Skip Navigation

Nicola A. Alexander
University at Albany

About the Author

Nicola A. Alexander is a doctoral student in the public administration and policy Ph.D. program at the Nelson A. Rockefeller Graduate School of Public Affairs and Policy at the University at Albany. She is interested in public finance, especially as it relates to elementary and secondary schools and resource allocation issues.

Ms. Alexander is a member of the American Education Finance Association, the Association for Public Policy Analysis and Management, the American Society of Public Administration, and the National Tax Association.

The Growth of Education Revenues
Between 1982-83 to 1991-92:
What Accounts for Differences Among States?
Nicola A. Alexander
University at Albany

Public elementary and secondary schools have experienced a tumultuous period since 1983, when A Nation at Risk was published. In response to legal challenges and criticisms raised by that report, states embarked on a substantial effort to improve their schools. This study focuses on one aspect of how states responded to the demand for better education systems: How much additional money was raised for schools. This focus does not imply that money is more important than the operation of schools. However, many studies have focused on the non-monetary aspects of school improvement efforts over the last decade, with relatively little research being devoted to school finances. This study examines revenue for schools that is provided by state and local governments, and which accounts for approximately 93 percent of total education revenue. It does not consider federal aid, which contributes approximately 7 percent.

NOTE: This research is a modification and extension of article: The Growth of School Spending During the Past Decade: What Accounts for Differences Among States? which was a preliminary examination of three major factors independently influencing education revenues. This research also analyzes how these influences jointly affect the level of revenue increases from 1982-83 to 1991-92.


Between school year 1982-83 to school year 1991-92, state and local education revenues increased 96 percent. Most of this increase was due to inflation. After removing the effects of price increases, education revenues rose 39.1 percent.¹ On a per-pupil basis, revenues increased 27.8 percent in real terms.

All states did not increase education revenues by the same amount. In fact, in seven states real per-pupil revenue rose more than 50 percent, while it fell in eight states.² The main objective of this paper is to analyze some of the most important factors related to the following differences:

Description of the Data Used in this Analysis

The data used in this analysis come from the National Center for Education Statistics (NCES) Common Core of Data (CCD) collection of public education revenues for the years 1982-83 through 1991-92. CCD is an annual collection of basic information on the population of public elementary and secondary schools in the United States derived from the administrative and fiscal records of state departments of education. Each year, states report to NCES the revenues their local education agencies receive from local, intermediate, state, and federal sources. As indicated, this analysis focuses only on state and local government education revenues.

These data have both advantages and disadvantages. A major advantage is that the information is subject to rigorous scrutiny by NCES and therefore, are highly reliable. However, the reliability of the data comes at a cost. CCD data tend to be less current than data from other sources. Another shortcoming of the data is their imputations of state contributions to teacher pensions, which are a major source of state support for education.

We adjusted the CCD state and local revenue data for years 1982-83 through 1991-92 in two ways to facilitate analysis. First, we divided the reported revenue aggregates by a measure of state student populations (1) to permit comparisons of different size states and (2) to control for fluctuations in enrollment size when measuring revenue change over time. Previous studies from the Bureau of the Census and the Council of Chief State School Officers (CCSSO), have noted variation in how states calculate and report average daily attendance (ADA). Consequently, this analysis uses NCES' annual enrollment figures by state as the student population measure.

Second, we adjusted the nominal figures reported by the state using the CPI index to permit analysis in constant 1991-92 dollars.³

This study relies heavily on work completed by Martin E. Orland, preliminary research done by Gold and Alexander, as well as a recent report produced by John Augenblick et al. for the Education Commission of the States. As with the present study, Orland also uses NCES data, but he adjusts them in an unconventional way. Instead of using a single national price adjustor, Orland made adjustments individually for each state using a two-step procedure. First, he adjusted for the varied rates of wage and salary change occurring in the different states. He used the average wage data of each state to convert 77.4 percent of the reported public education revenues--the percentage representing the proportion of education costs going to wages and salaries in 1985-86. Second, he adjusted for inflationary price increases in areas other than wages and salaries by converting the remaining 22.6 percent of state education revenues with a national constant price adjustor.

