Beginning with Serrano v. Priest in 1971, equity-based decisions issued by state supreme courts led to a decrease in cross-district inequality in per pupil expenditures. In subsequent years, more state supreme courts overturned existing systems of public school finance for failing to provide adequate education to students living in poor school districts. Adequacy-based decisions have not produced measurable changes in cross-district inequality in expenditures, but have led to higher overall levels of funding for public education. The nationwide increase in per pupil expenditures over the past several decades is, however, largely the product of growth in personal incomes and a decline in the relative size of the cohort of school-age children, and not of court-ordered finance reforms. In California, after Serrano and the most far-reaching equalization reforms implemented anywhere in the country, the association between the wealth of a school district and educational quality remains strong and persistent. If one′s concern is the quality of education that students receive and not the amount of money spent on them, the victories that reformers have won in the courts have been hollow victories. (JEL I210, I220, I240).

FINANCIAL SERVICES

1. Introduction

Elementary and secondary education in the United States has traditionally depended upon local property taxes as the primary source of revenue. Because of large differences in property values, basing public school finance upon property taxes generated large differences in spending across school districts. In Serrano v. Priest (1971), the California Supreme Court ruled that the public school finance system then in place was unconstitutional because of the disparities in expenditures that it generated.

In subsequent rulings the California court mandated that per pupil expenditures be set at virtually identical levels across school districts. Similar suits have since been filed in many other states, and in 20 of them the supreme court has overturned an existing system of public school finance. 2 To comply with such decisions, state legislatures

1 An earlier version of this articlewas presented at the 2011 Annual Meeting of the American Political Science Association in Seattle, Washington. We would like to thank Mike Alvarez, David Grether, Philip Hoffman, Jonathan Katz, Terry Moe, Peter Ordeshook, Douglas Reed, Michiko Ueda, Sara Elizabeth Dahill-Brown, Michael Hartney, Ryane McAuliffe Straus, Christina Wolbrecht, and the anonymous reviewers for their comments and suggestions. We would also like to thank Alicia Fernandez and Matthew Kiewiet for research assistance.

https://advance.lexis.com/api/document?collection=analytical-materials&id=urn:contentItem:5NRM-CWH0-0227-V247-00000-00&context=

mailto:[email protected]

Page 2 of 26

The Impact of State Supreme Court Decisions on Public School Finance

have been compelled to design new funding mechanisms that do not produce the disparities in educational expenditures that the courts have found objectionable.

Or have they? In his landmark study, Rosenberg (1991) argues that unless a fortuitous constellation of political forces is present, court decisions do not in and of themselves lead to significant social reform. This is because of the courts′ inherent lack of power to bring about policy change on their own. Courts must, Rosenberg contends, "… depend upon the actions of others for their decisions to be implemented" (p. 336). In this area of policy, "others" refers to state legislatures. Several studies following Rosenberg have shown that it is difficult to compel state legislatures, composed as they are of reelection-seeking legislators, to do something they do not want to do. They have exhibited considerable ingenuity in circumventing constitutional spending and borrowing restrictions (Kiewiet and Szakaly 1996), and have succeeded in blunting the impact of ballot initiatives, sometimes sidestepping them completely (Gerber et al. 2000).

When confronted with a court mandate to overhaul an existing system of public school finance, the legislature can employ a wide range of tactics to avoid compliance (Hanushek and Lindseth 2009). Some state legislatures have concocted new funding mechanisms that do not fully address the court′s original objections, or that fail to pass muster on other constitutional grounds. Lacking enforcement mechanisms, state supreme courts are reluctant to push hard against recalcitrant state legislatures (Leyden 1992). Subsequent judicial action has also limited the scope of funding reforms. In California, the courts have ruled that Serrano and its successor decisions apply only to funds derived from local property taxes. The several billion dollars allocated annually by the state government to local districts in the form of 120 or so categorical aid programs-roughly a third of the total budget for K-12 education-is not subject to equalization constraints (Kollars 2003).

