminating the
bias against low-effort states; (4) adjustments to the equity factor
should include rewarding states for increasing their level of equity in
education spending and considering more fully the differences in
educating higher-needs students; and (5) the allocation of title I funds
should be based on the number of children in poverty rather than on the
states' total number of school-age children.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: HEHS-96-142
TITLE: School Finance: Options for Improving Measures of Effort
and Equity in Title I
DATE: 08/30/96
SUBJECT: Educational grants
Economically depressed areas
Supplemental appropriations
Elementary education
Secondary education
Disadvantaged persons
IDENTIFIER: Education Finance Incentive Program
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO report. Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved. Major **
** divisions and subdivisions of the text, such as Chapters, **
** Sections, and Appendixes, are identified by double and **
** single lines. The numbers on the right end of these lines **
** indicate the position of each of the subsections in the **
** document outline. These numbers do NOT correspond with the **
** page numbers of the printed product. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
** A printed copy of this report may be obtained from the GAO **
** Document Distribution Center. For further details, please **
** send an e-mail message to: **
** **
** **
** **
** with the message 'info' in the body. **
******************************************************************
Cover
================================================================ COVER
Report to Congressional Requesters
August 1996
SCHOOL FINANCE - OPTIONS FOR
IMPROVING MEASURES OF EFFORT AND
EQUITY IN TITLE I
GAO/HEHS-96-142
Effort and Equity Measures
(104857)
Abbreviations
=============================================================== ABBREV
AFDC - Aid to Families With Dependent Children
COV - coefficient of variation
ESEA - Elementary and Secondary Education Act
IASA - Improving America's Schools Act
LEA - local educational agency
NCES - National Center for Education Statistics
TTR - total taxable resources
Letter
=============================================================== LETTER
B-274117
August 30, 1996
The Honorable Jeff Bingaman
The Honorable Christopher Dodd
The Honorable Carol Moseley-Braun
The Honorable Paul Simon
United States Senate
Disparities in per pupil funding for elementary and secondary
education within each state have long been a concern of parents,
teachers, state officials, the courts, and federal officials. Since
the early 1970s, these disparities have led poor districts in more
than 40 states to challenge the constitutionality of their states'
school finance systems. More than half of the state systems have
been challenged since 1989.\1
Since the 1960s, federal education funds have been allocated to areas
with high levels of need for additional educational services, such as
compensatory education or bilingual education. During the 1980s this
need increased as the nation's school-age population became
increasingly poor, racially and ethnically diverse, and at risk of
school failure.\2 Successfully educating at-risk populations depends,
in part, upon adequate and equitable funding. Schools have been
addressing the needs of at-risk children through a variety of
compensatory and education reform efforts, some of which use federal
education funds.
The largest single federal elementary and secondary education grant
program to local school districts--$6.7 billion in fiscal year 1996\3
--is title I of the Elementary and Secondary Education Act (ESEA) of
1965.\4 The purpose of title I is to improve the educational
opportunities of educationally deprived children by helping them
succeed in regular programs, attain grade-level proficiency, and
improve achievement in basic and more advanced skills. As
reauthorized by P.L. 103-382 in October 1994, these title I
educational services may be financed by four funding formulas for
this common purpose. The four formulas cover basic grants,
concentration grants, targeted grants, and Education Finance
Incentive Program grants.\5 The 1994 legislation, for the first time,
authorized funding for the Education Finance Incentive Program
beginning in fiscal year 1996, although no funding was specified in
fiscal year 1996 appropriations for that program or for targeted
grants. (See app. I for a description of these grant programs.)
Many complex policy and technical issues surround policymakers'
decisions about whether to provide title I funding through the
Education Finance Incentive Program formula in future years. The
Clinton administration proposed not funding this program for fiscal
year 1997 because "the formula would reward states that make a high
effort and are highly equalized, but it would not consistently target
funds on states with high concentrations of poor children."\6 Those
who support the Education Finance Incentive Program suggest that if a
state's spending for education increases and spending disparities
among a state's districts decrease, the federal government is able to
more effectively allocate title I funds to provide truly supplemental
educational resources to disadvantaged children.\7
Title I's Education Finance Incentive Program would provide
additional funds to states with high levels of "fiscal effort" for
education (that is, high state spending relative to the state's
ability to pay) and equity in per pupil spending.\8 In our June 1994
correspondence,\9 we identified potential weaknesses in the proposed
measures of effort and equity used in the title I program. In
addition, Members of Congress have also called for improvements in
these measures. Consequently, you asked us to provide a more in-
depth analysis of these measures and provide options for improving
them. Specifically, this report
-- examines the measures now included in title I's Education
Finance Incentive Program to reflect state fiscal effort for
education and equity in per pupil spending,
-- proposes several options for improving these measures,
-- describes the characteristics of states with higher levels of
effort and equity under both the current definitions and the
options we developed, and
-- proposes alternative ways the options we developed could be used
in allocating funds under the Education Finance Incentive
Program.
For our evaluation of the effort and equity factors, we reviewed the
relevant school finance literature and consulted with school finance
experts. On the basis of this review, we developed the following
criteria to assess the quality of the effort factor. The effort
factor should (1) include an indicator of states' ability to pay that
is comprehensive, reflecting all sources of potential revenue-
raising capacity states may use to fund education; (2) avoid using
indicators that are not directly related to state fiscal effort; and
(3) provide a direct incentive for increased effort.
To assess the quality of the equity factor, we adopted the following
criteria: (1) the measure of spending disparities should be
comprehensive, (2) the measure of students' needs should be as
comprehensive as possible, (3) spending differences should be
adjusted for differences in purchasing power across school districts,
and (4) the equity factor should provide an incentive for states to
further reduce spending disparities.
We consulted with experts to review the measures and options we
formulated.\10 See appendix II for a full discussion of scope and
methodology. We conducted our review from May 1995 through June 1996
in accordance with generally accepted government auditing standards.
--------------------
\1 School Finance: Three States' Experiences With Equity in School
Funding (GAO/HEHS-96-39, Dec. 19, 1995).
\2 School-Age Children: Poverty and Diversity Challenge Schools
Nationwide (GAO/HEHS-94-132, Apr. 29, 1994).
\3 Under P.L. 104-134, however, about $1.3 billion of this $6.7
billion will not be available for fiscal year 1996 awards (school
year 1996-97) until Oct. 1, 1996.
\4 We use the term title I to refer to title I, part A, Improving
Basic Programs Operated by Local Educational Agencies, of ESEA, as
added by the Improving America's Schools Act of 1994, P.L. 103-382.
\5 The basic grant formula allocates grants on the basis of the
number of poor children, while concentration grants are allocated to
areas with concentrations of poor children--those with a poverty rate
over 15 percent or more than 6,500 poor children. Targeted grants
further focus grants on the highest poverty areas by allocating the
greatest per pupil funding to areas with the highest poverty rates or
number of children in poverty.
\6 See Department of Education, Justifications of Appropriation
Estimates to the Congress: Fiscal Year 1997, Vol. I, p. B-20.
\7 For a discussion of related title I (formerly known as chapter 1)
issues, see Commission on Chapter 1, Making Schools Work for Children
in Poverty: A New Framework Prepared by the Commission on Chapter 1
(Washington, D.C.: Dec. 1992).
\8 An individual state, however, may be relatively high on one
measure and relatively low on the other. For example, Tennessee has
achieved a high level of equity in per pupil funding but exerts a low
level of fiscal effort for education.
\9 Title I Formula in S. 1513 (GAO/HEHS-94-190R, June 7, 1994).
\10 The following experts were involved in initial discussions or
reviewed drafts of this report: Michael Compson (Department of the
Treasury), William Fowler (Department of Education's National Center
for Education Statistics [NCES]), Frank Johnson (NCES), Martin Orland
(NCES), Lawrence Picus (University of Southern California), Wayne
Riddle (Congressional Research Service), William Sonnenberg (NCES),
Stephanie Stullich (Department of Education), and Deborah Verstegen
(University of Virginia).
RESULTS IN BRIEF
------------------------------------------------------------ Letter :1
The measures of both effort and equity that are used in the Education
Finance Incentive Program could be improved. Improved measures would
more efficiently reward states that increase their level of fiscal
effort in raising education revenue or successfully diminish spending
disparities among districts.
Under title I, the effort factor measures each state's spending per
pupil relative to its ability to pay, relative to this ratio for the
nation as a whole. A strength of this factor is that it considers a
state's ability to pay when assessing its effort in raising revenue
for education. We found, however, that the measures included in the
effort factor could be improved. We developed three alternative ways
to measure effort that are more comprehensive and do not penalize
states with high proportions of school-age children. Two of these
alternatives also reward states for improving their level of effort
over time. One of these two provides a further incentive for
low-effort states to increase their effort.
Similarly, the equity factor could also be improved. The current
definition of equity has two components: a measure of spending
disparities and a measure of student needs. The measure of spending
disparities among a state's districts takes into account spending in
every district in the state and rewards states for being equitable.
The measure of student needs, however, does not adjust for all major
differences in student needs across school districts. For example,
it adjusts for the extra costs associated with educating poor
students but does not adjust for the extra costs associated with
students who have disabilities or limited English proficiency. In
addition, the current measure does not adjust for differences in
purchasing power among a state's districts (for example, differences
in the costs of services, such as teacher salaries, that affect how
much a state's dollars can buy). Furthermore, the equity definition
could more effectively reward states for improving their level of
equity over time. We developed five alternative ways to address
equity that are based on two different measures of spending
disparities. These alternatives take into account differences in
student needs and purchasing power. Three of these alternatives also
more effectively reward states for improving their level of equity
over time.
Current definitions of effort and equity are more strongly related to
certain demographic characteristics than are the alternatives we
developed. For example, the options we developed do not penalize
states with high levels of child poverty the way the current effort
factor does.
The current formula for title I's Education Finance Incentive Program
allocates funds to states using, in addition to the effort and equity
factors, the state's total number of school-age children, rather than
children in poverty--those who are the focus of all other title I
allocation formulas. Consequently, some states could benefit from
the Education Finance Incentive Program even though they do not have
high levels of poverty. We constructed four alternative allocation
formulas using the alternative effort and equity factors we
developed. Two of the four alternatives we present are based on
numbers of poor children, rather than all school-age children.
BACKGROUND
------------------------------------------------------------ Letter :2
Total expenditures in all U.S. elementary and secondary schools in
school year 1993-94 reached an estimated $285 billion. Education is
the largest single expenditure category in state budgets, accounting
for about 20 percent of total state spending in fiscal year 1994.
