3
The Impact of Court-Mandated School Finance Reform

William N. Evans, Sheila E. Murray, and Robert M. Schwab

Introduction

Through the 1960s, local governments provided the majority of funds for public primary and secondary education in the United States. Because property taxes have traditionally been the primary source of local tax revenue, the resources devoted to education were to a large extent a function of the property tax base in a community. Critics argued that this finance system was inherently unfair and, following the success of Serrano v. Priest (1971), repeatedly challenged the constitutionality of local funding plans in court.

Table 3-1 summarizes the status of finance reform litigation.1 By 1998, supreme courts in 43 states had heard cases on the constitutionality of school finance systems. The courts have overturned systems in 19 states and upheld systems in 20; cases are still pending in the remaining 4. In addition, litigation has been filed in a number of states where the state supreme court had already ruled. In California, for example, there have now been three separate decisions in the Serrano case.

The legal grounds under which school finance systems have been challenged have varied over time. Heise (1995) has defined three "waves" of education finance cases. The first-wave cases, typified by Serrano , focused on the Equal Protection Clause of the United States Constitution. In 1973, the U.S. Supreme Court ruled in San Antonio Independent School District v. Rodriguez that local finance did not violate the U.S. Constitution and thus closed off this initial line of attack. The second wave began with a 1973 New Jersey case, Robinson v. Cahill , and looked to state constitutions for relief. In part, these cases appealed to state



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 72
72 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM 3 The Impact of Court-Mandated School Finance Reform William N. Evans, Sheila E. Murray, and Robert M. Schwab INTRODUCTION Through the 1960s, local governments provided the majority of funds for public primary and secondary education in the United States. Because property taxes have traditionally been the primary source of local tax revenue, the re- sources devoted to education were to a large extent a function of the property tax base in a community. Critics argued that this finance system was inherently unfair and, following the success of Serrano v. Priest (1971), repeatedly chal- lenged the constitutionality of local funding plans in court. Table 3-1 summarizes the status of finance reform litigation.1 By 1998, supreme courts in 43 states had heard cases on the constitutionality of school finance systems. The courts have overturned systems in 19 states and upheld systems in 20; cases are still pending in the remaining 4. In addition, litigation has been filed in a number of states where the state supreme court had already ruled. In California, for example, there have now been three separate decisions in the Serrano case. The legal grounds under which school finance systems have been challenged have varied over time. Heise (1995) has defined three “waves” of education finance cases. The first-wave cases, typified by Serrano, focused on the Equal Protection Clause of the United States Constitution. In 1973, the U.S. Supreme Court ruled in San Antonio Independent School District v. Rodriguez that local finance did not violate the U.S. Constitution and thus closed off this initial line of attack. The second wave began with a 1973 New Jersey case, Robinson v. Cahill, and looked to state constitutions for relief. In part, these cases appealed to state 72

OCR for page 72
73 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB TABLE 3-1 Summary of States with Court-Ordered Reforms, 1971-97a Alabamab 1993 Arizona 1994 Arkansas 1983 California 1971 Connecticut 1977 Kansas 1976 Kentucky 1989 Massachusetts 1993 Montana 1989 New Hampshirec 1997 New Jersey 1973 Ohio 1997 Tennessee 1993 Texas 1989 Vermont 1997 Washington 1978 West Virginia 1979 Wisconsin 1976 Wyoming 1980 aAdditional post-1997 reforms are discussed by Minorini and Sugarman (see Chapters 2 and 6 in this volume). bThe 1993 Alabama decision is a lower court decision that decided the finance system is unconstitutional. The state has indicated it will not appeal. cIn 1993, the New Hampshire State Supreme Court ruled that the state has a duty to fund public education. SOURCE: Summary of Legislative and Court-Ordered Reforms; Evans, Murray, and Schwab (1997a). Journal of Public Policy Analy- sis and Management, Copyright © (1997). Reprinted by permission of John Wiley & Sons, Inc. equal protection clauses and were brought in the interest of equity for school children. In addition, some also looked to state education clauses and argued that local funding violated states’ constitutional responsibility to provide efficient and adequate education. The third wave began in 1989 and focused almost exclu- sively on education clauses in state constitutions. These provisions are often ambiguous and ambitious; the New Jersey constitution, for example, calls for a “thorough and efficient system of free public schools” (New Jersey Constitution, article 8, section 4). The courts in the third-wave cases have relied on such language to require much more sweeping reform of states’ public school systems and to take far greater control of financing issues. Minorini and Sugarman (see Chapter 2 in this volume) give a thorough account of the history of school finance litigation.

OCR for page 72
74 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM We should note that although successful litigation was the impetus for many legislative reforms, some states adopted some form of finance reform without the court’s prodding. For example, Utah adopted finance reform on its own without judicial intervention. Other states, including Michigan, adopted reforms even though state courts upheld their finance systems. The remainder of this chapter has the following organization. We first examine the effect of the courts on the distribution and level of education spend- ing per pupil. We begin by looking at the California experience in the aftermath of Serrano and then turn to broader empirical studies and studies that have developed simulation models. The next two sections extend the existing litera- ture on the impact of court-ordered reform in three directions and focus on the effect of court-mandated reform on education spending. We first consider the consequences of adjusting expenditures for the costs of inputs (such as teacher salaries). We then look at the impact of successful litigation on low-income districts—as opposed to low-spending districts—and the consequences of reform for black and white students. A final section looks at some of the other important consequences of court decisions. We begin by examining the impact of the courts on education out- comes and then turn to private contributions to public schools and the demand for private schools. Although the research on court-mandated finance reform has been extensive, clearly not all issues have been resolved, and thus we offer a brief summary and conclusions and some suggestions for future research. EVIDENCE OF THE IMPACT OF THE COURTS ON EDUCATION FINANCE Some of the earliest literature on court-mandated reform focused on the California experience. Broader econometric models and simulations models have addressed several implications of court-mandated reform including the ef- fect on the distribution of resources and outcomes.2 California Case The general consensus from the California work has been that the shift toward state financing of education has led to a significant decrease in spending on education. Silva and Sonstelie (1995) try to estimate what proportion of this decline should be attributed to Serrano and ensuing policy changes such as Propo- sition 13, and how much should be attributed to other factors such as changes in income and number of students.3 They begin by estimating the determinants of education spending using data from all of the states other than California. Using this equation, they show that prior to Serrano in 1969-70, spending in California was similar to other states during the same period after adjusting for differences in family income and the tax price of an additional dollar of education. They