That state supreme court decisions have led to a reduction in the inequities and inadequacies of public school finance is thus something that cannot be presumed, but must rather be gauged empirically. It is also important to assess the consequences that state supreme court decisions

have had for other features of public school finance. Another major issue of concern is how court-ordered reforms affect the overall level of expenditures on public education. More equality in per pupil expenditures across school districts can be achieved by "leveling up." This is done by supplementing local property tax revenue in poor school districts with state revenue, derived from state taxes, to bring their expenditure levels up to (or at least closer to) the expenditure levels of wealthier districts. If states level up, court-ordered equalization reforms increase total expenditures on education. 3 But equalization can also be achieved by "leveling down." By constraining them or taxing them, in one way or another, wealthy school districts can be induced to spend less than they would otherwise, thus leading to lower total expenditures on education. Shifting responsibility for K-12 education from the local level to the state level, where public education must compete with many other salient spending priorities, may also lead to lower overall spending (Rubinfeld 1995; Mainwaring and Sheffrin 1997).

Whether states level up or level down matters a great deal. Most people most of the time favor greater equality in educational opportunities, but may be less supportive of this goal if it comes at the cost of a decline in overall financial support for public schools. Conversely, achieving greater equality in expenditures across local school districts by leveling up necessarily entails increasing state-level funding on education. This raises the issue of what tradeoffs must be made to do this, either in terms of higher taxes or decreased spending in other areas of government (Baicker and Gordon 2006).

the key judicial actors in this policy arena. Because there is much more inequality in average per pupil expenditures across states than within states, this decision also significantly limited the scope of finance equalization reform.

3 If states relied completely upon leveling up they would need to increase spending in all other districts up to that of the highest- expenditure district. This is not likely to be feasible. Leveling up is thus deemed to occur if average per pupil expenditures increase in a state, even though the equalization reforms that were implemented may have constrained expenditures in some districts. It should also be noted that expenditures on K-12 education in the United States have risen dramatically over the past several decades, from an average of $2606 (in 2007 dollars) in 1960 to $9910 in 2005 (Hanushek and Lindseth 2009). A better way to pose the question of whether states level-up or level-down is thus to ask whether states that have been subject to court- ordered equalization reforms exhibit faster or slower rates of expenditure growth than other states.

Page 3 of 26

The Impact of State Supreme Court Decisions on Public School Finance

A third question we seek to answer is whether the nature of the constitutional objection posed in public school finance litigation is consequential. Serrano v. Priest and the initial wave of state supreme court decisions that overturned existing systems of public school finance did so on the basis of "equity," that is, an unacceptable amount of inequality in per pupil expenditures across school districts that they generated. Between 1971 and 1983 plaintiffs in a half dozen states succeeded in persuading the state supreme court to overturn an existing system of finance on equity grounds, but failed to do so in several other states. This type of litigation was a tough sell, as it required appeal to a state constitution equal protection clause, as well as a finding by the court that district wealth is a

suspect classification and a strict scrutiny standard is therefore applicable (Roelke et al. 2004). Between 1983 and 1989, eight state supreme courts in a row rejected the plaintiffs′ claims and upheld the constitutionality of the existing funding system.

Beginning with the 1989 Kentucky Rose v. The Council for Better Education, Inc. decision, plaintiffs pursued a different line of argument that has proven to be more successful. Rather than directly attacking cross-district inequality in expenditures as unconstitutional, they instead sought to persuade the courts that existing finance systems do not provide an adequate level of funding for students in poor school districts. In the parlance of the field, "adequacy" cases supplanted the earlier "equity" cases. Education clauses in many state constitutions tend to be more supportive of demands that the provision of public education be adequate, and the remedy called for in adequacy cases-more state grants to poor districts-does not entail the wholesale reformulation of a public school finance system. Adequacy reforms also do not require that restrictions or disincentives be applied to high- expenditure districts. In some cases adequacy cases were filed because previous equity reforms were perceived to have been ineffective in aiding poor school districts.