Elementary and secondary schools receive most of their funds from
state and local revenues. Federal aid has mainly focused on
providing services to educationally disadvantaged children through
categorical, program-specific grants. In school year 1992-93, state
and local shares of total education spending were roughly equal,
estimated at 45.6 percent ($113 billion) for states and 47.4 percent
($118 billion) for local educational agencies (LEA). The federal
share was 6.9 percent ($17 billion).\11
Although most of the activities promoting equity in education take
place within states, the federal role in supporting equity has been
discussed since the 1960s, when concern was voiced over states'
inappropriate use of federal funds intended to improve equity on
behalf of disadvantaged children. In summer 1993 the Senate held
hearings on the federal role in school finance equalization.\12
Subsequently, the Congress amended ESEA to further help disadvantaged
children by improving the targeting of title I funds to local
education agencies and schools with relatively high levels of poor
students.
Title I is a federal program that provides remedial education
services to low-achieving students in high-poverty elementary and
secondary schools. Title I funds are intended to supplement, not
supplant, local and state education funding. Until ESEA was amended
by the Improving America's Schools Act (IASA) of 1994, title I grants
to local education agencies were distributed under two formulas--the
basic grant formula and the concentration grant formula. In an
effort to increase the amount of aid going to the neediest children,
between 1988 and 1994 the statute required that 10 percent of the
appropriations to LEAs were to be distributed using the concentration
grant formula. That formula generally allocated funds only to those
LEAs in counties where eligible children\13 equaled either at least
6,500 or 15 percent of the total population aged 5 through 17. The
rest of the appropriation was distributed under the basic grant
formula (which is based on numbers of poor school-age children
multiplied by a cost factor reflecting a state's per pupil spending).
In 1994, the Congress sought to provide greater targeting of title I
aid although opinions varied as to the best way of doing it. In the
end, IASA made some technical changes to the basic and concentration
grants and added two more title I funding streams--targeted grants
and the Education Finance Incentive Program. The targeted grant
formula may use money appropriated for title I in excess of the
fiscal year 1995 level. Targeted grants are similar to basic grants
except that poor and other children counted in the targeted formula
are assigned weights based on the county's or LEA's child poverty
rate and the number of poor school-age children. This formula
generally reflects the recommendations for using weighted pupil
formulas in title I made by the Commission on Chapter 1, the RAND
Corporation, and GAO.\14
Under the targeted grant formula, the higher the poverty rate or
number of poor children in the county or LEA, the higher the title I
grants per formula child.\15
The 1994 reauthorization of title I included an effort and equity
bonus--additional dollars--through a new Education Finance Incentive
Program to encourage states to have more equitable education finance
systems. Although part of title I, this program is funded
separately. The 1994 reauthorization authorized $200 million for the
Education Finance Incentive Program for fiscal year 1996.\16 As of
July 1996, however, no funding has been appropriated. (See app. I
for more background information related to title I and appropriation
issues.)
The Education Finance Incentive Program defines effort as the ratio
of state spending for elementary and secondary education per pupil to
the per capita income of state residents. The effort factor,
however, can be no less than 95 percent and no more than 105 percent
of the national average. Per capita income is used as a proxy for a
state's ability to pay for education spending. Thus, the effort
factor measures each state's actual spending as a percentage of its
ability to spend.
The equity factor measures the variation in per pupil spending across
a state's districts divided by the state's average per pupil
spending. Additional weight is given to the number of poor pupils to
reflect the higher cost of educating these children. The equity
factor is subtracted from 1.3 for use in the allocation formula.
--------------------
\11 School Finance: Trends in U.S. Education Spending
(GAO/HEHS-95-235, Sept. 15, 1995).
\12 See, for example, William L. Taylor, Testimony Before the
Subcommittee on Education, Arts and the Humanities, Senate Committee
on Labor and Human Resources, Aug. 3, 1993.
\13 Known as "formula children," these were children aged 5 through
17 (1) in poor families, according to the latest decennial census and
applying the Bureau of the Census's standard poverty income
thresholds; (2) in families above the poverty income level for a
family of four but receiving Aid to Families With Dependent Children
(AFDC) payments; and (3) in certain institutions for neglected or
delinquent children. Poor children constitute approximately 96
percent of the formula children.
\14 Commission on Chapter 1, Making Schools Work for Children in
Poverty; Iris C. Rotberg and James J. Harvey, Federal Policy
Options for Improving the Education of Low-Income Students (Santa
Monica, Calif.: The RAND Corporation, 1993); Remedial Education:
Modifying Chapter 1 Formula Would Target More Funds to Those Most in
Need (GAO/HRD-92-16, July 28, 1992). For a discussion of these
weights, see Wayne C. Riddle, Education for the Disadvantaged:
Analysis of 1994 ESEA Title I Amendments Under P.L. 103-382, Report
for Congress No. 94-968 EPW (Washington, D.C.: Congressional
Research Service, Nov. 18, 1994).
\15 Riddle, Education for the Disadvantaged.
\16 For fiscal years 1997 through 1999, no specific authorization
limits have been set.
IMPROVING EFFORT MEASURES
------------------------------------------------------------ Letter :3
One of the strengths of the effort factor used in the Education
Finance Incentive Program is that it considers a state's ability to
pay when determining the level of effort. However, the current
effort factor could be improved for three reasons. First, its
measure of ability to pay--per capita personal income--is not as
comprehensive as another that is available: total taxable resources
(TTR). Second, because its measure of spending, which is per
student, is related to a measure of ability to pay which is per
capita, the factor penalizes states with high proportions of
school-age children. Third, the effort factor does not adequately
reward states for improving their level of effort over time.
The effort factor's measure of ability to pay is not as comprehensive
as it could be because per capita income excludes many taxable
resources that states are able to use for financing education. To
the extent these additional sources of funding capacity are not
equally available in all states, the per capita income measure
overstates the funding capacity of some states and understates it for
others.
The current effort factor also penalizes states with high percentages
of school-age children in their populations. It measures the ratio
of state spending, expressed on a per pupil basis, and state funding
capacity, expressed on a per capita basis. This calculation
introduces the percentage of the state's school-age population into
the measure of state effort. The effect is to inappropriately
penalize states with a high percentage of school-age children because
the percentage of a state's population that is school-age is
unrelated to its level of effort.\17
Moreover, the current effort factor could more effectively encourage
states to increase their level of effort over time by (1) including a
bonus based on rate of increase or decrease in effort over time and
(2) eliminating the current requirement that the effort factor be at
least 95 percent and no more than 105 percent of the average effort
of all states. A fiscal incentive, or bonus, would reward states for
increasing their level of educational effort over time, rather than
rewarding only those that already have high effort. The 95-percent
floor undermines the incentive for low-effort states to increase
their effort in funding elementary and secondary education. Under
current law, states with very low effort may increase their level of
effort considerably, yet receive no additional dollars.
Two of the three alternative effort factors we developed include a
bonus based on the rate of increase or decrease in effort over time,
in addition to other modifications we have made to address the
potential drawbacks of the current effort factor (see app. III).
One of these two alternatives also eliminates the requirement that
the effort factor be at least 95 percent and no more than 105 percent
of the average effort of all states.
Table 1 compares the characteristics that define the current measure
of effort in title I with the measures we propose as options A, B,
and C. Option B is more comprehensive than option A because it
considers both the current level of effort and the change in effort
for each state over time. Option C is the same as option B, except
that the effort factor is not constrained by the current requirement
to be between 95 and 105 percent of the average effort. For a
state-by-state breakout on each of these options, see appendix III.
Table 1
Comparison of Current Title I Effort
Definitions With Alternative Definitions
Characteristic Title I Option A Option B Option C
---------------------- ---------- ---------- ---------- ----------
Includes a X X X
comprehensive measure
of states' ability to
pay
Is not biased against X X X
states with a high
proportion of school-
age children
Provides a direct X X
incentive for
improvement in effort
over time
Provides an incentive X
for low-effort states
to increase their
effort (no minimum)
----------------------------------------------------------------------
--------------------
\17 Researchers have noted that states with a large percentage of
their population that is school-age may have a more difficult time
raising needed resources for education than those with a smaller
percentage that is school-age. "All else being equal, states who are
able to spread their educational costs across a larger population
base (i.e., those with higher population to pupil ratios), can more
easily generate a given level of per pupil spending than can states
with greater numbers of students relative to their population,"
according to Martin Orland and Carol Cohen, Meeting the Challenge of
Devolution: How Changing Demographic and Fiscal Contexts Affect
State Investments in Education (Washington, D.C.: The Finance
Project, Feb. 1996), p. 2.
IMPROVING EQUITY MEASURES
------------------------------------------------------------ Letter :4
The equity definition in title I's Education Finance Incentive
Program contains two components. Of the two, the measure of spending
disparities is more comprehensive than the measure of student needs.
The measure of spending disparities\18 is a good overall measure
because it takes into account per pupil spending in all of each
state's school districts. The measure of student needs, however,
only explicitly takes into account the greater needs of one type of
pupil--those who are poor--in determining per pupil expenditures.
Although the definition allows other types of higher needs students
to be considered, such as students with disabilities or limited
English proficiency, they are not explicity included.
Table 2 compares the current equity measure and the five options we
developed on the following characteristics: (1) the
comprehensiveness of the measures of variation in education spending
levels among districts in the state, that is, whether they include
all districts and consider low-spending districts; (2) the ability of
measures of student needs to take into account the cost differences
of educating different target populations (students who are poor,
have limited English proficiency, or have disabilities); (3) the
inclusion of a comprehensive measure of purchasing power; (4) the
inclusion of a direct incentive for states to improve their levels of
equity over time; and (5) the presence of minimum and maximum limits.
Of the five options we provide, E, G, and H are the more
comprehensive. Each considers both the current level of equity a
state has achieved and the recent progress the state has made toward
achieving equity in education spending. To the extent possible
within the limitations of the data currently available, we took into
account differences in student needs related to numbers of students
who were poor, had limited English proficiency, or had
disabilities.\19 Whether policymakers prefer option E or option G
depends on their interest in measuring variation in spending levels
for all school districts in the state (option E), or focusing on a
state's ability to bring low-spending school districts up to the
median (option G). Option H is the same as option E, except that it
does not contain the limits each of the other equity options we
developed does.\20 (For more information on each of these options,
see app. IV.)
Table 2
Comparison of Title I Equity Definition
With Alternative Definitions
Option Option Option Option Option
Characteristic Title I D E F G H
---------------- ------- ------- ------- ------- ------- -------
Considers X X X X
spending levels
for all
districts in the
state
Considers extent X X
to which
district
spending levels
fall below the
median
Considers
differences in
student needs
Poor X X X X X X
Limited English X X X X X
proficiency
Disabilities X X X X X
Includes a X X X X X
comprehensive
measure of
differences in
purchasing power
Provides a X X X
direct incentive
for improvement
in equity
Includes limits X X X X
(minimum and
maximum)
----------------------------------------------------------------------
--------------------
\18 The coefficient of variation (COV) is the standard deviation (a
common statistical measure of variation) of state and local per pupil
spending for primary and secondary education within the state,
divided by the average level of per pupil spending in the state.