OCR for page 72
75 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB found a very different story in 1989-90. Spending was significantly lower in California than they would have predicted. They conclude that roughly one-half of the decline in spending in California can be attributed to Serrano. The change in spending in California in the post-Serrano era has been dra- matic. Rubinfeld (1995) shows that in 1971-72 California spending per pupil was 98 percent of the national average and that California ranked 19th among the states; by 1991-92, California spending was only 86 percent of the national average and the state had fallen to 39th. Broader Econometric Studies It is difficult to know the extent to which the California experience can be generalized to all states or to any particular state. The California courts set a particularly strict standard in their definition of equality, requiring that the differ- ences in per-pupil spending among nearly all districts be no greater than $100 (in 1971 dollars). It is quite possible that the Serrano decision dramatically reduced public support for education—Fischel (1989), for example, argues that Serrano led to the passage of Proposition 13—but that milder reform that simply required higher state support for the poorest districts would have led to a very different reaction. Broader empirical work can address many of the shortcomings of case stud- ies. By looking at data from many states, these empirical studies allow us to look at more general responses to school finance reform efforts. Manwaring and Sheffrin (1995) use a panel data set from 1970-90 to examine the role of equaliza- tion litigation and reform in determining the level of education funding in a dynamic model. They found that on average, successful litigation or legislative education reform raises education spending significantly. The Downes and Shah (1995) analysis is similar in some ways to the Manwaring and Sheffrin (1995) model. They show that the stringency of constraints on local discretion deter- mines the effects of reforms on the level and growth of spending. Further, for any particular type of reform, the characteristics of a state’s school population deter- mine the direction and magnitude of the post-reform changes in spending. We took a different approach in Murray et al. (1998). In that paper, we looked at the impact of court-ordered reform on the distribution of spending within states as well as the average level of spending across states. Our study was based on data for the more than 10,000 unified elementary and secondary school districts over the 20-year period 1972-92.4 Table 3-2 presents some summary statistics from that paper. The first section looks at the level and sources of funding. Real resources per student grew at an average rate of 2.1 percent per year during this period.5 The distribution of resources changed substantially over these 20 years. Revenues from state sources rose very quickly during 1972- 87 and, as a consequence, the states’ share of total resources increased from 38.3 percent to 49.3 percent. Revenues from the states then grew slowly during 1987-

OCR for page 72
76 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM TABLE 3-2 Summary of Current Education Expenditures, 1972-92 Measure 1972 1977 1982 1987 1992 Funding per Student ($1992) Local 1,923 1,881 1,799 2,163 2,621 State 1,394 1,708 1,903 2,451 2,587 Federal 325 346 297 315 368 Total 3,642 3,935 3,999 4,929 5,576 Measures of Inequality 95/5 ratio 2.72 2.37 2.22 2.53 2.40 Coefficient of variation 30.8 28.1 25.6 29.6 29.9 Gini coefficient (×100) 16.3 15.0 13.8 15.8 15.5 Theil Index (×1000) 43.7 37.1 31.0 40.7 40.5 Theil Index Decomposition Within states 13.7 14.4 14.0 12.6 13.4 Between states 30.0 22.8 17.0 28.2 27.1 National 43.7 37.2 31.0 40.7 40.5 Variance Decomposition Within states 32.2 41.5 47.5 32.8 35.3 Between states 67.8 58.5 52.5 67.2 64.7 National 100.0 100.0 100.0 100.0 100.0 SOURCE: Funding per student from National Center for Education Statistics, Digest of Education Statistics, 1994 (1994b). Education expenditure inequality measures are authors’ calculations from Bureau of the Census, Census of Government School System Finance File (F-33), various years. Calculations exclude school districts from Alaska, District of Columbia, Hawaii, Montana, and Vermont. 92. Local funding increased throughout this period, including the last five years; in 1992, local governments contributed 47.0 percent of all of public education resources. The federal government played a small and shrinking role throughout 1972-92. The second section in Table 3-2 gives several measures of inequality in district spending at the national level. Each of these measures rises when spend- ing inequality rises.6 The measures we chose give a bound on the effect of court- mandated reform. The ratio of the 95th percentile in per-pupil spending to the 5th percentile in spending is a simple ranking that treats transfers to the top or bottom of the distribution the same; changes in spending in the rest of the distribution change the 95th to 5th ratio. Changes throughout the distribution of spending contribute to the values of the coefficient of variation and the Gini coefficient. The Theil coefficient gives more weight to changes in the tails of the distribution.