This shift in the nature of public school finance litigation and the court decisions that flow from it could well lead to different outcomes. It is reasonable to hypothesize that court-ordered reforms based upon adequacy considerations, which do not attack the existence of inequality in expenditures per se, produce less equalization than equity-based reforms mandates. On the other hand, adequacy-based reforms, which direct more expenditures to poor districts whereas not requiring lower expenditures in wealthy districts, could also be expected to produce a higher average level of public school expenditures. It is our intention to determine if the shift from equity to adequacy considerations has had a measurable effect upon the distribution, as well as upon the overall amount, of public school expenditures.

In a book that provided many of the intellectual underpinnings of the plaintiff′s argument in Serrano, Coons et al. (1970) accept inequality as inevitable. They also accept that some parents are more willing and able than others to spend money on their children′s education. When school districts depend heavily upon their local tax base for revenue, however, children in poor school districts, who tend to be from poor families, are further disadvantaged. What Coons et al. reject is not inequality per se, but the idea that the state should accept institutional arrangements that reinforce already strong market-based tendencies toward inequality.

In their view, the goal of educational finance reform should not necessarily be the equalization of per pupil expenditures, but rather what they call "fiscal neutrality" (see also Long 1973). In a fiscally neutral funding system, the quality of education that schools provide is not correlated with the amount of wealth present in the district. The final task of this study is to determine the extent to which equalization of per pupil expenditures

leads to equalization in the quality of education that public schools are able to provide to their students.

Educational quality is often measured in terms of performance outputs such as scores on standardized tests- students who score higher on such tests presumably received higher quality education. We have known since the time of the Coleman Report, however, that the impact of per pupil expenditures or other school district characteristics is dwarfed by the impact of family characteristics (particularly the parents′ level of education) and other socioeconomic factors. For that reason we prefer quality indicators that are educational inputs, and which should therefore be a function of the financial resources available to school districts. A crucially important educational input is the person at the head of the classroom. The indicators of teacher quality that are available to

Page 4 of 26

The Impact of State Supreme Court Decisions on Public School Finance

us are the years of experience that teachers have, and whether or not they have attained an advanced (post- baccalaureate) degree.

Previous research has led many to doubt that these or any other readily observable teacher characteristics are valid measures of educational quality, primarily because they are often found to be uncorrelated with student test scores. 4 Our primary concern here, however, is to determine the degree to which these indicators are associated with the wealth of school districts. If fiscal neutrality can be achieved through funding equalization, following such reforms low-wealth districts should be able to compete effectively with high-wealth districts in the market for teachers-a market in which those with more experience and more education earn higher salaries. If so, the wealth of a school district would be uncorrelated with the amount of education and years of experience of the district′s teachers. But all else is rarely equal. Even if poor districts are provided with the financial resources to pay teachers the same salaries as wealthy districts, there are many reasons, for example, the safety and pleasantness of the classroom environment, location attractiveness, and cultural amenities, that would still make wealthy districts more attractive. In this case, teachers with more education and more experience will be more likely to seek and to gain employment in wealthy districts, and to stay there once they are hired. Finding that wealthier school districts are staffed by teachers with more education and more experience after equalization reforms have been implemented would thus be prima facie evidence that equalizing expenditures does not result in an equalization of educational quality, and that the reformers′ goal of fiscal neutrality has not been realized.