\19 In five states data were not available on the number of pupils
with disabilities; however, it is expected that data will be provided
for all but two states when data become available for 1993-94 from
the Common Core of Data collected by NCES. Moreover, the number of
pupils with limited English proficiency is based on the decennial
census and may underestimate the current number.
\20 For the other equity options, we used limits of at least 0.95 and
no more than 1.30.
CHARACTERISTICS OF STATES WITH
HIGHER LEVELS OF EFFORT AND
EQUITY
------------------------------------------------------------ Letter :5
We examined the demographic characteristics of states with higher
levels of effort and equity under the current title I definition and
under the new definitions we developed. We looked at the
relationship between these effort and equity factors and (1) state
median household income, (2) variations in median household income
among districts, (3) state percentage of school-age children in
poverty, and (4) variations in percentage of school-age children in
poverty among districts.
Under the current effort factor, we found that states with higher
levels of child poverty rates do significantly less well than those
states with lower levels of child poverty. There was no significant
relationship between a state's rate of child poverty and the options
we developed; thus, the options we developed do not penalize
high-poverty states the way the current effort factor does. For
further discussion of these analyses, see appendix III.
When we examined the correlation between the various equity factors
and variability in median household income among districts, we found
that the lower the variability in median household income across
districts in the state, the higher the equity factor, and vice versa.
The strength of the association was strongest, however, for the
current equity factor and weakest for the three measures that
considered improvement in equity over time--options E, G, and H (see
app. IV, table IV.2).
IMPROVING THE EDUCATION FINANCE
INCENTIVE PROGRAM FORMULA
------------------------------------------------------------ Letter :6
The current formula for title I's Education Finance Incentive Program
allocates funds to a state based on the effort factor and the equity
factor multiplied by the state's total number of school-age children,
rather than the number of children in poverty--those who are the
focus of all other title I allocation formulas. Consequently, some
states could benefit from the Education Finance Incentive Program
even though they do not have high levels of poverty.
The alternative allocation formulas we developed not only use the
alternative effort and equity factors we developed, but also propose
using the number of children in poverty. Two of the four
alternatives we present are based on the number of poor school-age
children, rather than all school-age children. (For illustrative
alternative allocation formulas using the effort and equity measures
we developed, see app. V.\21 ) Eight of the 10 poorest states would
receive greater funding using the two alternative formulas we propose
that are based on children in poverty than they would under the most
targeted of title I formulas, the targeted grant formula.
--------------------
\21 These measures also have implications for another federal
program, Impact Aid, title VIII of IASA, the purpose of which is to
help pay the operating costs of LEAs that are affected by federal
activities. (See app. VI.)
CONCLUSIONS
------------------------------------------------------------ Letter :7
The definitions of effort and equity in title I's Education Finance
Incentive Program could be improved in a number of ways. The
definition of effort used in this program could be improved by (1)
using a more comprehensive measure of ability to pay, (2) eliminating
the bias against states with high proportions of school-age children,
(3) providing a direct incentive for states to improve their level of
effort over time, and (4) eliminating the lower limit for the effort
factor. The definition of equity used in current law could be
improved by (1) more fully considering differences in students' needs
among districts, (2) considering differences in purchasing power
among a state's districts, and (3) rewarding states for improving
their level of equity in education spending, not just for already
being equitable. The formula for allocating funds under this program
would better target funds to states with higher proportions of
children in poverty if it were based on the numbers of poor children
rather than all school-age children.
MATTERS FOR CONGRESSIONAL
CONSIDERATION
------------------------------------------------------------ Letter :8
Should the Congress decide to fund title I's Education Finance
Incentive Program, it may want to improve the effort and equity
measures and the way they are used in the allocation formula by
considering the options we have presented in this report.
Specifically, we believe that the Congress may wish to consider
-- reducing the floor on the effort factor so that low-effort
states are rewarded for increased effort;
-- modifying the effort factor to eliminate the penalty on states
where a high percentage of the population is school-age;
-- using, in the effort factor, a more comprehensive measure of
states' revenue raising capacity, such as the total taxable
resources indicator published by the Secretary of the Treasury;
-- including in the effort and equity factors a bonus for
improvement over time;
-- expanding the needs component of the equity factor to include
children with limited English proficiency and children with
disabilities;
-- adjusting the equity factor for differences in the cost of
educational services across each state's districts; and
-- basing the allocation formula on the number of poor school-age
children rather than all school-age children.
AGENCY COMMENTS AND OUR
EVALUATION
------------------------------------------------------------ Letter :9
The Department of Education provided written comments on a draft of
this report (see app. VII). The Department expressed concern about
our analysis of the impact of the Education Finance Incentive Program
on the targeting of title I funds and whether the incentive formula
can be expected to provide a meaningful incentive for states to
change their school funding systems.
The Department of Education was concerned that the Education Finance
Incentive Program, even with the refinements we proposed, would tend
to redistribute title I funds away from many higher poverty states
and school districts. They stated that the Education Finance
Incentive Program formula has a devastating impact on targeting
because of a combination of factors, including (1) states with low
fiscal effort tend to be high-poverty states with fewer resources,
(2) the equity factor draws funds away from some high-poverty states
while benefiting some low-poverty states, and (3) the incentive
formula allocates funds based on the total number of school-age
children rather than numbers of poor children.
Regarding the Department's first point (that states with low fiscal
effort tend to be high-poverty states), we found that although this
was true using the current effort measure, it is not true using the
measures of effort we developed. Our analysis shows that while there
is a significant negative correlation between the current effort
measure and a state's poverty level (that is, poorer states would get
fewer dollars), this is not the case with each of the three effort
options we developed.
Contrary to the Department's second point (that the equity factor
draws funds away from some high-poverty states while benefiting some
low-poverty states), as our report points out, we found no
correlation between a state's score on the current equity measure and
a state's poverty level. We did find, however, that states with high
levels of variation in income levels and poverty rates across their
districts did less well using the current equity measure than states
with lower levels of variation; our equity measure options
ameliorated this problem somewhat.
With regard to the third point (that the incentive formula allocates
funds based on the total number of school-age children rather than on
the number of poor children as in the other title I formulas), we
agree that this is true with the current formula. In our draft
report, however, one of the three allocation alternatives we
identified uses numbers of poor, rather than all, school-age children
as a basis for allocating funds under the Education Finance Incentive
Program. We have also added a fourth allocation alternative that
uses numbers of poor children. Both of those allocation alternatives
would target more dollars for poor states such as Alabama, Arkansas,
Kentucky, Mississippi, New Mexico, South Carolina, Tennessee, and
West Virginia--8 of the 10 poorest states--than would the targeted
grant formula.
The Department also stated that the report sidesteps the issue of
whether the Education Finance Incentive Grant formula can be expected
to provide a meaningful incentive for states to change their school
funding systems. Although this issue was not the focus of our study,
we state in our report that some experts question whether the level
of funding that may or may not be appropriated for the Education
Finance Incentive Program would be of sufficient size to have any
effect on the plans of state or local educational agencies to provide
greater levels of effort or equity (see app. I).
We concur with the Department's comment that the provisions
restricting a state's effort factor to between 95 and 105 percent of
the national average result in weakening the incentive for states to
increase their level of effort. In response to the Department's
comment, we developed another option for the effort measure that does
not include these minimum and maximum constraints, and included this
in our analysis. This fourth alternative allocation formula based on
the number of poor school-age children also includes this
unconstrained effort option as well as an unconstrained equity
option. This allocation alternative, as previously noted, would
result in targeting more dollars to 8 of the 10 highest poverty
states than would the targeted grant formula.
---------------------------------------------------------- Letter :9.1
We are sending copies of this report to the Secretary of Education,
appropriate congressional committees, and other interested parties.
If you wish to discuss the contents of this report, please call me on
(202) 512-7014 or Eleanor Johnson, Assistant Director, on (202)
512-7209. Major contributors to this report are listed in appendix
VIII.
Carlotta C. Joyner
Director, Education and
Employment Issues
BACKGROUND
=========================================================== Appendix I
Under title I, federal funds are authorized to school districts to
provide supplementary educational services for low achievers in areas
with children in poverty. As reauthorized by P.L. 103-382 in
October 1994, these title I educational services may be financed by
four funding formulas for this common purpose. The four funding
formulas are for basic grants, concentration grants, targeted grants,
and Education Finance Incentive Program grants. In fiscal year 1996,
approximately $6.7 billion was appropriated for two of these funding
formulas: basic grants and concentration grants.\22
No funds have been appropriated for either targeted grants or the
Education Finance Incentive Program for fiscal year 1996.
Basic grants are generally allocated based on numbers of poor
school-age children multiplied by a "cost factor"--a measure based on
a state's average per pupil spending.\23 Concentration grants are
based on numbers of school-age children in areas with high
concentrations of poverty--where over 6,500 or 15 percent of the
children are poor--and a measure of per pupil spending. Targeted
grants provide an even greater focus on allocating funds to the
highest poverty areas because they target the greatest per pupil
funding to the areas with the highest poverty rates or numbers of
children in poverty.
--------------------
\22 However, under P.L. 104-134, about $1.3 billion of these basic
and concentration grant funds will not be available for fiscal year
1996 awards (school year 1996-97) until Oct. 1, 1996.
\23 In a previous review of the title I formula, we noted that the
current cost factor needs improvement as it rewards states for
spending more on education without considering differences across
states in their ability to spend on education. See Remedial
Education: Modifying Chapter 1 Formula Would Target More Funds to
Those Most in Need (GAO/HRD-92-16, July 28, 1992).
APPROPRIATION ISSUES RELATED TO
TITLE I'S EDUCATION FINANCE
INCENTIVE PROGRAM
--------------------------------------------------------- Appendix I:1
Many complex policy and technical issues surround congressional
policymakers' decisions about whether to provide some title I funding
through the Education Finance Incentive Program formula. It remains
unclear whether additional title I funds will be appropriated in the
near future and, if so, whether they will be made available for the
two title I formulas created in the 1994 reauthorization: targeted
grants and Education Finance Incentive Program grants.\24 Funds
appropriated in excess of the fiscal year 1995 title I appropriation
may be spent for targeted grants. However, in part because the
excess amounted to only 0.5 percent of the basic and concentration
grant appropriations for fiscal year 1996, no funds were spent for
targeted grants for fiscal year 1996.\25 In addition, no funds were
earmarked for the Education Finance Incentive Program, although $200
million was authorized for fiscal year 1996. If funding for title I
increases in future years, it remains unclear whether the Congress
will appropriate funds for targeted grants or the Education Finance
Incentive Program.