Both Gold et al. and Augenblick used expenditure data reported by the National Education Association (NEA). Augenblick's analysis adjusts these data in two ways. The first method is a more conventional adjustment; it involves deflating revenue to eliminate the effects of inflation. The second method attempts to account for differences among states in the cost of education. For this purpose, Augenblick used an index developed by Howard Nelson of the American Federation of Teachers (AFT). The Nelson index estimated, for example, that the price level in Alabama was 10 percent below the national average; therefore, to make its spending per-pupil comparable to that in other states, Augenblick multiplied the state's actual spending per-pupil by 1.11. On the other hand, Connecticut's price level was estimated to be 27 percent above average, so its spending per-pupil was multiplied by 0.79.

The Nelson (and other cost-of-living) adjustments are irrelevant for most of the questions discussed in this analysis because they do not affect the percentage increase in per-pupil revenues. The adjustments do, however, affect consideration of the relationship between the level of revenue in 1982-83 and increases that occurred during the next 9 years. Table 1 shows the adjusted and unadjusted levels of per-pupil revenues in 1982-83. One of the biggest differences is in Connecticut, which ranked fifth in unadjusted revenues but ranked only 26th based on the Nelson index. Hawaii, Massachusetts, New Jersey, and Rhode Island had a ranking that was at least 10 places lower after revenue was adjusted. There are no states with large pre- and post-adjustment differences that had higher rankings using the Nelson adjustments.

The numbers reported here differ somewhat from other published data on this topic (e.g., NCES, 1988; and Education Commission of the States, 1993). The source of these differences will vary by study, but are generally caused by one or more of the following factors:

Table 1.Real per-pupil school revenues actual and adjusted for differences in price levels among states, 1982-83

                      Unadjusted        Adjusted          Difference Change in
                      Per-Pupil         Per-pupil         in dollar  Rank
State                          Rank              Rank     Unadj.-Adj Unadj.-Adj.
   National average  $3,896            $3,896             
Connecticut           4,828     5       3,793    29       1,035         -24
Hawaii                3,806    25       2,997    43         809         -18
Massachusetts         4,779     7       3,775    30       1,004         -23
New Jersey            5,497     4       4,251    19       1,246         -15
Rhode Island          4,543    10       4,111    21         432         -11

NOTE: Price level adjustment is based on index developed by Howard Nelson, 1989.

Revenue Increases and Changes of Enrollment

Table 2 shows the growth of inflation-adjusted education revenues. Three measures have to be considered to obtain a complete picture of how education revenues increased: total and per-pupil revenues and changes in enrollment. In West Virginia, for example, total education revenues rose considerably slower than average, but per-pupil revenues rose 51.8 percent, the seventh fastest increase in the nation. This is because West Virginia's enrollment actually fell 14.3 percent, the biggest decline in the United States. Consequently, though the growth of total revenues in West Virginia lagged behind the national average, this was more than offset by the substantial decrease in enrollment. On the other hand, Florida had below average per-pupil revenue increases. The fact that enrollment rose 32.5 percent in this state--the third highest in the nation--made it difficult for Florida to increase its per-pupil revenues. The states with the biggest increases in revenue per-pupil were Arkansas, Kentucky, Maine, New Jersey, and North Carolina. The states with the smallest increases were Colorado, New Mexico, North Dakota, Utah, and Wyoming.

Table 2. Growth of inflation-adjusted school revenues, total and per-pupil, 1982-83 to 1991-92