OCR for page 72
77 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB The Theil index is attractive in part because it is relatively easy to decompose the Theil into disparity in spending between and within states. All of the inequality measures in Table 3-2 follow a similar pattern. Spend- ing at the 95th percentile was 2.72 times higher than spending at the 5th percen- tile in 1972. This ratio then fell to 2.40 in 1992, suggesting a narrowing of the differences in spending across students. The Theil index fell during the 1970s and early 1980s, rose sharply between 1982 and 1987, and then remained roughly constant. Inequality according to all four measures was higher in 1992 than in 1982 and somewhat lower than in 1972. The next two sections of Table 3-2 break spending inequality into two com- ponents: inequality due to differences in spending within states and inequality due to differences across states. That part of Table 3-2 makes a number of interesting points. We found that between-state inequality is much larger than within-state inequality. In 1992, variation across the states represented 64.7 percent of the total variance in per-pupil spending; between-state inequality ac- counted for 66.9 percent of the national Theil index that year. Putting these trends together, Table 3-2 shows that more than 90 percent of the reduction in the Theil index during 1972-92 was due to a reduction in inequality between states; there was little change in inequality within states. This would suggest that the litigation that began with Serrano is limited in its ability to equalize the education resources available to students. We then estimated a series of econometric models to explain state-level inequality. We used two different variables to mark the timing of reform. Ini- tially, we included a simple indicator variable Court Reform that equals 1 in all years after court-ordered education finance reform, and zero otherwise. Because we suspect reform will take some time to alter inequality, we also used a second variable, Years After Court Reform, which equals the number of years since the state supreme court overturned a finance system. Thus, for example, this variable always equals 0 in those states without successful litigation. We came to three main conclusions. First, court-mandated education fi- nance reform can decrease within-state inequality significantly. Depending on the way we measure inequality, our results imply that reform in the wake of a court decision reduces spending inequality within a state by anywhere from 19 to 34 percent. Second, our results suggest that court-ordered reform reduces in- equality by raising spending at the bottom of the distribution while leaving spend- ing at the top unchanged. As a result of court-ordered reform, we found that spending would rise by 11 percent in the poorest school districts, rise by 7 percent in the median district, and remain roughly constant in the wealthiest districts. Third, finance reform leads states to increase spending for education and leave spending in other areas unchanged, and thus, by implication, states fund the additional spending on education through higher taxes. As a consequence, the state’s share of total spending rises as a result of court-ordered reform. Using our estimates of the reduction in inequality due to court-mandated

OCR for page 72
78 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM reform and the average within state inequality presented in Table 3-2, we can also consider what would have happened to inequality in the absence of court-man- dated reform. Our estimates imply that inequality as measured by the Theil index is 34 percent lower in reform states than in nonreform states. For the reform states we constructed a predicted Theil for each state by combining the actual Theil and the inequality lost due to reform. For the nonreform states, our pre- dicted Theil remained the same as the actual. We then recalculated the weighted average of our within-state inequality measure. Without a court mandate, aver- age within-state inequality would increase 9 percent (from 13.4 to 14.6) in 1992. Thus, while it is true that within-state inequality was essentially unchanged be- tween 1972 and 1992, our estimates suggest that it would have risen sharply in the absence of court-ordered finance reform. In Evans et al. (1997a) we tried to separate the responses of state and local governments to court mandates. We began by identifying the district at the 5th, 25th, 50th, 75th, and 95th percentile of the distribution of local resources in each state. We then used a district’s per-pupil local, state, and total revenues as the dependent variable in a series of 30 econometric models (5 points in the distribu- tion × 3 sources of funds × 2 reform variables). Table 3-3 presents some of our results from that paper. We find in Table 3- 3 that, following court-mandated reform, total revenues rose significantly in dis- tricts with the lowest local revenues. All of this increase represents additional funds from the state government; we find some mixed evidence that some of this additional state money replaces local revenue. We also find that revenues in the districts with highest local revenues are essentially unchanged by court-ordered reform. Consider first the results based on Court Reform (the middle three columns in Table 3-3). Following reform, real revenue per student rose by $560 for the district at the 5th percentile of the distribution (16 percent of the mean expendi- tures for that district) and by $500 (13 percent) for the district at the 25th percen- tile. Local revenue fell in those districts, but the effect is not significantly differ- ent from zero, and thus all of the increase in revenues represents additional funds from the states. Estimated changes in revenues at the 75th and 95th percentile are much smaller than at the bottom of the distribution and are statistically insignifi- cant. The results based on Years After Court Reform (the last three columns of Table 3-3) are similar, though there are some interesting differences. We again find large and significant increases in revenue in the lowest-revenue districts and small and insignificant changes in the highest-revenue districts. We also find that, in the districts at the 5th and 25th percentile of the distribution, increases in state revenue in part offset decreases in local revenue. Our results suggest, for example, that successful litigation will lead a state government to provide the lowest-revenue districts additional state aid of $700 per student 10 years after reform. These districts reduced local revenue by $190, and thus total revenue

OCR for page 72
TABLE 3-3 Impact of Court-Mandated Finance Reform on the Distribution of Education Resources, 1972-92 Coefficient {Standard Error} Coefficient {Standard Error} on Court Reform on Years after Court Reform Percentile of Per-Pupil Means {Standard Deviations} Dependent Variable Dependent Variable Revenues from Local Per-Pupil Revenues ($1992) Per-Pupil Revenues ($1992) Per-Pupil Revenues ($1992) Sources by Source by Source by Source Local State Total Local State Total Local State Total 5th 1,074 2,510 3,584 –152.5 712.45 559.95 –19.40 70.06 50.66 {584} {1,092} {1,173} {83.95} {204.7} {192.5} {6.93} {16.88} {16.02} 25th 1,521 2,209 3,730 –99.93 600.34 500.41 –21.07 101.50 80.44 {741} {907} {1,106} {95.41} {178.6} {193.2} {7.84} {13.39} {15.31} 50th 1,913 2,063 3,976 5.96 310.72 316.68 –8.763 45.71 36.95 {907} {815} {1,106} {120.2} {155.0} {182.1} {10.03} {12.64} {15.09} 75th 2,446 1,833 4,279 115.69 128.24 243.93 3.326 21.62 24.94 {1,195} {836} {1,232} {170.7} {163.2} {184.1} {14.29} {13.56} {15.35} 95th 3,396 1,622 5,018 136.56 135.38 271.94 –9.407 26.99 17.58 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB {1,741} {817} {1,596} {266} {165.3} {271.5} {22.24} {13.70} {22.72} SOURCE: Impact of Court-Mandated Finance Reform on the Distribution of Education Resources, 1972-1992; Evans, Murray, and Schwab (1997a). Journal of Public Policy Analysis and Management, Copyright (1997). Reprinted by permission of John Wiley & Sons, Inc. 79