For fiscal year 1997, the Clinton administration has proposed funding
targeted grants at $1 billion, while decreasing funds for basic
grants by about $500 million. The proposal is intended to enhance
the ability of the poorest communities to provide supplementary
instructional services to disadvantaged students. The administration
proposes not to fund the Education Finance Incentive Program for
fiscal year 1997 because "the formula would reward states that make a
high effort and are highly equalized, but it would not consistently
target funds on states with high concentrations of poor children."\26
In addition, some experts question whether the level of funding that
may or may not be appropriated for the Education Finance Incentive
Program would be of sufficient size to have any effect on the plans
of state or local educational agencies to provide greater levels of
effort or equity.
Those who support funding the Education Finance Incentive Program may
note that title I funds are seen as supplementing a basic level of
state and local funding for instructional services to compensate for
the additional educational needs that accompany concentrations of
poverty.\27 If, however, there are spending disparities among a
state's school districts, title I funding "may only help make up some
of the gap in resources available to disadvantaged children compared
to those received by the advantaged."\28 To the extent that a state's
spending for education increases and spending disparities among a
state's districts decrease, the federal government is able to more
effectively target funds to provide truly supplemental educational
resources to disadvantaged children.
If funds are appropriated for the Education Finance Incentive Program
in the future, under current law each state's share would be
determined by the formula in figure I.1.
Figure I.1: Formula for
Determining State Share
(See figure in printed
edition.)
Notes: This formula differs from other title I, part A, formulas in
two ways: (1) it uses school-age population, not poor school-age
population, as a basis for allocation; and (2) it does not include an
expenditure factor, as other formulas do, that rewards high-spending
states. For more information, see Riddle, Education for the
Disadvantaged.
Throughout the rest of this report, we will use the term "equity
factor" to refer to the "1.3 minus the equity factor."
Funds would then be allocated to districts within each state based on
their share of the total of other title I funds for school districts:
basic grants, concentration grants, and targeted grants. Each state
is to be allotted at least 0.25 percent of the total appropriation.
The effort and equity factors in the Education Finance Incentive
Program are intended to provide additional dollars to those states
that have relatively high fiscal effort (in order to provide adequate
levels of funding for education) and those states with relatively low
disparities in per pupil funding across districts.
--------------------
\24 For more information on the title I basic, concentration, and
targeted grant formulas, see Wayne Riddle, Education for the
Disadvantaged: Analysis of 1994 ESEA Title I Amendments Under P.L.
103-382, Report for Congress No. 94-968 EPW (Washington, D.C.:
Congressional Research Service, Nov. 18, 1994).
\25 See Wayne Riddle, Title I, ESEA: Current Status and Issues,
Report for Congress No. 96-380 EPW (Washington, D.C.: Congressional
Research Service, Apr. 26, 1996).
\26 See Department of Education, Justifications of Appropriation
Estimates to the Congress: Fiscal Year 1997, Vol. I, p. B-20.
\27 For a discussion of related title I (formerly known as chapter 1)
issues, see Commission on
Chapter 1, Making Schools Work for Children in Poverty: A New
Framework Prepared by the Commission on Chapter 1 (Washington, D.C.:
Dec. 1992).
\28 See Riddle, Education for the Disadvantaged, p. 17.
DEFINITION OF EFFORT IN TITLE
I'S EDUCATION FINANCE INCENTIVE
PROGRAM
--------------------------------------------------------- Appendix I:2
The Education Finance Incentive Program defines effort using two
components: state spending for education and state ability to pay
(see app. III). The measure for state spending for education is the
state's average per pupil expenditure for public elementary and
secondary education. The measure of the state's ability to pay for
services is the state's average per capita income.\29 Those states
with high effort, that is, high state spending relative to their
ability to pay, are rewarded for this under the Education Finance
Incentive Program.
A state's level of effort is compared with that for the nation as a
whole to develop an "effort factor" to be used in the Education
Finance Incentive Program formula. If the index is 1.00, the state's
level of spending for education, relative to its per capita income,
is the same as it is for the nation as a whole. Those states with
spending for education relative to their ability to pay that is
greater than that for the nation receive a factor higher than 1.00;
those with spending relative to ability to pay that is lower than
that for the nation receive a factor lower than 1.00. No state,
however, may have an effort factor higher than 1.05 or lower than
0.95. Limiting the range from 0.95 to 1.05 limits the degree to
which states receive a smaller share because of their much lower
effort or a larger share because of their much higher level of
effort.
--------------------
\29 Under current law, averages used to calculate the effort factor
are determined using 3 years of data to minimize the effect of
changes from year to year.
DEFINITION OF EQUITY IN TITLE
I'S EDUCATION FINANCE INCENTIVE
PROGRAM
--------------------------------------------------------- Appendix I:3
The definition of equity used in title I's Education Finance
Incentive Program includes two components: a measure of spending
disparities and a measure of student needs. The law also contains a
number of other adjustments that take into account complexities
arising from various types of school districts (for example,
elementary, secondary, and unified), extremely small school
districts, and other factors (see app. IV).
The first component measures the level of disparity in current per
pupil expenditures across the state's school districts. There are a
variety of ways to measure spending disparities; title I uses the
coefficient of variation (COV). The COV is the standard deviation (a
common statistical measure of variation) in spending per student
among all districts within the state, divided by the average level of
spending per student in the state.
The second component is a partial accounting, through use of a
weighting factor, for differences in student needs across districts
within the state; it specifically focuses on differences in the
number of poor students. Students with no need for additional
services are weighted by a factor of 1.0; poor students are weighted
by a factor of 1.4.\30 Such a weighting system recognizes that it may
not be best to have equal spending per pupil across districts if the
needs of the children in those districts are not equal. For example,
additional local, state, and federal dollars might be targeted to
districts with high concentrations of poor children to provide
services to compensate for their greater educational needs.
The equity factor is then constructed by subtracting the measure of
spending disparities--the COV adjusted by a weighting component for
poor students--from 1.3.\31 For most states, the equity factor ranges
from 1.2 to 1.0. One state, Hawaii, has only one school district
and, therefore, no variation in spending, so it receives an equity
factor of 1.3. For similar reasons, Washington, D.C., and Puerto
Rico also receive an equity factor of 1.3.\32 Current law also
includes a provision that those states that meet the disparity
standard under Impact Aid\33 --Alaska, Kansas, and New
Mexico--receive an equity factor of 1.2.
--------------------
\30 In addition, the law allows the Secretary of Education to take
into account the needs of other types of pupils, such as pupils with
limited English proficiency or disabilities. While data were not
available on numbers of children with disabilities for eight states
at the time the law was passed, data will soon be available for all
but two states. The law also allows the Secretary to consider
differences in purchasing power across the states' districts when
determining the level of equity. Indexes on differences in
purchasing power across the states' districts were not available at
the time the law was enacted but have been developed recently.
\31 The law also allows the Secretary to incorporate other valid and
accepted methods to measure equity not reflected in the COV method.
\32 One expert suggested that Hawaii, Washington, D.C., and Puerto
Rico receive an undeserved advantage because of this situation and
that they should receive an average score on the equity factor,
rather than the highest score.
\33 See app. VI for a discussion of the Impact Aid program.
SCOPE AND METHODOLOGY
========================================================== Appendix II
We reviewed the formula used in title I's Education Finance Incentive
Program, focusing on the components of the formula related to effort
and equity, which the Congress refers to as effort and equity
factors. Our study was designed to answer the following questions:
(1) How can the current effort factor be improved? (2) How can the
current equity factor be improved? (3) How are state demographic
characteristics, such as poverty rate or median income, related to
state scores on the effort and equity factors? (4) How can the
options we developed be used in alternative ways to allocate funds
under the Education Finance Incentive Program?
To answer these questions, we analyzed the strengths and weaknesses
of title I's effort and equity factors using criteria we developed
after reviewing the literature and consulting with experts. On the
basis of our review of the literature and consultation with experts,
we developed alternative effort factors for possible use in the
Education Finance Incentive Program using a universe sample of school
district data from the Department of Education's National Center for
Education Statistics (NCES) for school year 1991-92. School district
spending data were collected by the Bureau of the Census for NCES.
We also developed alternative equity factors that include two
different measures of spending disparities and consider differences
in purchasing power across a state's school districts. See
appendixes III and IV for a description of the methods used to
produce the alternatives for the effort and equity factors for title
I's Education Finance Incentive Program.\34 To the extent possible
within the limitations of the data currently available, we took into
account differences in student needs related to numbers of students
who were poor, had limited English proficiency, or had
disabilities.\35 We also developed illustrative state allocations
under current and alternative title I Education Finance Incentive
Program formulas as well as under the targeted grant formula (see
app. V).\36
--------------------
\34 See app. VI for a discussion of some of the methodological
strengths and weaknesses of the definition of equalization used in
Impact Aid.
\35 In five states, data were not available to take into account the
number of pupils with disabilities; however, it is expected that data
will be provided for all but two states when data become available
for school year 1993-94. Moreover, the number of pupils with limited
English proficiency is based on the decennial census and
underestimates the current number.
\36 Although existing NCES data are sufficient for the purposes of
illustrating the general effects of changes in the effort and equity
factors, should the Congress wish to pursue these strategies, NCES
may need to address a number of cases of missing data or
irregularities in the school district database.
MEASURING EFFORT
========================================================= Appendix III
The effort factor in the Education Finance Incentive Program provides
additional title I aid to those states with education spending
relative to their ability to pay that is higher than other states.
Current law defines this factor as the state's average per pupil
expenditure divided by the state's average per capita income relative
to that for the nation as a whole. In the current law, averages are
determined using 3 years of data to minimize the effect of changes
from year to year. However, no state's effort factor can be less
than 95 percent of the nation's average or more than 105 percent.
Under P.L. 103-382, funds are to be allocated to states based on the
state's number of children aged 5 through 17 multiplied by both the
effort factor and the equity factor. Each state is to be allotted at
least 0.25 percent of the total appropriation.
EFFORT FACTOR IN CURRENT LAW
COULD BE IMPROVED
------------------------------------------------------- Appendix III:1
The effort factor in the current law, which considers states'
education spending relative to their ability to pay, could be
improved. First, the measure of ability to pay used in the law, per
capita income, is not as comprehensive as the one we used: total
taxable resources (TTR). Second, the current effort factor penalizes
states with high proportions of school-age children. Finally, the
existing effort factor could more directly reward states for
increasing their level of effort, not just for having high effort.