                       Total            Per-pupil
State                  Revenues Rank    revenues  Rank  Enrollment  Rank
  National average     39.12             27.78             8.87
Alabama                19.93    38       19.68    33       0.21     41
Alaska                 22.63    36       -1.68    45       24.72     5
Arizona                79.82     2       35.28    13       32.92     2
Arkansas               59.49     9       56.1      4        2.17    34
California             70.39     4       34.1     16       27.07     4
Colorado               10.29    44       -2.39    46       12.99    14
Connecticut            58.17    11       54.65     6        2.28    32
Delaware               30.43    29       14.29    36       14.13    11
Florida                68.76     5       27.38    27       32.49     3
Georgia                44.27    18       25.59    30       14.88    10
Hawaii                 47.11    15       34.5     15       9.37     19
Idaho                  46.37    16       30.37    20       12.27    16
Illinois               34.26    26       32.81    18       1.09     37
Indiana                40.41    21       43.88    11       -2.41    49
Iowa                   7.12     46       7.65     40       -0.49    43
Kansas                 18.97    40        6.77    41       11.43    17
Kentucky               65.62     6       63.69     1        1.18    36
Louisiana              7.69     45        7.99    39       -0.28    42
Maine                  63.22     7       58.17     3        3.19    30
Maryland               47.27    14       33.88    17       10.00    18
Massachusetts          24.08    35       26.81    29       -2.15    48
Michigan               20.46    37       22.89    32       -1.98    47
Minnesota              32.65    28       17.86    34       12.55    15
Mississippi            40.93    20       30.1     21        8.32    22
Missouri               39.53    22       29.15    24        8.03    23
Montana                2.88     47       -1.21    43        4.14    29
Nebraska               19.84    39       13.28    38        5.8     26
Nevada                107.27     1       39.84    12       48.21     1
New Hampshire          70.7      3       49.77     8       13.97    13
New Jersey             56.75    12       59.15     2       -1.51    45
New Mexico             13.1     43       -3.37    47       17.04     9
New York               30.32    30       29.6     23        0.56    40
North Carolina         58.53    10       55.05     5        2.25    33
North Dakota           -8.21    49       -9.39    49        1.3     35
Ohio                   27.71    33       29.91    22       -1.69    46
Oklahoma               -0.3     48       -1.26    44       0.97     39
Oregon                 30.31    31       14.21    37       14.09    12
Pennsylvania           27.37    34       28.88    25       -1.17    44
Rhode Island           34.18    27       27.29    28        5.41    27
South Carolina         51.52    13       44.61    10        4.77    28
South Dakota           14.04    42        4.28    42        9.36    20
Tennessee              34.81    25       31.05    19        2.87    31
Texas                  45.52    17       23.05    31       18.26     8
Utah                   16.78    41       -4.79    48       22.65     6
Vermont                62.11     8       48.72     9        9.01    21
Virginia               43.69    19       34.52    14       6.81     25
Washington             39.16    23       14.29    35       21.76     7
West Virginia          30.19    32       51.85     7      -14.26    50
Wisconsin              37.99    24       28.88    26        7.07    24
Wyoming               -28.00    50      -28.76    50        1.07    38

SOURCE: U.S. Department of Education, National Center for Education Statistics. Digest of Educational Statistics.

Enrollment rose in 41 states and fell in 9 states between school years 1982-83 and 1991-92. The biggest increases in enrollment were in Nevada (48.2 percent), Arizona (32.9 percent), Florida (32.5 percent), California (27.1 percent), and Alaska (24.7 percent). The largest decreases in enrollment were in West Virginia (-14.3 percent), Indiana (-2.4 percent), Massachusetts (-2.2 percent), Michigan (-2.0 percent), and Ohio (-1.7 percent).

Table 3 examines the relationship between increases in revenue per-pupil and enrollment growth. In 20 states the trade-off between enrollment increases and revenue growth was particularly marked. Eleven of these states had high enrollment increases and low revenue growth; nine states had declines or low growth of enrollment and high revenue increases.

Table 3. Highlights of inflation-adjusted per-pupil education revenues and change of enrollment, 1982-83 to 1991-92


High increase in per-pupil revenue and decrease or small increase in enrollment (9states)
Arkansas        Connecticut             Illinois
Indiana         Kentucky                New Jersey
North Carolina  Tennessee               West Virginia
Low increase in per-pupil revenue and high increase in enrollment (11 states)
Alaska          Colorado                Delaware                Kansas
Minnesota       New Mexico              Oregon                  South Dakota
Texas           Utah                    Washington
High per-pupil revenue increase despite high enrollment increase (7 states)
Arizona         California              Hawaii
Idaho           Maryland                Nevada
New Hampshire
Low per-pupil revenue increase despite enrollment decrease or small increase (7 states)
Alabama         Iowa                    Louisiana
Michigan        North Dakota            Oklahoma

SOURCE: U.S. Department of Eduction, , Digest of Education Statistics.