OCR for page 72
80 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM rose by $510. Court-ordered reform, according to these estimates, is in part a way of providing tax relief to the lowest-spending school districts in the state. Card and Payne (1997) also consider the possibility that increases in state funding for education were offset by decreases in local funding. They find that this fiscal substitution effect was small, though their estimate is imprecise and they thus cannot rule out the possibility that some of the equalizing effect of state finance reform is “undone” by changes in local revenue. Hoxby (1996) offers a very different view of the impact of court-ordered school finance reform. There are two key elements to Hoxby’s approach. First, she treats school finance reform not as an event but rather as a change in the structure of school state aid. These changes in state aid change the incentives facing local schools districts. This framework allows Hoxby to exploit differ- ences in the “price” of a $1 of education across states, across districts within states, and across time to estimate the effects of different types of reform. By price, Hoxby means the cost to the residents if they choose to increase total education spending (inclusive of state and federal aid) by $1. The econometric part of the paper then estimates total spending as a function of price, income, and demographic variables. Second, in Hoxby’s paper school districts can increase spending by raising either the property tax rate or the property tax base. Changes in the structure of state aid change the optimal tax base and rate. Hoxby’s results are provocative. The paper argues that near-equality of per- pupil resources cannot be achieved without substantial decreases in the average level of per-pupil resources. In fact, it finds that districts with low income or low property value end up with lower per-pupil spending under equalization schemes that achieve near equality. These results follow directly from her estimates of the impact of price on school spending. Strong equalization plans raise the price of education for most districts and as a consequence local spending falls sharply; in many cases she finds that lower local spending more than offsets potentially higher state aid. We have a number of concerns about this paper, in part because Hoxby’s conclusion does not seem to be consistent with the spending patterns that have emerged in the aftermath of reform in most states. We discussed the evidence on this point above, but we can make this point in a somewhat more straightforward way. There were court decisions in 12 states between 1971 and 1992 (the last year of our study). Consider a simple “difference in difference calculation” where we compare the growth in expenditures per student after reform in a state to the growth in expenditures per student in the entire country. For example, the Connecticut decision was in 1977. From 1977 through 1992, expenditures per student in Connecticut rose 9.3 percent per year. For the nation as a whole, expenditures rose 7.5 percent per year during that period, and thus Connecticut expenditures rose 1.9 percentage points faster than the U.S. average following school finance litigation. If Hoxby’s argument is correct, then we would expect Connecticut to be an outlier, or at the minimum we would expect to find a number

OCR for page 72
81 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB of states where expenditures grew slowly or fell after a court decision. But Connecticut is not an outlier. In 10 of the 12 states we examined, expenditures rose at least as fast as the U.S. average in the aftermath of school finance litiga- tion; California and Wyoming are the two exceptions to this rule. We would also argue that Hoxby’s characterization of reform in terms of the price of education is not completely convincing. While conceptually this frame- work might make sense, it is not at all clear that this sort of model could capture the impact of court-ordered reform. In general, these models rarely describe the effects of aid from higher levels of government very well. For example, this model would predict that a $1 lump-sum grant from the state would have the same effect on spending as would a $1 increase in income; in fact, grants are almost invariably found to have a much larger effect on spending than the theory would predict (the flypaper effect). Reform does not always fit neatly into the economist’s view of the world. The paper’s treatment of the California experi- ence illustrates this problem. Hoxby (1996:20) assigns California the state re- quired tax rate on the grounds that otherwise the marginal price of education would be $1. But in fact, given the rigidities of local school finance in post- Serrano, post-Proposition 13 California, the marginal price of education to a local district is probably infinite; districts have essentially lost all control over spending. We are also concerned about a second conceptual issue. As we noted above, in Hoxby’s paper school districts can increase spending by raising either the property tax rate or the property tax base. This notion of the districts choosing their tax base is odd. If we followed this line of reasoning to its logical limit then we would conclude that there are no property-rich or property-poor districts, only districts that find it optimal to have a large tax base and others that find it optimal to have a small tax base. There are avenues through which districts can change property values; the paper points to capitalization, for example. But as a first approximation, it seems much more sensible to take tax base as given when thinking about the price of education than to think about a district consciously trying to change its tax base in order to manipulate state aid. This endogeneity of the tax base has important implications in Hoxby’s paper for the estimated tax price. Consider, for example, district power-equaliza- tion programs (DPE).7 The standard analysis of a DPE assumes implicitly that a school district’s tax base is fixed. Under a DPE, the state would choose a tax base per student V*. District j would then act “as if” its tax base were V* rather than Vj (assuming for the moment that V* is greater than Vj). That is, if it sets a tax rate tj it will raise tjVj from local sources, receive state aid Aj = tj(V* – Vj), and thus spend Ej = tjV* on education. To put things slightly differently, if a district wishes to spend Ej it would choose tj = Ej / V* and therefore receive Aj = Ej(1 – Vj / V*). Thus, under a guaranteed tax base grant it costs local governments only Vj / V* (which is less than 1) to raise an additional dollar of resources.8 In the Hoxby analysis, the community could choose (optimally) to increase

OCR for page 72
82 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM Vj in order to increase spending. But increases in Vj will lower state aid, and therefore the local price of education could be greater than $1 even if V* is greater than Vj. Hoxby estimates that aggressive equalization programs (in her analysis, Florida, Illinois, Minnesota, New Jersey, New York, Oklahoma, Texas, and Utah) typically have prices in the range of $1.20, though those states contain districts that face very different tax prices. These estimated tax prices play a crucial role in Hoxby’s paper; if we are not comfortable with this conceptual framework, then we need to consider her conclusions very cautiously. In all, we would argue that Hoxby’s paper represents an important method- ological contribution to the literature. Our approach cannot capture the differ- ences in the nature of the reforms across states. These differences were some- times substantial; as we noted above, the California Supreme Court in the Serrano decision, for example, required that spending per student vary by no more than $100 per student in the vast majority of districts. Thus, our finance reform variables reflect the average effect of a court decision but cannot capture the effects of a particular decision. Hoxby’s approach allows her to draw distinctions among reforms. Most of the available evidence from econometric studies of the experience of the states where the courts have overturned the system of school finance suggests that court-ordered reform has, in general, (1) raised the average level of spending and (2) successfully reduced inequality in spending. We believe that those conclusions are correct. Simulation Models The final source of evidence of the effects of school reforms on the level of spending and education and economic outcomes comes from simulation models. Nechyba (1997), Fernandez and Rogerson (1997), and Epple and Romano (1995) have all developed computable general equilibrium models that allow them to explore the effects of education finance reform. The Nechyba and Epple and Romano models are similar in spirit. Both models look at a metropolitan area and include housing markets, mobility, and a public sector where policy is deter- mined as in the median voter model. In both, educational output depends on education spending and peer-group effects. The Fernandez and Rogerson model does not consider peer group effects, but does offer one important extension. It is a dynamic model, and education is not a consumption good but an investment good. The current elderly care about the current young’s income, which is determined in part by education investments. All three models look at a funda- mental question: What happens in a general equilibrium setting when we move from decentralized to centralized school finance?9 The answers these three models offer are intriguing. Fernandez and Rogerson find that a centralized education finance system could generate significant wel- fare gains. Centralized finance reduces heterogeneity in education spending and can change the income distribution as well as the average income in society.