ALTERNATIVE TITLE I EFFORT
FACTORS
------------------------------------------------------- Appendix III:2
We developed three alternative effort factors for title I's Education
Finance Incentive Program using TTR as a measure of a state's ability
to finance services. TTR, defined and compiled by the Department of
the Treasury, considers both personal income and the gross state
product for each state. TTR takes into account all income produced
within a state, whether received by residents or nonresidents, or
retained by business corporations. TTR is a more comprehensive
indicator of taxable resources than personal income alone, in part
because it also considers income produced in a state but received by
nonresidents.
In our alternative effort factors, we consider both spending and
ability to pay per student, rather than spending per student and
ability to pay per capita, as the current law does. The ability to
pay per student is a better measure of a state's ability to finance
educational services for students than ability to pay per capita.
Moreover, using ability to pay per capita, rather than per student,
results in lower Education Finance Incentive grants for states with
high proportions of school-age children--those states usually
intended to benefit from education-related grants.
In one of the alternatives we developed, option B, we increase the
level of the reward to states that increase their level of effort
over time. For example, two states may have the same level of effort
in the current year. But while one state increased its level of
effort from the previous year, the other state decreased its level of
effort. In the first of the alternatives we present, option A, these
two states would have the same effort factor because they each
currently have the same level of effort--as would happen under
current law. In the second alternative, option B, the state that
increased its level of effort from the previous year would have a
higher score than the state that decreased its level of effort.
Option C is similar to B but without the limits of 0.95 and 1.05, so
that low-effort states would have a greater incentive to increase
their effort.
The details of the three alternative effort factors we developed are
as follows. The first alternative, which we refer to as option A, is
based on state and local funding for elementary and secondary public
education (kindergarten through grade 12) divided by the state's
TTR.\37 To develop an index, we divided the state figure by a
comparable figure for the nation as a whole. This effort index, as
is the case for the current law index, is limited to no less than
0.95 and no more than 1.05. Option A is based on data for school
year 1992-93.
The second alternative, which we refer to as option B, is an index
that also considers the rate of change in effort over time. We
determined the rate of change in effort over time by comparing option
A with a similar effort factor for the previous school year, 1991-92.
For example, if the state's effort factor was 1.01 under option A,
and it had improved by 3 percent from the previous year, under option
B its effort factor would be 1.01 multiplied by 1.03 (1 plus 0.03),
or 1.04. In contrast, if the state's effort factor was 1.01 under
option A, but its effort had decreased by 3 percent from the previous
year, under option B its effort factor would be 1.01 multiplied by
0.97 (1 minus 0.03) or 0.98. Again, the factors were limited within
the bounds of 0.95 and 1.05.
The third option we developed, option C, is similar to option B in
that it provides an additional incentive for change over time. In
addition, option C eliminates the lower limit of 0.95 and the upper
limit of 1.05. We allowed only half of the index to vary, however,
because otherwise the gap between the extremes would be too wide. In
other words, option C equals 0.5 multiplied by an unconstrained
effort factor, plus a constant of 0.5. Under option C, the effort
factor would range from a low of 0.72 for the District of Columbia to
a high of 1.21 for West Virginia. Table III.1 compares current and
alternative effort factors.
Table III.1
Current and Alternative Effort Factors
Current
effort
State factor\a Option A Option B Option C
----------- --------- ---------- ---------- ----------
Alabama 0.95 0.95 0.95 0.90
Alaska 1.05 1.05 1.01 1.11
Arizona 0.95 1.05 1.04 1.02
Arkansas 0.95 1.01 1.01 1.01
California 0.95 0.95 0.95 0.94
Colorado 0.95 0.97 0.98 0.99
Connecticut 1.05 0.99 0.97 0.98
Delaware 1.05 0.95 0.95 0.88
District of 1.05 0.95 0.95 0.78
Columbia
Florida 1.00 0.95 0.95 0.94
Georgia 0.95 0.96 1.02 1.01
Hawaii 0.95 0.95 0.96 0.91
Idaho 0.95 1.01 0.97 0.98
Illinois 0.96 0.95 0.95 0.94
Indiana 1.02 1.05 1.05 1.09
Iowa 0.98 1.05 1.05 1.07
Kansas 0.96 1.03 1.03 1.02
Kentucky 0.98 0.99 0.98 0.99
Louisiana 0.98 0.95 0.95 0.94
Maine 1.05 1.05 1.05 1.17
Maryland 1.05 0.97 0.97 0.98
Massachuset 1.05 0.95 0.95 0.94
ts
Michigan 1.05 1.05 1.05 1.16
Minnesota 1.00 1.05 1.04 1.02
Mississippi 0.95 0.95 0.95 0.92
Missouri 0.95 0.95 0.95 0.95
Montana 1.05 1.05 1.00 1.09
Nebraska 0.95 1.03 1.05 1.04
Nevada 0.95 0.95 0.95 0.89
New 0.95 1.00 1.00 1.00
Hampshire
New Jersey 1.05 1.05 1.04 1.09
New Mexico 0.95 1.05 1.03 1.02
New York 1.05 1.05 1.04 1.05
North 0.95 0.95 0.95 0.90
Carolina
North 0.95 0.96 0.95 0.97
Dakota
Ohio 1.05 1.05 1.05 1.09
Oklahoma 0.95 1.05 1.05 1.03
Oregon 1.05 1.05 1.05 1.11
Pennsylvani 1.05 1.05 1.04 1.03
a
Puerto Rico 0.95 0.95\b 0.95\b 0.78\b
Rhode 1.05 1.05 1.05 1.05
Island
South 0.98 1.02 1.01 1.00
Carolina
South 0.95 0.95 0.97 0.97
Dakota
Tennessee 0.95 0.95 0.99 0.86
Texas 0.95 1.04 1.00 1.00
Utah 0.95 1.05 1.05 1.08
Vermont 1.05 1.05 0.98 1.11
Virginia 0.95 0.95 0.95 0.96
Washington 0.95 1.05 1.05 1.04
West 1.05 1.05 1.05 1.21
Virginia
Wisconsin 1.05 1.05 1.05 1.11
Wyoming 1.05 1.05 0.96 1.09
----------------------------------------------------------
\a Liane White, Education Finance Incentive Grant Under ESEA Title I,
Report for Congress 95-963 EPW (Washington, D.C.: Congressional
Research Service, Sept. 12, 1995).
\b Comparable data for our comprehensive measure of states' ability
to pay, TTR, developed by the Department of Treasury, were not
available for Puerto Rico. Therefore, we gave Puerto Rico the same
score under options A, B, and C that it would receive under the
definition in the current law--the lowest effort factor calculated
for any state.
--------------------
\37 This is mathematically equivalent to dividing a state's spending
per pupil by the state's TTR per pupil.
CORRELATIONS BETWEEN CURRENT
AND ALTERNATIVE EFFORT FACTORS
AND DEMOGRAPHIC CHARACTERISTICS
------------------------------------------------------- Appendix III:3
We examined the correlations between effort and states' demographic
characteristics such as state median income, within-state deviations
in median household income, state percentage of school-age children
in poverty, and within-state deviations in percentage of school-age
children in poverty. We examined the extent to which title I's
current effort factor and the alternative effort factors we developed
were related to these demographic factors.
We found that states with higher levels of child poverty rates do
significantly less well with the current effort factor than those
states with lower levels of child poverty. However, there was no
significant relationship between a state's rate of child poverty and
the options we developed. Thus, the options we developed do not
penalize high-poverty states the way the current effort factor does.
When we examined the correlation between the various effort factors
and variability in median household income across a state's
districts, we found that the variable "within-state deviations in
median household income" was positively correlated with the current
effort factor.\38 That is, the higher the variability in median
household income across districts in the state, the higher the
current effort factor, and vice versa. There was no significant
correlation, however, between this variable and the options we
developed (see table III.2). Moreover, neither the current effort
factor nor the options we developed were significantly correlated
with state median household income or within state deviations in
child poverty levels.
Table III.2
Correlations Between the Effort Factors
and State Demographic Characteristics
Current
Characteristic measure Option A Option B Option C
---------------- ---------- -------- -------- --------
State median -0.06 -0.22 -0.16 -0.21
household
income
Within-state 0.29\a 0.09 0.08 0.10
deviations in
median
household
income
Percentage of -0.34\a -0.20 -0.15 -0.21
school-age
children in
poverty
Within-state 0.03 0.05 0.09 0.04
deviations in
percentage of
school-age
children in
poverty
----------------------------------------------------------
\a Significant at the 0.05 level.
--------------------
\38 Two variables, state median household income and within-state
deviations in median household income, were adjusted for differences
in purchasing power.
MEASURING EQUITY
========================================================== Appendix IV
The equity factor of the Education Finance Incentive Program formula
provides that additional title I aid go to those states with
relatively low disparities in per pupil expenditures among local
educational agencies (LEA).\39 The particular statistical measure
used to determine the level of spending disparities in each state
under current law is the coefficient of variation (COV). The COV is
defined as the standard deviation divided by the mean. Assuming a
normal distribution, approximately two-thirds of all school districts
will fall within one standard deviation of the mean, or average. For
example, if the state's average per pupil spending is $6,000 and the
standard deviation is $2,000, then approximately two-thirds of
districts in the state will be spending between $4,000 and $8,000 per
pupil. However, average spending levels vary greatly from state to
state. To provide fair comparisons across the states, each state's
standard deviation is divided by its average spending level.\40 Using
this example, the COV is $2,000 divided by $6,000 or 0.33.
In addition to defining the measure of equity as the COV, the law
specifies that the COV should be weighted by the number of poor
children in each school district for each state. The Congressional
Research Service has pointed out that "the effect of the additional
weighting for poor children is that expenditure disparities in favor
of LEAs with relatively large numbers of poor children would reduce a
State's measured COV, while expenditure disparities in favor of LEAs
with relatively low numbers of poor children would increase a State's
COV."\41
To provide a simplified example of how weighting might work, assume a
state has two school districts--one with only poor students and
another with no poor students. Also assume that the per pupil
spending is the same for the two districts, except that 40 percent
more funding per poor student is provided to fund additional
services. Such a state would appear to have a significant spending
disparity. However, a weighted COV takes into account the
differences in student needs.