The seven states with low increases in revenue despite decreases or small increases in enrollment had economies dependent on oil production, automobile manufacturing, agriculture, or industries that were depressed throughout much of the period.

Convergence of Revenue Levels

Intergovernmental competition tends to even out differences in education revenue among states. Because employers seek well-trained workers, states that have relatively weak school systems are at a significant disadvantage in attracting investment that is crucial to a prosperous economy. As a result, it would be expected that states with relatively low per-pupil revenue in 1982-83 would tend to have faster than average revenue increases over the next nine years. Moreover, because of fiscal pressures and a higher initial base, it would be expected that those states with relatively high per-pupil revenues in 1982-83 would have lower than average revenue increases over the same period.

The convergence hypothesis can be tested in two ways, using actual per-pupil revenues in 1982-83 or using a measure of revenue that is adjusted for differences in costs. Table 4 examines the relationship between 1982-83 revenues and subsequent revenue growth using revenue adjusted only for inflation. In general, table 4 supports the hypothesis that a "catch up" phenomenon occurred. Some distinct regional tendencies are apparent. Of the 12 states with relatively low education revenues in 1982-83 and subsequent large increases over the next nine years, 11 are in the Southeast or Western quadrant of the country. By contrast, all four states with large increases in per-pupil revenues despite already having high revenue in 1982-83 are in the Northeast.

Table 4. Highlights of inflation-adjusted per-pupil education revenue, 1982-83 to 1991-92, and 1982-83 revenue per-pupil

Low 1982-83 per-pupil revenue and high increase (12states)
Arkansas        Arizona         California        Idaho
Indiana         Kentucky        Maine             Nevada
North Carolina  South Carolina  Tennessee         West Virgina
High 1982-83 per-pupil revenue and low increase (12 states)
Alaska          Colorado        Delaware        Iowa
Kansas          Mississippi     Minnesota       Montana
North Dakota    Oregon          Washington      Wyoming
Low 1982-83 per-pupil revenue and low increase (5 states)
Alabama         Louisiana
South Dakota    Utah
High 1982-83 per-pupil revenue and high increase (4 states)
Connecticut     Maryland
Rhode Island    New Jersey

SOURCE: U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics.

Economic Growth and Revenue Increases

The growth of state and local education revenues depends heavily on the health of a state's economy. It would be expected that states with strong economies would provide larger increases in per-pupil revenues. In the United States, indicators of state fiscal capacity are often grounded in measures of personal income. A limitation of these measures is their inability to reflect the diversity of tax and revenue sources, as well as their failure to capture the ability of states to export taxes.4 Notwithstanding these shortcomings, measures of personal income are often more current and more readily available than other indicators. Further, states generally do not vary significantly in their rankings among the various fiscal capacity measures, except in the case of energy-rich and tourist-rich states.

Table 5 examines the relationship between increases in personal income per-pupil and school revenue per-pupil. In general, it supports the notion that a strong economy leads to increased funding for schools. The relationship between per-pupil revenue increases and the growth of personal income per-pupil is more pronounced than the relationships described above.

Table 5. Highlights of inflation-adjusted per-pupil education revenue, 1982-83 to 1991-92, and growth of real personal income per-pupil, 1982-83

High increases in personal income per-pupil and high growth in per-pupil revenue  (13 states)
Arkansas        Connecticut             Indiana         Hawaii
Kentucky        Maine                   Maryland        New Hampshire
New Jersey      North Carolina          South Carolina  Vermont
Low increases in personal income per-pupil and low growth in per-pupil revenue (12 states)
Alaska          Colorado                Kansas          Louisiana
Montana         Nebraska                New Mexico      North Dakota
Oklahoma        Oregon                  Texas           Utah
Low increases in personal income per-pupil and high increases in per-pupil revenue (4 states)
Arizona         California
Idaho           Tennessee
High increases in personal income per-pupil and low increases in per-pupil revenue (2 states)
Alabama         Michigan

SOURCE: U.S. Department of Education, National Center for Education Statitstics, Digest of Education Statistics. U.S. Department of Commerce, Bureau of Economic Analysis.