OCR for page 72
88 TABLE 3-5 Impact of Court-Mandated Finance Reform on Per-Pupil Revenue by Source of Revenue and Distribution of Household Income of Districts Coefficient {Standard Error} Coefficient {Standard Error} on Court Reform on Years after Court Reform Means {Standard Deviations} Dependent Variable: Dependent Variable: Per-Pupil Revenues ($1992) Per-Pupil Revenues ($1992) Per-Pupil Revenues ($1992) by Source by Source by Source Quartile of Median Household Income Local State Total Local State Total Local State Total First 1,640 2,367 4,007 -242 804 563 -40 102 62 {805} {1,036} {1,288} {153} {214} {238} {12} {15} {18} Second 1,842 2,005 3,847 170 370 539 -2 59 56 {924} {848} {1,230} {206} {165} {262} {17} {12} {20} Third 2,013 1912 3,926 173 379 552 2 49 51 {1,021} {812} {1,295} {248} {162} {269} {20} {12} {21} Fourth 2,456 1,756 4,211 435 155 590 4 21 25 {1,436} {735} {1,505} {345} {154} {332} {28} {12} {27} SOURCE: Authors’ calculations. THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM

OCR for page 72
89 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB mean estimate of the βit overall state/year pairs is 0.027. We then use the esti- mates for β values as dependent variables in a regression of the form β it = Dit γ + λi + θ t + vit (3) where λ and θ are state and year effects, respectively, vit is a random error, and Dit is one of our court reform variables. When Dit is defined as Court Reform, the estimate of γ (standard error) is –0.033 (0.037), whereas when the variable is Years after Court Reform, the coefficient is –0.004 (0.002). Thus, the evidence here is somewhat mixed. When we use our Court Reform variable, the impact of the courts on the link between income and spending is negative but insignificant; when we use our Years after Court Reform variable, the impact is negative and statistically significant. This method and results are similar to the work of Card and Payne (1997), who also find that finance reform has weakened the link between income and spending. We have also looked at the impact of court-mandated reform on spending for black and white students. Because black students tend to live in low-income districts, we would expect that court-mandated reform would redistribute re- sources toward black students. Using the distribution of black and white students in a district, we constructed race-specific shares of student enrollments. We then used these shares as weights to estimate race-specific, average per-pupil revenues for each state. We then used the average black and white per-pupil local, state, and total revenues as the dependent variables in a series of six regression models. These estimates are presented in Table 3-6. The results in Table 3-6 on race parallel our results on income in Table 3-5. We find in Table 3-6 that, following court-mandated reform, state aid directed TABLE 3-6 Impact of Court-Mandated Finance Reform on Per-Pupil Revenue by Source and Race Coefficient {Standard Error} on Mean {Standard Deviation} Court Reform Dependent Variable Per-Pupil Revenues ($1992) Per-Pupil Revenues ($1992) by Source by Source Racial Group Local State Total Local State Total Black 1,876 2,232 4,108 –216 664 448 {882} {922} {1,349} {178} {216} {247} White 1,978 2,015 3,993 140 434 574 {986} {820} {1,291} {200} {155} {249} SOURCE: Authors’ calculations.

OCR for page 72
90 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM toward black students rose significantly. However, we find some evidence that districts substitute state aid for their own revenues. State aid directed toward black students increased by an estimated $664 per student following reform; local spending on black students fell by $216. Thus, total per-pupil revenue for black students increased by $448 per pupil while per-pupil revenue for white students increased by $574. Here again we need to interpret these results cautiously since many of the parameters in Table 3-6 are imprecise estimates. OTHER CONSEQUENCES OF COURT-MANDATED EDUCATION FINANCE REFORM To this point, we have focused on the impact of court-mandated reform on education spending. Here we turn to other consequences of court decisions. We look first at the effect of the courts on education outcomes. We then focus on private contributions and the demand for private schools. Education Outcomes There are at least two ways to approach the question of the impact of court- mandated reform on education outcomes. The first is to argue that if finance reform changes outcomes, it does so largely by redistributing resources, and therefore we can answer this question by looking at evidence of the link between resources and outcomes. Several papers have taken this approach. Husted and Kenny (1996b) use data on 1987-88 and 1992-93 SAT scores from 37 states. They conclude that the mean SAT score is higher in those states with greater within-state variation in spending. Hoxby (1996) finds that an aggressive power- equalization plan would raise the dropout rate by 3 percent and that state-financed schools would raise the dropout rate by 8 percent. In general, as we argued above, the link between spending and outcomes is a contentious issue that has received a great deal of attention but remains unresolved nonetheless. The second approach is to look at changes in outcomes that follow court- mandated reform. This second approach has certain advantages. In particular, it captures any impacts of reform that are unrelated to changes in resources. For example, it is quite plausible to think that some of the education reforms in Kentucky following the Rose decision might in the end have a larger impact on outcomes than will the change in resources. On the other hand, perhaps schools will become less accountable following reform as the state role in education grows and the local role shrinks; in that case, the change in outcomes would be smaller than would be predicted based on changes in resources. A few papers have taken this second approach. Downes (1992) looked at the California experience following Serrano. He found that greater equality in spend- ing was not accompanied by greater equality in measured student performance. Downes and Figlio (1997) use individual level data from the National Longitudi-