In the current law, the 1.4 weighting per poor student stems, in
part, from title I, which authorizes an additional 40 percent in per
pupil spending to provide services to educationally disadvantaged
children in high poverty areas. However, appropriations for title I
have generally been less than half of what is authorized. As a
result, in a recent study, researchers used 1.2 as a weighting for
poor students.\42
The law also contains a number of other adjustments that take into
account complexities arising from various types of LEAs (for example,
elementary, secondary, and unified), extremely small LEAs, and other
factors. Under current law, separate COVs are used for elementary,
secondary, and unified school districts. A statewide COV is then
determined by calculating a weighted average based on the number of
weighted students (that is, counting poor students as 1.4) for each
type of district occurring within the state. Some states have only
unified school districts serving students from kindergarten through
grade 12; other states have elementary, secondary, and unified school
districts. The law excludes spending in extremely small school
districts, such as those in remote areas, because spending in these
school districts may be atypical.
--------------------
\39 To simplify, we have used the term "equity factor" rather than
"1.3 minus the equity factor" in this and subsequent appendixes.
\40 This measure also helps to provide comparability over time. For
example, if we wanted to examine the level of disparity of many years
ago, when spending levels in nominal dollars were half of what they
are now, such figures would be comparable. For example, if the state
mean had been half of what it is now, $3,000, and, similarly, the
standard deviation was $1,000, also half, the COV would be $1,000
divided by $3,000 or 0.33, the same as above.
\41 Riddle, Education for the Disadvantaged, p. 16.
\42 For more information, see Thomas B. Parrish, Christine S.
Matsumoto, and William J. Fowler, Disparities in Public School
District Spending: 1989-90, Statistical Analysis Report No. 95-300
(Washington, D.C.: National Center for Education Statistics, Feb.
1995).
METHODOLOGY FOR DEVELOPING
ALTERNATIVES
-------------------------------------------------------- Appendix IV:1
Our work is based on statistical analyses of fiscal and demographic
data from the nation's school districts for school years 1991-92 and
1989-90, the latest years for which expenditure data for the universe
of school districts were available. In determining alternative
equity factors, we also treated the various types of school
districts--elementary, secondary, and unified--separately and
computed a statewide weighted average. As is the case under current
law, we excluded school districts with fewer than 200 students
enrolled and districts that reported they had no schools. We also
excluded districts with expenditures that were likely to be atypical,
such as those devoted primarily to vocational or special education.
In examining expenditure equity, we used total current expenditures,
which do not include expenditures for debt services or capital
outlay.
Data limitations include an underreporting of the number of children
who have limited English proficiency and those with disabilities.
Although NCES data are sufficient for the purposes of illustrating
alternatives to the current effort and equity factors, NCES may want
to address a number of cases of missing data or irregularities in the
school district database. The data available to us on school
district finances and numbers of special needs children by school
district were compiled by NCES and the Bureau of the Census. Data on
numbers of children with limited English proficiency came from
parents' reports to the Bureau of the Census about whether their
children speak English "not well" or "not at all." In addition, five
states in school year 1991-92 and eight states in school year 1989-90
did not provide numbers of children with disabilities.\43 In
developing an equity factor for these states without data, we were
not able to take into account differences among school districts in
the number of children with disabilities.
--------------------
\43 We use the term children with disabilities to refer to children
with an individualized educational plan under the Individuals With
Disabilities Education Act. Data on numbers of such children were
not available for five states in 1991-92: Kentucky, Ohio, Oklahoma,
Pennsylvania, and Virginia. In 1989-90, data were not available for
Alaska, Kansas, Vermont, and Washington, in addition to Kentucky,
Ohio, Oklahoma, and Pennsylvania. In 1993-94, data will be available
for all but two states: Kentucky and Ohio.
DEVELOPMENT OF ALTERNATIVE
DEFINITIONS OF EQUITY
-------------------------------------------------------- Appendix IV:2
To improve the current measures of equity used in the Education
Finance Incentive Program, we examined (1) the comprehensiveness of
the various measures of equity available, (2) the comprehensiveness
of the measures accounting for differences in student needs across
districts, (3) the effect of including a measure of purchasing power
across districts, and (4) the effect of including a direct incentive
for states to improve their levels of equity over time.
THE COMPREHENSIVENESS OF
VARIOUS EQUITY MEASURES
------------------------------------------------------ Appendix IV:2.1
We reviewed literature related to measures of school finance equity,
most particularly a set of expert papers prepared for the Department
of Education in 1992 to evaluate whether the measures of equity used
in the Department's Impact Aid program could be improved.\44 Although
focused on Impact Aid (see app. VI), these papers informed the
discussion related to equity measures used in title I as well.
Some of these experts generally agreed that an equity measure would
be better if it took into account a large portion of each state's
school districts in determining the disparity in per pupil spending
across the state, as the COV does. Robert Berne and Leanna Stiefel
also suggested that another measure of spending disparities be
used--the McLoone Index.
The COV takes into account per pupil spending in all of each state's
school districts and, therefore, is a comprehensive measure of
equity. Another measure of equity, the McLoone Index, focuses on
equity for school districts that spend less than the median.\45
This index is the ratio of the sum of expenditures for districts
below the median to what the expenditures in these districts would be
if they were able to spend at the median level per pupil. Where per
pupil expenditures are equal for all the districts in the state that
are at or below the median, the McLoone Index
is 1.0.
--------------------
\44 These papers, published in a special issue of the Journal of
Education Finance, Vol. 18, No. 1 (Summer 1992), were written by
the following experts on equity measures in school finance: Robert
Berne and Leanna Stiefel, K. Forbis Jordan, Allan Odden, and Richard
Salmon. Joel Sherman provided a summary article as editor for the
special issue.
\45 It is also possible to construct a modified McLoone Index that
examines all districts with spending below, for example, the 75th or
80th percentiles rather than below the 50th percentile, or median.
THE COMPREHENSIVENESS OF
MEASURES OF DIFFERENCES IN
STUDENT NEEDS
------------------------------------------------------ Appendix IV:2.2
Experts suggested that taking into account student needs would
improve current measures of equity. For example, if one district in
a state has many pupils that are poor, have limited English
proficiency, or otherwise need special educational services, it may
be appropriate for the state to provide more aid for that district
than for districts without high proportions of these at-risk pupils.
Therefore, it may be necessary to make adjustments to consider that
one district has greater student needs by weighting its students
according to their need for additional services. If, for example,
the cost of educating a student with limited English proficiency is,
on average, 20 percent more than the cost of educating a student
without additional needs, these students would then be weighted 1.2.
If the district were able to spend at the level needed to cover these
additional services, its expenditures per weighted pupil would show
that it was spending at a level comparable with districts with fewer
students with additional needs.
We chose to use a set of weights developed for the NCES report,
Disparities in Public School District Spending: 1989-90, that takes
into account differences in student needs across school districts.\46
The researchers assigned students with disabilities a weight of 2.3
because the cost of educating such children is generally 2.3 times
the cost of educating children who do not need special educational
services, although the cost of educating children with specific types
of disabilities varies widely.\47 The report used weights of 1.2 for
children from poor families or those who have limited English
proficiency. This additional 0.2 weighting for students in poverty
stems from an estimate based on the average title I allocation per
student divided by average revenues per student. The rationale for
using a weight of 1.2 for children with limited English proficiency
is based on an expectation that they will need additional educational
services, comparable with those for poor children, although school
districts may generally spend less than this currently.
We also consulted with the Congressional Research Service on the
issue of whether to use 1.2 or 1.4 as a weight for the number of poor
children and those with limited English proficiency. While the
current equity measure uses the higher weight of 1.4 to adjust for
the greater needs of poor children, it does not adjust for the
greater needs of other students needing additional services, such as
students with disabilities or limited English proficiency. But
because we were taking into account the additional costs associated
with educating these students and because there may be some double
counting, that is, students may be weighted twice if they are both
poor and have limited English proficiency (or have disabilities and
limited English proficiency), we decided the weight of 1.2 was more
appropriate in this case.\48 More precise estimates are not available
on the cost of educating students who may have multiple types of
special needs; moreover, data currently available do not allow us to
estimate numbers of such children by school district.
--------------------
\46 Parrish, Matsumoto, and Fowler, Disparities in Public School
District Spending: 1989-90.
\47 This cost estimate is based on analysis of data from a nationally
representative sample. For more information, see M.T. Moore, E.W.
Strang, M. Schwartz, and M. Braddock, Patterns in Special Education
Service Delivery and Cost, Contract No. 3000-84-0257 (Washington,
D.C.: Decision Resources Corp., 1988). More recent studies have
resulted in a similar figure.
\48 We note, however, that the pros and cons of using specific
weights are debated by some experts and program officials. For
example, some experts would argue that a higher weight of 1.4 would
be appropriate for both poor students and those with limited English
proficiency. (See, for example, Deborah Verstegen and Kent McGuire,
"The Dialectic of Reform," Educational Policy, Vol. 5, No. 4,
pp. 386-411.) One program official suggested that it may be more
appropriate to use a lower weight for students with limited English
proficiency than for poor students, in part because of the double
counting.
ADJUSTING FOR DIFFERENCES IN
PURCHASING POWER
------------------------------------------------------ Appendix IV:2.3
We believe that adjusting for differences in purchasing power across
a state's districts is useful in providing more comparable measures
of spending levels, or spending disparities, across districts. For
example, district A may be able to hire teachers of the same quality
at 80 percent of the cost of district B because district A may be in
a part of the state that offers lower housing costs, greater
availability of desirable services, or better weather. If each
district spends $4,000 per pupil, district B will not be able to
provide the same level of services to its students as district A.
Therefore, to provide greater comparability, we adjusted the spending
levels of the various districts to take into account differences in
purchasing power reflected in the cost of hiring and retaining
teachers.
We used a teacher cost index recently developed for the National
Center for Education Statistics.\49 While an index that examines
differences in the cost of living is also available by district, we
believe that the NCES teacher cost index is better suited to our
purpose of providing comparability across districts because it
considers the purchasing power of districts in determining
personnel-related costs, a major cost to school districts. Our focus
is on the district's ability to provide comparable educational
services to its students, rather than on whether teachers' salaries
are adequate given the cost of living in their area.
Not all costs, however, vary within the state. For example, the
costs of books, instructional materials, and other supplies and
equipment tend to vary little within a state or, for some items, the
nation. Therefore, we used the teacher cost index to adjust only the
portion of expenditures generally estimated to be related to
personnel costs.\50 We used an estimate developed by Stephen Barro
for NCES; he calculated that 84.8 percent of total current
expenditures are personnel costs, including salaries, fringe
benefits, and some purchased services.\51
--------------------
\49 See Jay Chambers and William Fowler, Public School Teacher Cost
Differences Across the United States, Analysis/Methodology Report No.
95-758 (Washington, D.C.: National Center for Education Statistics,
Oct. 1995).