State Versus Local Revenues

The final issue considered in this study is somewhat different from those discussed previously. Enrollment changes, initial revenue levels, and personal income growth all can be viewed as determinants of how much revenue increased. In this section, we consider where the money came from--was it mainly state or local governments that provided the bulk of increased funding in the states with particularly large or small revenue increases? Growth in state funding tended to outpace that provided by local governments in those states which had the smallest increases in per-pupil revenues. The opposite occurred for those states with particularly high increases.

Revenue Increases and the Joint Impact of Enrollment, 198283 Per-Pupil Education Revenue, and Personal Income Per-Pupil

Regression analysis5 generally supports the model of education revenue increases postulated above. Controlling for regional variation, as well as the economic and demographic factors previously discussed, the model accounts for 63.4 percent6 of the variation in the growth of per-pupil education revenues from 1982-83 to 1991-92. The regression findings indicate that for each 1 percent increase in personal income per-pupil, per-pupil education revenue rises by 0.612 percent, holding all other factors constant. This estimated coefficient is significant at the 0.05 percent level.

The unadjusted 1982-83 per-pupil revenue negatively influences the subsequent growth of per-pupil school revenue from 1982-83 to 1991-92, ceteris paribus. For each 1 percent rise in the level of 1982-83 education revenue per enrolled student, per-pupil revenue decreases by 0.006 percent. This estimated coefficient is significant at the 0.05 percent level. The magnitude of the coefficient implies, however, that 1982-83 per-pupil education revenue does not have a substantive impact on the growth of education revenue per-pupil.

Surprisingly, the results do not reflect the expected negative effect of enrollment. Instead, they indicate that for each 1 percent increase in enrollment, education revenue per-pupil rises by 0.104 percent, all things equal. This estimated coefficient is not significant at the 0.05 percent level.

The results of the regression analysis support the suggestion that a pattern of regional variation in per-pupil revenue growth exists. In general, the states in the New England, Mid-Atlantic, Great Lakes, and Far West regions have a relatively higher growth of per-pupil education revenue than in the Southeast, all things equal. By contrast, the Plains, Southwestern, and Rocky Mountain states have a relatively lower growth of education revenue from 1982-83 to 1991-92 than the Southeast. The estimated coefficients of the Plains and Rocky Mountain regions are the only coefficients significant at the 0.05 percent level. See table 6 for a summary of these regression findings.

Table 6. Joint impact of personarevenue on the growth offor regional variation

Variables                               Coefficient             T-ratio
Constant                                  50.027                 4.791
Personal income per-pupil                  0.586                 2.844*
Enrollment                                 0.078                 0.377
Adjusted 1983 PPR                         -0.008                -3.625*
New England                                9.148                 1.445
Mid Atlantic                               2.698                 0.374
Great Lakes                                2.335                 0.346
Plains                                   -12.668                -2.035*
Southwest                                 -9.522                -1.223
Rocky Mountain                           -16.989                -2.258*
Far West                                  -1.531                -0.201
R-squared:                                 0.735
Adj. R-squared:                            0.666
Degrees of freedom:                        10,39
*:                                   coefficient is significant at the 0.05 percent level.

NOTE: The Southeastern states are the comparative regional base for this model. The regression results indicate, for example, that the growth of per-pupil school revenue in New England was 9.15 percentage points higher than in the Southeast, holding personal income per-pupil, enrollment, and adjusted 198283 per-pupil revenue constant. This result was not significant at the 0.05 percent level.

SOURCE: U.S. Department of Education, National Center for Education Statistics, Digest of Education Statistics. Washington, DC: Government Printing Office. U.S. Department of Commerce, Bureau of Economic Analysis.