OCR for page 72
91 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB nal Survey of the High School Class of 1972 (NLS-72) and the National Educa- tional Longitudinal Survey (NELS). Their evidence suggests that court-man- dated school finance reforms do not significantly change either the mean level or the distribution of student performance on standardized tests of reading and mathematics. They do find, however, that legislative reforms that are not a result of a court decision lead to higher test scores in general; the estimated effect was particularly large in initially low-spending districts. Card and Payne (1997) focus on the impact of finance reform on SAT scores. They carefully deal with an important selectivity problem that arises in any study that uses SAT scores; students who take the test are not a random sample of all students. The participation rate of high school seniors ranges from only a few percent (in states where relatively few students attend college and the SAT is not required for admission to the state university) to over 60 percent (in states where a high fraction attend college and the test is required by the state university). Moreover, students who take the SAT are far more likely to be from wealthy families and to be ranked highly in their graduating class. They conclude that the evidence points to a modest equalizing effect of school finance reforms on the test score outcomes for children from different family backgrounds (though they would agree that the evidence is not decisive). Their most precise estimates imply that, on average, court-mandated reform in 12 states over the 1980s closed the gap in average SAT scores between children of highly educated and poorly educated parents by about 10 points. Voluntary Contributions The demand for education spending varies across districts, in part because wealth and income vary across districts. As a consequence, almost inevitably any effort to reduce inequality in education spending will force some districts to spend less on education than they would like. How might families in those districts respond? They might try to supplement public funds with private dona- tions. Some districts put clear limitations on such donations. New York City, for example, has what has come to be called the “Greenwich Village rule.” Parents of children in P.S. 41 in Greenwich Village raised money to hire an additional fourth grade teacher. The head of New York schools, however, blocked their plan; in New York, he ruled, parents can raise money for all sorts of purposes (fixing the roof, buying new uniforms for the football team) but they cannot hire personnel.14 California gives parents more latitude, and Brunner and Sonstelie (1997) have shown that a number of districts have used private donations to reduce the impact of the limitations imposed by Serrano. They focus on the growth of local educational foundations, nonprofit organizations designed to channel voluntary contributions to local schools. There were six of these foundations in 1971 (the year of the Serrano decision); by 1995, there were 537. They also found that in

OCR for page 72
92 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM 1992, PTAs, local education foundations, and other nonprofits raised nearly $100 million for California public schools, the bulk of which was concentrated in a few school districts that were particularly constrained by the Serrano spending limita- tions. As Brunner and Sonstelie (1997) explain, these results are in some ways surprising. The “free rider” problem would seem to suggest that donations would be much smaller than they actually are in California. Suppose we all make decisions as to how much to contribute independently. If I contribute $1 and there are n children in the school, then my child will receive only $1/n in benefits from my contribution; for any reasonable size school, this is essentially zero. If my child will receive none of the benefits from my donation, perhaps I should donate nothing yet still enjoy the benefit of your contribution, i.e., it would make sense for me to act as a free rider. There is substantial literature that suggests that in general this free-rider problem is rarely as severe as theory suggests it should be; perhaps California families are more cooperative and are better able to avoid the free-rider trap than our textbooks tell us they should be. It is also important to keep the Brunner and Sonstelie results in proper perspective. While contributions may have allowed wealthy parents to increase spending a bit at the margin, these contributions were far too small to undo the effects of Serrano. In the early 1990s, California spent approximately $24.9 billion annually on its 5.3 million public school K-12 students. Brunner and Sonstelie’s $100 million thus represents spending of about $19 per student (roughly .4 percent). Even if all of the money were concentrated among the wealthiest 20 percent of all students, contributions were just $95 per student (2 percent). California PTAs would have to sell far more muffins at bake sales than they have in order to return California to a pre-Serrano world. The Demand for Private Schools How else might families respond to court-mandated spending equalization plans? They might choose to abandon the public school system entirely and send their children to private schools. As Brunner and Sonstelie (1997) explain, equalization throws some families off of their demand curve for education and thus generates a deadweight loss. Families could move back to their demand curve by choosing a private school, but of course they would then need to pay private school tuition. A family will choose private schools if the deadweight loss from equalization is greater than the additional cost of the private school. The empirical evidence on this question is mixed. Most of the research, not surprisingly, has focused on California. The raw data suggests that Serrano has not led many families to choose private schools. Brunner and Sonstelie (1997) show that about 9 percent of California school children were enrolled in private schools in 1973-74, as compared to roughly 10 percent in the rest of the country. By 1992-93, private school enrollment had increased to about 10 percent in

OCR for page 72
93 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB California and 12 percent in the rest of the country. They conclude that private school enrollment in California basically followed the national trend. Downes and Schoeman (1996), however, come to a different conclusion. They argue that even if the supply of private schools did not increase, Serrano could account for nearly half of the actual movement from public to private schools in California over the 1970-80 period. Husted and Kenny (1996a) have also looked at the impact of the court- ordered reform on the demand for private schools. Using data from the 1970, 1980, and 1990 census for 160 metropolitan areas, they find that as average spending per pupil rises in a state and as spending becomes more equal, private school enrollments fall. SUMMARY AND CONCLUSIONS The literature on the impacts of court-mandated finance reform is an impor- tant element of the research on education finance. This is not a surprising devel- opment. The passage of a quarter century since the landmark Serrano decision affords researchers the perspective (and data) by which court reforms can be evaluated. Since the primary emphasis of most reforms was to redistribute dol- lars, it is also not surprising that the vast majority of papers to date have examined whether these reforms have altered the distribution and level of education spend- ing. We believe the bulk of the evidence suggests that court-ordered reform has achieved its primary goal of fundamentally restructuring school finance and gen- erating a more equitable distribution of resources. In most cases, this was achieved by states’ directing more resources to districts with low local revenues. This last result is however not uniform, as the experience in California points out. The California evidence suggests that greater equality was achieved by reducing spending at the top of the distribution. In contrast, research on the other intended and unintended consequences of finance reform is just beginning to take shape. Again, this is not a surprising time line. If reform had no impact on the distribution or level of resources, there would be no need to look at many of these additional outcomes. Given the limited number of studies in this area, it is also not shocking that there is no consensus on whether court-ordered reform improved outcomes. We view re- search on the outcomes of finance reform to be the important next step. We realize that gains may be difficult to quantify. We may have attacked an easier set of questions through our own work simply because dollars are an easier variable to measure. More importantly, however, any change in outcomes will have to come about by a fundamental change in some characteristic of the dis- trict, such as a change in resources. Since outcomes are one more step down the causality chain, changes in these variables will obviously be more difficult to detect.