\50 Data on district level teacher costs were not available in Oregon
for school districts comprising about 27 percent of students enrolled
in the state; cost data were also not available for districts
enrolling about 15 percent of students in Alaska, about 8 percent of
students in New York, and about 6 percent in Maryland and New Jersey.
Since our analysis was completed, we have been informed by NCES that
data correcting these problems are now, or will soon be, available.
\51 Stephen M. Barro, Cost-of-Education Differentials Across the
States, Working Paper No. 94-05, (Washington, D.C.: National Center
for Education Statistics, July 1994).
INCLUSION OF A BONUS FOR
IMPROVING EQUITY OVER TIME
------------------------------------------------------ Appendix IV:2.4
In two of the five alternative measures of equity we developed,
options D and F, states are rewarded solely on the basis of their
current level of equity, as is the case under current law. Another
way to compare states is to include a measure of whether and to what
extent states have improved their level of equity in recent years.
Our three other alternative measures of equity, options E, G, and H,
take into account rate of change in the level of equity from school
year 1989-90 to school year 1991-92, the most recent comprehensive
data available on school district level finances.
Our first alternative, option D, uses a COV, as the current title I
equity measure does. In addition, option D takes into account the
needs of students with limited English proficiency or disabilities as
well as those who are poor. We used weights of 1.2 for poor students
and those with limited English proficiency, and 2.3 for students with
disabilities, rather than the weighting of 1.4 for only poor students
as in the current measure of equity. We also used a teacher cost
index to adjust for purchasing power differences across school
districts. Like the current measure, we subtracted this adjusted COV
from 1.3 so that the two measures are comparable. We limited this
measure, along with three of the other four we developed, so that no
state's measure is less than 0.95 or more than 1.3. (The current
equity factor presently ranges from 0.99 to 1.3; see table IV.1.)
This limitation affected few states and resulted in relatively minor
changes.
Option E is a variation of option D that takes into account a state's
improvement in equity over time, from school year 1989-90 to school
year 1991-92. First we calculated a factor similar to option D,
using 1989-90 data, and determined the rate of change from 1989-90 to
1991-92. We then multiplied option D by 1 plus the rate of change.
Thus, for example, if the state's equity factor improved by 3 percent
over that time, option E would yield a 3- percent increase in the
factor over option D. If a state's level of equity decreased by 3
percent, option E would be 3 percent lower than option D. For
example, if a state currently had an equity factor of 1.10 under
option D and it improved by 3 percent, option E would be equal to
1.10 multiplied by 1.03, or 1.13. If, instead, it decreased by 3
percent over this time, option E would be equal to 1.10 multiplied by
0.97, or 1.07.
We also developed two options, F and G, based on the McLoone Index,
which measures the extent to which the state brings up the
expenditures of those districts spending below the median. Option F
is based on the McLoone Index and includes the same adjustments for
differences in student needs and purchasing power across a state's
school districts. Under the current law, states that are fully
equalized receive a score of 1.30; states that are fully equalized
using the McLoone Index receive a score of 1.00. To ensure that the
equity factors we developed were comparable with the current equity
factor, option F is equal to the adjusted McLoone Index plus 0.30.
(The 0.30 is determined by subtracting 1.00 from 1.30.) In this way,
states that are fully equalized would receive a score of 1.30, just
as they currently do under the existing equity factor.
We also developed an equity factor based on the McLoone Index that
takes into account the rate of change in equity over time, which we
refer to as option G. We used the option F method to calculate
indexes using data for school years 1989-90 and 1991-92. Again,
states that increase their level of equity over time receive an
increase in their equity factor under option G, while those whose
level of equity declines receive a lower score under option G than
option F.
As noted earlier, options D through G include limits such that the
equity factor cannot drop below 0.95 or rise above 1.30; the use of
these limits affected few states. We also developed option H, which
is identical to option E except that it uses no limits. As shown in
table IV.1, the equity factors for options E and H are identical
except for four states: California, Louisiana, New York, and Rhode
Island.
Table IV.1
Current and Alternative Title I Equity
Factors
Current
equity Optio Optio Optio Optio Optio
State factor\a n D n E n F n G n H
-------------- ---------- ----- ----- ----- ----- -----
Alabama 1.17 1.19 1.18 1.21 1.20 1.18
Alaska 1.20\b 0.99 1.03 1.02 0.95 1.03
Arizona 1.20 1.04 1.08 1.12 1.17 1.08
Arkansas 1.17 1.13 1.12 1.19 1.18 1.12
California 1.18 0.95 0.95 1.07 1.04 0.83
Colorado 1.18 1.12 1.11 1.18 1.18 1.11
Connecticut 1.16 1.02 1.01 1.05 1.06 1.01
Delaware 1.23 1.22 1.25 1.24 1.26 1.25
District of 1.30 1.30 1.30 1.30 1.30 1.30
Columbia\c
Florida 1.22 1.09 1.10 1.12 1.12 1.10
Georgia 1.14 1.16 1.15 1.19 1.19 1.15
Hawaii\c 1.30 1.30 1.30 1.30 1.30 1.30
Idaho 1.17 1.07 1.07 1.16 1.13 1.07
Illinois 1.11 1.00 1.00 1.24 1.23 1.00
Indiana 1.18 1.15 1.16 1.19 1.19 1.16
Iowa 1.22 1.15 1.15 1.22 1.23 1.15
Kansas 1.20\b 1.05 1.06 1.21 1.24 1.06
Kentucky 1.19 1.18 1.20 1.18 1.16 1.20
Louisiana 1.18 1.04 0.95 1.21 1.20 0.92
Maine 1.19 1.14 1.14 1.19 1.19 1.14
Maryland 1.14 1.20 1.22 1.19 1.19 1.22
Massachusetts 1.07 0.97 0.97 1.08 1.08 0.97
Michigan 1.09 1.14 1.16 1.21 1.20 1.16
Minnesota 1.17 1.12 1.13 1.21 1.27 1.13
Mississippi 1.18 1.17 1.17 1.22 1.21 1.17
Missouri 0.99 1.02 0.98 1.18 1.19 0.98
Montana 1.14 1.10 1.10 1.21 1.22 1.10
Nebraska 1.17 1.09 1.10 1.14 1.11 1.10
Nevada 1.21 1.16 1.18 1.30 1.30 1.18
New Hampshire 1.14 1.10 1.10 1.16 1.17 1.10
New Jersey 1.14 0.97 0.99 1.02 1.03 0.99
New Mexico 1.20\b 1.01 1.01 1.06 1.02 1.01
New York 1.05 1.05 1.16 1.18 1.06 1.26
North Carolina 1.21 1.20 1.21 1.23 1.25 1.21
North Dakota 1.17 1.11 1.12 1.21 1.22 1.12
Ohio 1.07 1.09 1.08 1.16 1.14 1.08
Oklahoma 1.20 1.11 1.12 1.21 1.19 1.12
Oregon 1.16 1.09 1.09 1.15 1.15 1.09
Pennsylvania 1.10 1.09 1.10 1.16 1.15 1.10
Puerto Rico\c 1.30 1.30 1.30 1.30 1.30 1.30
Rhode Island 1.19 0.95 0.95 0.95 0.95 0.92
South Carolina 1.19 1.19 1.19 1.23 1.22 1.19
South Dakota 1.16 1.12 1.16 1.23 1.25 1.16
Tennessee 1.15 1.16 1.16 1.19 1.20 1.16
Texas 1.18 0.95 0.95 1.06 1.09 0.95
Utah 1.19 1.19 1.19 1.24 1.21 1.19
Vermont 1.12 1.10 1.10 1.18 1.21 1.10
Virginia 1.08 1.17 1.16 1.25 1.27 1.16
Washington 1.22 1.08 1.09 1.15 1.14 1.09
West Virginia 1.24 1.23 1.24 1.26 1.29 1.24
Wisconsin 1.18 1.13 1.13 1.18 1.17 1.13
Wyoming 1.15 1.12 1.12 1.25 1.28 1.12
-------------------------------------------------------------
\a Liane White, Education Finance Incentive Grant Under ESEA Title I,
Report for Congress No. 95-963 EPW (Washington, D.C.: Congressional
Research Service, Sept. 12, 1995).
\b The current law provides that those states that qualify as
equitable under the Impact Aid program will receive a minimum equity
factor of 1.20. Without this 1.20 minimum, the equity factors would
be 0.95 for Alaska, 1.16 for Kansas, and 1.16 for New Mexico.
\c The current equity factor for the District of Columbia, Hawaii,
and Puerto Rico is 1.30 because they each have only one school
district and, therefore, no spending disparities across school
districts. For the same reason, these states receive a score of 1.30
under the alternatives we developed.
CORRELATIONS BETWEEN EQUITY AND
DEMOGRAPHIC CHARACTERISTICS
-------------------------------------------------------- Appendix IV:3
We examined the correlations between equity and states' demographic
characteristics, such as state median income, within-state deviations
in median household income, state percentage of school-age children
in poverty, and within-state deviations in percentage of school-age
children in poverty. We examined the extent to which title I's
current equity factor and the alternative equity factors we developed
were related to these demographic factors.\52
When we examined the correlation between the various equity factors
and variability in median household income across a state's
districts, we found that the variable "within state deviations in
median household income" was negatively correlated with each equity
factor.\53 That is, the lower the variability in median household
income across districts in the state, the higher the equity factor,
and vice versa. The association was strongest, however, for the
current equity factor and weakest for the three alternatives that
consider improvement in equity over time--options E, G, and H (see
table IV.2).
Table IV.2
Correlations Between the Equity Factors
and State Demographic Characteristics
Optio Optio Optio Optio
Characteristi Current n n n n Option
c measure D E F G H
------------- ------- ----- ----- ----- ----- ------
State median -0.11 - - - - -0.05
household 0.03 0.04 0.09 0.01
income
Within-state - - - - - -
deviations 0.57\a 0.50\ 0.43\ 0.48\ 0.45\ 0.36\a
in median a a a a
household
income
Percentage of 0.13 0.07 - 0.13 0.08 -0.02
school-age 0.01
children in
poverty
Within-state - - - - - -
deviations 0.37\a 0.44\ 0.40\ 0.42\ 0.37\ 0.35\a
in a a a a
percentage
of school-
age children
in poverty
----------------------------------------------------------
\a Significant at the 0.05 level.
Each of the equity factors we developed, however, as well as the
current one, was negatively correlated with the variable "within-
state deviations in percentage of school-age children in poverty." We
found no correlation between a state's percentage of school-age
children in poverty and the five alternative equity factors we
developed; similarly, there was no correlation for the current equity
factor. In addition, there was no significant correlation between a
state's median household income and the current or alternative equity
factors.