The question of whether increasing the financial resources available to school districts allows them to hire more experienced and more educated teachers raises a larger and more general point about the consequences of educational finance reform. Those advocating reform assume that quality of education is a function of expenditures, and that providing school districts with more money should lead to improvements in educational quality. Again, previous research indicates that this assumption is problematic. The large secular increase in per-pupil expenditures in the United States over the past several decades has done little to increase average test scores. Dozens of studies have failed to find a significant correlation between expenditure levels and student performance. Additional expenditures made in poor districts as a consequence of court-ordered equalization have done little to close the achievement gap between white and minority students (Hanushek and Lindseth 2009). True, some researchers, most notably Kremer (1995) and Verstegen and King (1998), conclude that increasing educational expenditures does lead to better student performance. In our view, however, the evidence concerning education production functions generally supports Hanushek′s (2008) assessment: "Commonly purchased inputs to schools- class size, teacher experience, and teacher education-bear little systematic relationship to student outcomes, implying that conventional input policies are unlikely to improve achievement" (p. 1).

Some amount of controversy concerning the impact of expenditures on student performance is sure to continue. What we can safely conclude from previous research is that increasing expenditures on education does not necessarily lead to gains in educational attainment, and that any gains that are achieved are generally too small to be detected in studies based upon conventional sample sizes. How much money is spent on education and how it is distributed across school districts are important matters of public policy, and it certainly is important to understand how these policies are affected by state supreme court decisions. It must be acknowledged, though, that if educational expenditures did have clearer, more demonstrable effects upon student performance, the state supreme court decisions we are examining and the litigation that led up to them would be far more consequential for the educational experience and life opportunities provided to public school students in this country.

Our study is not the first to bring evidence to bear on the questions we have posed. We are, however, in a position to make a more definitive assessment of court-ordered equalization reforms than has been previously possible. This is due primarily to the simple passage of time. In the most comprehensive study in this area to date, Murray et al. (1998) had available five panels of school district expenditure data, collected at 5-year intervals, thus spanning a

4 Instead of estimating the effect of measurable characteristics such as teacher experience and education upon student achievement scores, some researchers estimate the effect of teacher quality in quasi-experimental settings. After randomly assigning students and teachers to different classrooms, subsequent gains (or losses) in test scores achieved by students in individual teachers′ classrooms are used as "value-added" measures of teacher quality (Hanushek and Rivken 2010).

Page 5 of 26

The Impact of State Supreme Court Decisions on Public School Finance

20-year period. We now have eight panels of these data spanning 35 years. The additional data allow us to make more precise estimates of the impact of court-ordered education finance reforms. We are also able to gauge how quickly reforms occur in response

to court decisions, and to distinguish between short-term and long-term impacts. It is also the case that the 1989 Rose decision was issued only a few years before the last year (1992) in Murray et al.′s data series. With data extending out to 2007, we can differentiate between the consequences of equity-based reforms and those made in response to adequacy-based considerations. Finally, the passage of time means that we are also better able to investigate the long-term impact of public school finance reform upon the goal of fiscal neutrality. More specifically, we can determine the extent to which educational attainment remains a function of district wealth a generation after Serrano.

Answers to these questions are important in their own right, but the findings of our analyses also add to the relatively meager empirical record concerning the policy-making role of state supreme courts. As Brace, Hall, and Langer (2001) put it, "… there is a remarkable and unfortunate asymmetry between the political importance of state supreme courts and the attention given to them by the research community" (p. 81). There are reasons why state supreme courts have not received as much attention as they should; a major one is the wide-ranging nature of their jurisdiction. These courts issue so many rulings in so many areas of law and policy that it is difficult to compare the output of one court with another. Nearly every state supreme court in the country, however, has ruled on the validity of existing systems of public school finance, and on the basis of the same criteria. It is thus possible in this policy area to make comparisons across states, to identify patterns in the data, and to make statistically sound inferences about them.

2. Previous Research

Studies that have estimated the impact of court rulings upon expenditure inequalities agree that the resultant funding reforms reduce cross-district inequality, but differ in their assessments of the magnitude of this effect. Murray et al. (1998) report that states in which the supreme court overturned an existing system of public school finance experienced significant reductions in inequality in cross-district expenditures. Springer et al. (2009) find a similar pattern, but estimate the magnitude of the effect to be about half of what Murray et al. reported. They also test the hypothesis that reforms made in response to equity-based lawsuits, which target inequality per se, result in larger reductions in inequality than do reforms made in response to demands for adequate levels of expenditures in all districts. Some of their data support this hypothesis, but other data derived from other sources do not.