OCR for page 72
94 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM NOTES 1. In our models a state becomes a reform state when its supreme court issues its first opinion, even if the decision concerned the finance system more generally or if the decision was based on a legal issue, such as a declaratory judgment, and not on the merits of the case (declaratory judgment is a statutory remedy that allows a plaintiff to bring a suit if he is unsure of his legal rights). Others might have treated the data somewhat differently. Dayton (1996), for example, considers only state supreme court decisions that specifically address the disparity in education resources on the merits of the case. Under that defini- tion, we would not have considered Kansas as a reform state since the Kansas decision was a declaratory judgment. Nor would we have classified Wisconsin as a reform state since the Wisconsin system was overturned in 1976 on the grounds that the state’s equalization formula violated the state constitution’s tax article. 2. See Evans et al. (1997b) for a further discussion of these issues. 3. Proposition 13 was an amendment to the California State Constitution that limited property tax rates and property valuations, thereby limiting local governments’ access to the main source of funding for education. 4. See Murray et al. (1998) and Murray (1995) for a description of the data used in those studies. 5. Table 3-2 makes no adjustment for between-district differences in the cost of education. Thus, real resources in that table equal current dollar revenues deflated by the national Consumer Price Index. We discuss cross-section varia- tion in costs in a later section of this paper. 6. See Berne and Stiefel (1984) for a thorough discussion of the properties of measures of equity in public school resources. 7. See Coons et al. (1970). The school finance literature sometimes draws distinctions between guaranteed tax base, guaranteed yield, and district power equalization programs. These differences are minor, and we therefore refer to all three as district power equalization programs. 8. What happens in wealthy districts where V* is less than Vj? A pure DPE scheme includes “recapture.” All districts would receive Aj = tj(V* – Vj), which could be either positive or negative, i.e., school districts where the tax base is larger than the guaranteed base would be required to return to the state the excess tax revenue that it raised. Theoretically, DPE plans could therefore be self- financing; it is possible to set the guaranteed base so that the funds collected from the wealthy districts could be redistributed to poorer districts and therefore ΣAj would equal 0. In practice, however, no states with DPE require recovery; in fact, in most states even the wealthiest districts receive at least some state aid (Reschovsky, 1994). 9. Epple and Romano (1995) and Nechyba (1997) are also concerned with school choice issues.

OCR for page 72
95 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB 10. Burtless (1997) includes an interesting set of papers on this debate. 11. As Minorini and Sugarman (see Chapter 2 in this volume) explain, the original decision in Serrano allowed unequal spending, but did require that spending be uncorrelated with wealth. The California State Supreme Court es- tablished the spending rule in its third ruling in Serrano. 12. While these cost indices are the best available, it is not clear that they successfully capture the full difference in the costs of education across districts. Ideally, a cost index would account for the difference in wages that a central city school district would have to offer in order to attract teachers with the same qualifications, ability, and training that wealthy suburban districts attract. We suspect that these indices do not capture those differences and that it is therefore likely that they overstate the resources available to central city students. The available indices look at differences in the cost of inputs, but do not address variation in student needs; see Duncombe et al. (1996) for an important discus- sion of this issue. 13. We were also able to reestimate our decomposition using the individual district TCI available from the National Center of Education Statistics. The results of that decomposition are very similar to the estimates in Table 3-4; 57 percent of the overall inequality as measured by the Theil index were due to differences in resources between states. 14. In the end a “compromise” was reached; the city hired an additional teacher, and the donations were returned to the parents. ACKNOWLEDGMENTS This research was supported by the National Science Foundation under Grant #SPR-9409499. We thank NSF for its support. We also thank Helen Ladd, Allan Odden, and Christopher Jencks for very helpful comments on an earlier version of this chapter. REFERENCES Berne, R., and L. Stiefel 1984 The Measurement of Equity in School Finance: Conceptual, Methodological, and Em- pirical Dimensions. Baltimore: The Johns Hopkins University Press. Betts, J.R. 1995 Does school quality matter? Evidence from the national longitudinal survey of youth. Review of Economics and Statistics 77:231-50. Brunner, E., and J. Sonstelie 1997 School Finance Reform and Voluntary Fiscal Federalism. Unpublished working paper, November.