--------------------
\52 We did not include the District of Columbia, Hawaii, or Puerto
Rico in any of the correlations, as they each comprise only one
school district.
\53 Two variables, state median household income and within-state
deviations in median household income, were adjusted for differences
in purchasing power.
ILLUSTRATIVE ALTERNATIVE
ALLOCATIONS USING NEW MEASURES
=========================================================== Appendix V
To show how improved measures might be used, this appendix provides
illustrative allocations using the current formula for the Education
Finance Incentive Program along with alternative formulas we
developed; we also provide a state-by-state estimate of allocations
under the targeted grant formula for comparison purposes. While NCES
data are sufficient for the purposes of illustrating the general
effects of changes in the effort and equity factors, NCES may want to
address a number of cases of missing data or irregularities in the
school district databases.
As noted earlier, if funds were appropriated for the Education
Finance Incentive Program, each state's grant would be determined by
the formula in figure V.1.
Figure V.1: Formula for
Determining State Share
(See figure in printed
edition.)
We noted earlier that two of the effort factor options we developed,
options B and C, have several benefits compared with the current
effort factor. For example, options B and C (1) are more
comprehensive, (2) are not biased against states with high
proportions of school-age children, and (3) include an incentive for
states that improve their level of effort over time. Equity options
E, G, and H also have several benefits compared with the current
equity factor: (1) they consider the additional education costs
related to numbers of students with limited English proficiency or
disabilities, in addition to poor students; (2) they consider
differences in purchasing power across districts; and (3) they
include a bonus for states that become more equitable over time.
Preferences for options E or H versus option G depend on interest in
measuring variation in spending levels for all school districts in
the state (options E or H) or focusing on a state's ability to bring
low-spending school districts up to the median (option G).
Figure V.2 shows the three illustrative alternative formulas based,
in part, on the alternative effort and equity factors we developed.
Figure V.2: Examples of
Alternative Formulas
(See figure in printed
edition.)
Table V.I
Illustrative State Allocations Under
Current and Alternative Title I
Formulas--Education Finance Incentive
Program and Targeted Grants
(Dollars in Thousands)
Current Targeted
State formula 1 2 3\a 4\a grants\b
--------------- ----------- ----------- ----------- ----------- ----------- -----------
Alabama 3,099 3,275 3,135 4,525 4,306 3,498
Alaska 612 529 500 500 500 458
Arizona 3,225 3,344 3,437 3,492 3,439 3,210
Arkansas 1,850 1,993 1,986 2,741 2,753 2,146
California 23,378 19,839 20,539 19,728 17,177 26,812
Colorado 2,808 2,860 2,869 2,104 2,135 1,706
Connecticut 2,428 2,055 2,022 1,182 1,199 1,404
Delaware 573 554 527 500 500 476
District of 500 500 500 545 500 552
Columbia
Florida 10,034 9,030 8,668 8,311 8,260 8,044
Georgia 5,174 5,937 5,781 6,219 6,184 4,414
Hawaii 922 981 925 609 579 500
Idaho 996 985 982 766 777 500
Illinois 8,261 7,726 8,982 7,627 7,579 10,766
Indiana 4,565 4,897 4,734 3,806 3,968 2,410
Iowa 2,309 2,450 2,480 1,858 1,901 1,012
Kansas 2,082 2,090 2,301 1,597 1,588 1,092
Kentucky 2,953 3,143 2,850 4,328 4,391 3,594
Louisiana 3,715 3,047 3,637 5,500 5,269 6,242
Maine 1,014 1,029 1,011 759 849 500
Maryland 3,787 3,928 3,607 2,321 2,355 2,350
Massachusetts 4,027 3,464 3,643 2,485 2,469 3,306
Michigan 7,426 8,341 8,128 8,325 9,237 8,964
Minnesota 3,824 4,045 4,264 2,671 2,631 1,914
Mississippi 2,205 2,286 2,243 4,413 4,292 4,268
Missouri 3,377 3,505 4,022 3,286 3,300 3,168
Montana 763 742 772 746 817 580
Nebraska 1,289 1,419 1,344 1,044 1,039 664
Nevada 1,069 1,099 1,144 597 562 500
New Hampshire 819 878 877 500 500 492
New Jersey 5,771 5,222 5,133 3,224 3,394 3,648
New Mexico 1,457 1,395 1,332 1,976 1,965 1,812
New York 12,276 14,243 12,268 16,092 17,645 21,120
North Carolina 5,097 5,388 5,232 4,794 4,561 2,994
North Dakota 512 513 530 500 500 500
Ohio 8,296 8,817 8,772 8,513 8,876 7,896
Oklahoma 2,607 2,850 2,847 3,189 3,142 2,186
Oregon 2,503 2,465 2,447 1,833 1,946 1,444
Pennsylvania 8,658 9,025 8,928 7,816 7,775 8,306
Puerto Rico 3,757 3,959 3,732 8,000 8,000 6,928
Rhode Island 752 633 597 500 500 534
South Carolina 2,830 3,066 2,968 3,609 3,589 2,386
South Dakota 608 653 663 702 705 518
Tennessee 3,618 4,026 3,907 4,567 3,984 3,140
Texas 14,941 13,372 14,505 17,091 17,194 19,694
Utah 1,985 2,296 2,204 1,393 1,439 810
Vermont 500 500 500 500 500 460
Virginia 4,167 4,705 4,832 3,339 3,388 2,320
Washington 4,184 4,379 4,302 3,032 3,016 2,412
West Virginia 1,491 1,573 1,539 2,341 2,709 1,810
Wisconsin 4,406 4,449 4,354 3,404 3,614 3,054
Wyoming 500 500 500 500 500 482
---------------------------------------------------------------------------------------------
\a For alternatives 3 and 4, Puerto Rico's allocation has been set to
a maximum of $8 million.
\b These estimates of allocations for targeted grants are based on
estimates in Riddle, Education for the Disadvantaged, pp. 19-22.
While we adjusted the estimates to account for allocating $200
million rather than $100 million, we have not updated these
estimates, for example, to take into account state per pupil spending
data that have become available in the past year. However, we do not
anticipate major shifts in allocations from one year to the next.
IMPLICATIONS FOR IMPACT AID
========================================================== Appendix VI
In addition to an overview of the Impact Aid program and the way that
equalization is defined for this program, this appendix discusses
some of the strengths and weaknesses of that definition.
OVERVIEW OF THE IMPACT AID
PROGRAM
-------------------------------------------------------- Appendix VI:1
The Impact Aid program is intended to compensate school districts for
either a loss of tax revenues, because federal property is tax
exempt, or increased expenditures because of federal activity, for
example, the cost of educating the children of military personnel.
Under the Impact Aid program, if a state meets a certain equalization
level, it may reduce state aid payments to offset the Impact Aid
received by school districts.\54 In this way, these states can ensure
that Impact Aid funds will not contribute to creating greater
inequalities within the state. If the state does not pass Impact
Aid's test of equalization, it may not consider federal Impact Aid
payments to its localities in determining state aid (which would be
likely to result in decreasing state aid to those districts) because
the Impact Aid payments are meant for localities, not states.
--------------------
\54 However, states that qualify as equalized can only consider the
Impact Aid as local revenues "in proportion to the share that local
revenues covered under a state equalization program are of total
local revenues."
WAYS TO DETERMINE EQUALIZATION
-------------------------------------------------------- Appendix VI:2
Prior to the reauthorization of the Impact Aid program in 1994, the
Department of Education asked education finance experts to examine
the way that equalization is determined in the program.\55
These experts suggested, among other things, two improvements: (1)
use measures of spending disparities that are more comprehensive than
the Federal Range Ratio, and (2) consider differences in purchasing
power and student needs more systematically. Both issues are still
relevant.
The measure currently used to determine the level of equalization in
per pupil spending in the Impact Aid program, the Federal Range
Ratio, is the difference in per pupil expenditures between two
districts--a high-spending district (95th percentile) and a
low-spending district (5th percentile)--divided by the per pupil
expenditure of the low-spending district. One major drawback of this
measure is that it does not consider spending in a majority of each
state's school districts; it only considers spending in two of the
state's school districts (at the 95th and 5th percentiles).
Consequently, two states with fairly different spending patterns may
have similar Federal Range Ratio scores. For example, one state may
have per pupil spending clustered around the average spending level
with little variation between the two extremes of the 95th and 5th
percentiles, while another state may have per pupil spending that
varies greatly between these two points.
Also, the Impact Aid program's system for determining a state's level
of equalization does not consider the additional funds states may
provide to high-need areas. For example, some states provide
additional funds to take into account the greater needs of some types
of students (such as those who are poor or have limited English
proficiency or disabilities) or some types of districts (such as
those in sparsely populated areas). On the one hand, not including
such funds is a strength of the way that equalization is measured
because states that contribute additional funds to high-need areas
are not penalized for these greater spending disparities. On the
other hand, the overall method of determining equalization is
weakened by not considering such funds (and making related
adjustments) because it may not adequately take into account the
circumstances of districts in high-need areas.
A number of implications arise from an analysis of the current method
for determining equalization under the Impact Aid program. First,
the measure of spending disparities may be misleading. Second, the
way in which states treat districts in high-need areas is not fully
addressed. Third, few states qualify as equalized under this
measure. Fourth, to the extent that Impact Aid may allow some
districts to be compensated twice for the "impact" of a federal
presence--once by the federal government and once by the state
government--it may actually contribute to creating less, rather than
more, equalization in the state.\56
(See figure in printed edition.)Appendix VII
--------------------
\55 These experts included Robert Berne and Leanna Stiefel, K.
Forbis Jordan, Allan Odden, and Richard Salmon.
\56 See James Fox, "The Equity and Efficiency of Intergovernmental
Grants: The Case of State Equalization Plans and Impact Aid," paper
prepared for the Annual Conference of the American Education Finance
Association (Salt Lake City, Utah: Mar. 20-23, 1996), p. 2.
COMMENTS FROM THE DEPARTMENT OF
EDUCATION
========================================================== Appendix VI
GAO CONTACTS AND ACKNOWLEDGMENTS
======================================================== Appendix VIII
GAO CONTACTS
Eleanor Johnson, Assistant Director, (202) 512-7209
Ellen Kehoe Schwartz, Evaluator-in-Charge, (202) 512-7052
Veronica Scott, Evaluator, (202) 512-7039
STAFF ACKNOWLEDGMENTS
Robert Dinkelmeyer, Senior Analyst; Wayne Dow, Supervisory Operations
Research Analyst; Jerry Fastrup, Supervisory Economist; and Deborah
McCormick, Senior Social Science Analyst, provided technical advice
regarding statistical and school finance issues.
*** End of document. ***