The above analyses imply that the growth of a state's economy as measured by personal income per-pupil is closely tied to per-pupil revenue increases. A clear pattern emerges where the faster the growth of a state's economy, the more likely that state is to have per-pupil education revenues rise faster than average. While this general pattern holds for most states, key exceptions do exist. In Michigan, for example, despite the relatively high growth of personal income per-pupil, this state still had fairly low growth of per-pupil education revenue between 1982-83 to 1991-92. This reflects the particularly small increase in the contribution of the state government to education during the period studied.7

Differences in the level of per-pupil revenues in 1982-83 also seem to play a role in the variation among states in the growth of education revenues per student. The lower the education revenue was in 1982-83, the faster it tended to increase in the subsequent nine years. Distinct regional tendencies are readily apparent in this aspect of the analysis. The four states with large increases in per-pupil revenue despite already having high education revenue in 1982-83 are all located in the Northeast.

The greatest surprise of the study is the seemingly positive impact of the growth of enrollment on per-pupil education revenue increases. The direction of the enrollment effect contrasts with previous research on the growth of education revenue, as well as the expectations of this study. This unusual finding and the insignificant estimated coefficient suggest that student population would be better captured using another variable, such as the annual fall membership count.

Finally, the findings seem to indicate that those states with the highest growth of per-pupil education revenues tended to rely on increases from local sources rather than on growth of state funding. By contrast, those states with particularly small increases in per-pupil education revenue tended to receive more of their additional funding from state, rather than local, governments. However, the growth of per-pupil revenue does not denote the portion of education funding by government source. Consequently, nothing can be said of the states' existing relative reliance on a particular level of government.


Augenblick, John, et al. 1993. How much are schools spending?: A 50-state examination of expenditure patterns over the last decade. Denver, CO: Education Commission of the States.

Baily, Albertina. 1991. School finance litigation: A renewed impulse for reforms. Alexandria, VA:
National School Boards Association.

Fossett, James W., and James H. Wyckoff. 1993. Has medicaid growth crowded out state spending on education? Department of Public Administration and Policy. University at Albany.

Gold, Steven D. and Nicola A. Alexander. July 1993. The growth of school spending during the past decade: What accounts for differences among states? Center for the Study of the States. Prepared for Presentation at the Annual Summer Conference.

Nelson, F. Howard. 1991. "An interstate cost-of-living index." Education Evaluation and Policy 13: 103-11.

Orland, Martin E. 1988. Trends in real public elementary and secondary school revenues 1981-82 to 1985-86. Analysis report. Washington, DC: . ERIC 299 705.

U.S. Department of Education, . Digest of Education Statistics. Washington, DC: Government Printing Office.


1/ Inflation was measured by the Consumer Price Index (CPI), which rose 41 percent between school year 1982-83 to school year 1991-92. The CPI was chosen to facilitate comparisons with other recent studies; inflation, according to the implicit deflator for state and local governments, was slightly lower (40.5 percent).

2/ The seven states with per-pupil revenue increases greater than 50 percent were: Arkansas, Connecticut, Kentucky, Louisiana, New Jersey, North Carolina, and West Virginia. The eight states that had declines in per-pupil revenue were Alaska, Colorado, Montana, New Mexico, North Dakota, Oklahoma, Utah, and Wyoming.

3/ Many economists, including Alan Greenspan, chairman of the Federal Reserve, think that the CPI generally overstates the level of inflation that exists in an economy. Greenspan estimates that this exaggeration is anywhere between one-half and one and one-half percent. If this is true, then the real growth of per-pupil revenue and personal income per-pupil are consistently underestimated in this study.

4/ For a more detailed discussion of the limitations of using personal income and possible alternatives, see ACIR. 1993. "RTS 1991: State revenue capacity and effort." Washington, DC: Government Printing Office.

5/ The report conducted the regression analysis using the Econometrics Toolkit (ET) software.

6/ The percent of variation explained by the model increases to 66.7 percent when using adjusted per-pupil education revenues (see table 6).

7/ Michigan ranked 47th in the nation in terms of the growth of state education revenues from 1982-83 to 1991-92.

[Meeting the Challenge of Devolution: How Changing Demigraphic and Fiscal Contexts Affect State Investments in Education] Previous Table of Contents Next[A Study of Administrative Expenditures in Texas Public Schools]