OCR for page 72
96 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM Bureau of the Census 1972 Census of Government School System Finance File, (F33). Machine readable data files. Washington, DC: U.S. Department of Commerce, Bureau of the Census and U.S. Depart- ment of Education, National Center for Education Statistics. 1973 Census of Population and Housing, 1970. Special Fifth Count Summary Tapes, Techni- cal Documentation. Washington, DC: U.S. Department of Commerce, Bureau of the Census. 1977 Census of Government School System Finance File, (F33). Machine readable data files. Washington, DC: U.S. Department of Commerce, Bureau of the Census and U.S. Depart- ment of Education, National Center for Education Statistics. 1982a Census of Government School System Finance File, (F33). Machine readable data files. Washington, DC: U.S. Department of Commerce, Bureau of the Census and U.S. Depart- ment of Education, National Center for Education Statistics. 1982b Census of Population and Housing, 1980. Summary Tape File 3F, School Districts, Technical Documentation. Washington, DC: U.S. Department of Commerce, Bureau of the Census and U.S. Department of Education, National Center for Education Statistics. 1987 Census of Government School System Finance File, (F33). Machine readable data files. Washington, DC: U.S. Department of Commerce, Bureau of the Census and U.S. Depart- ment of Education, National Center for Education Statistics. 1992 Census of Government School System Finance File, (F33). Machine readable data files. Washington, DC: U.S. Department of Commerce, Bureau of the Census and U.S. Depart- ment of Education, National Center for Education Statistics. Burtless, G., ed. 1997 Does Money Matter? The Effect of School Resources on Student Achievement and Adult Success. Washington, DC: The Brookings Institution. Card, D., and A.B. Krueger 1992 Does school quality matter? Returns to education and the characteristics of public schools in the United States. Journal of Political Economy 100:1-40. Card, D., and A.A. Payne 1997 School Finance Reform, the Distribution of School Spending, and the Distribution of SAT Scores. Working Paper #387. Princeton, NJ: Princeton University Industrial Relations Section. Coleman, J.S., E.Q. Campbell, C.J. Hobson, J. McPartland, A.M. Mead, F.D. Weinfeld, and R.L. York 1966 Equality of Educational Opportunity. Washington, DC: U.S. Department of Health, Education and Welfare. Coons, J.E., W.H. Clune, and S.D. Sugarman 1970 Private Wealth and Public Education. Cambridge, MA: Harvard University Press. Dayton, J. 1996 Examining the efficacy of judicial involvement in public school funding reform. Journal of Education Finance 22(1):1-27. Downes, T.A. 1992 Evaluating the impact of school finance reform on the provision of education: The Cali- fornia case. National Tax Journal 45:405-419. Downes, T.A., and D.N. Figlio 1997 School Finance Reforms, Tax Limits, and Student Performance: Do Reforms Level Up or Level Down? Unpublished working paper. Downes, T.A., and D. Schoeman 1996 School Financing Reform and Private School Enrollment: Evidence from California. Unpublished working paper.

OCR for page 72
97 WILLIAM N. EVANS, SHEILA E. MURRAY, AND ROBERT M. SCHWAB Downes, T.A., and M.P. Shah 1995 The Effect of School Finance Reforms on the Level and Growth of Per-Pupil Expendi- tures. Unpublished working paper, March. Duncombe, W., J. Ruggiero, and J. Yinger 1996 Alternative approaches to measuring the cost of education. Pp. 327-356 in Holding Schools Accountable, H.F. Ladd, ed. Washington, DC: The Brookings Institution. Epple, D., and R.E. Romano 1995 Public School Choice and Finance Policies, Neighborhood Formation, and the Distribu- tion of Educational Benefits. Unpublished working paper, July. Evans, W., S. Murray, and R. Schwab 1997a Schoolhouses, courthouses, and statehouses after Serrano. Journal of Policy Analysis and Management 16(1):10-31. 1997b Toward increased centralization in public school finance. Pp. 139-172 in Intergovern- mental Fiscal Relations, R.C. Fisher, ed. New York: Kluwer Academic. Ferguson, R.F., and H.F. Ladd 1996 How and why money matters: An analysis of Alabama schools. Pp. 265-298 in Holding Schools Accountable, H.F. Ladd, ed. Washington, DC: The Brookings Institution. Fernandez, R., and R. Rogerson 1997 Education finance reform: A dynamic perspective. Journal of Policy Analysis and Man- agement 16:67-84. Fischel, W.A. 1989 Did Serrano cause Proposition 13? National Tax Journal 42:465-473. Hanushek, E.A. 1986 The economics of schooling: Production and efficiency in public schools. Journal of Economic Literature 24:1141-1177. Heise, M. 1995 State constitutional litigation, educational finance, and legal impact: An empirical analy- sis. University of Cincinnati Law Review 63:1735-1765. Hoxby, C.M. 1996 All School Finance Equalizations Are Not Created Equal: Marginal Tax Rates Matter. Unpublished working paper, March. Cambridge, MA: Harvard University. Husted, T.A., and L.W. Kenny 1996a The Legacy of Serrano: The Impact of Mandated Equal Spending on Private School Enrollment. Unpublished working paper, March. 1996b Evidence from the States on the Political and Market Determinants of Efficiency in Edu- cation. Unpublished working paper, October. Manwaring, R.L., and S.M. Sheffrin 1995 Litigation, School Finance Reform, and Aggregate Educational Spending. Unpublished paper. McMahon, W.W., and S. Chang 1991 Geographical Cost of Living Differences: Interstate and Intrastate Update 1991. MacArthur/Spencer Series Number 20. Normal, IL: Illinois State University, Center for the Study of Educational Finance. Murray, S.E. 1995 Two Essays on the Distribution of Education Resources and Outcomes. Unpublished Ph.D. dissertation, University of Maryland, College Park, MD. Murray, S.E., W.N. Evans, and R.M. Schwab 1998 Education finance reform and the distribution of education resources. American Eco- nomic Review 88(4):789-812.

OCR for page 72
98 THE IMPACT OF COURT-MANDATED SCHOOL FINANCE REFORM National Center for Education Statistics 1974 Elementary-Secondary General Information System, Part B-1 Local Education Agency Fiscal Report for Fiscal Year 1972. Machine readable data files. Washington, DC: U.S. Department of Education, National Center for Education Statistics. 1994a School District Data Book for 1990 Census Data, CD-ROM. Washington, DC: U.S. Department of Education, National Center for Education Statistics. 1994b Digest of Education Statistics, 1994. Washington, DC: U.S. Department of Education, National Center for Education Statistics. Nechyba, T.J. 1997 Public School Finance in a General Equilibrium Tiebout World: Equalization Programs, Peer Effects, and Competition. Unpublished working paper. Reschovsky, A. 1994 Fiscal equalization and school finance. National Tax Journal 47:185-197. Rubinfeld, D.L. 1995 California Fiscal Federalism: A School Finance Perspective. Unpublished Working Pa- per, June. Silva, F., and J. Sonstelie 1995 Did Serrano cause a decline in school spending? National Tax Journal 47:199-216.