Murray et al. also investigate the impact of court-ordered reforms upon the overall level of educational expenditures. Their study, supplemented by Evans et al. (1997) and Corcoran et al. (2003), weighs in unequivocally in favor of the leveling-up hypothesis: "… the state government raises spending for education in the poorest school districts, leaves spending in

the wealthiest districts unchanged, leaves spending outside of education unchanged, and thus by implication funds the additional aid to the poorest districts by raising state taxes" (806). Using state-level data collected between 1970 and 1990, Manwaring and Shefrin (1997) report that states subject to court-mandated reform exhibited a small net gain in average per pupil expenditures. Their findings are thus also consistent with the leveling-up hypothesis, but they also report that there is a great deal of variance in post-reform expenditures. Some states experienced gains, but others, for example, California and Washington, experienced declines.

Several other studies in this area have assessed the impact of educational finance reform in a single state- particularly California, which is where the school finance reform movement began with the Serrano decision. Silva and Sonstelie (1995) postulate that a shift from local to state financing produces a tax price effect that tends to increase expenditures, but also an offsetting income effect that discourages expenditures. They observe that after Serrano per pupil spending in California fell relative to other states, so the income effect presumably dominated. Fernández and Rogerson (1999, 2003) model public school expenditure levels under four different regimes: purely local finance, centralized (state-level) finance, foundation systems, and power equalization formulae. The latter two systems utilize both state and local revenue sources. Most states have long had foundation systems, in which

Page 6 of 26

The Impact of State Supreme Court Decisions on Public School Finance

states guarantee a particular level of per pupil funding as long as local school districts impose a certain rate of local property taxation (Verstegen and Jordan 2009). Fernández and Rogerson′s model predicts that when the distribution of income and other parameters are calibrated with realistic values, local financing yields higher levels of education expenditures than state financing, and a foundation system more than either purely local or purely state-level financing. Because Serrano led to the replacement of a foundation system with centralized state-level funding, they concur with Silva and Sonstelie that the Serrano decisions produced a decrease in expenditures on public schools in California.

The conclusions of these studies, that is, that Serrano and follow-up rulings led to an equalization of per pupil expenditures but contributed to a decline in overall spending on public education in California, has to a large extent become the conventional wisdom. But Gerber et al. (2000) take issue with the leveling-down scenario. They agree that per pupil expenditures in California fell in the years after Serrano, but only because tax and expenditure limitation initiatives, Proposition 13 and Proposition 4, limited state spending overall. The share of total (state and local) spending allocated to K-12 education actually increased in the years immediately following the Serrano decisions (p. 107). They further report that the impact of the tax and expenditure limitations had dissipated by the mid-1980s, and that the decline in per pupil expenditures in California relative to other states is primarily attributable to the drop in personal income relative to other states.

Hoxby and Kuziemko (2004) focus on public school finance reform in Texas. In an earlier study, Hoxby (2001) observes that states face many competing demands upon their revenues, and so state legislatures tend to favor equalization reforms that level down, that is, that limit expenditures in districts with high per pupil expenditures. Hoxby′s model also incorporates a key feature of public school finance missing from other models: school district quality is capitalized into property values. If equalization is achieved by constraining spending in high-expenditure districts, leveling down will depress property values in these districts and thus yield even less revenue for public schools. Hoxby and Kuziemko (2004) estimate this model and find that the equalization reforms implemented in Texas cost the state approximately $27,000 per student in taxable property values.

The final question our analysis addresses is that of fiscal neutrality: does equalization in per pupil expenditures weaken the correlation between school district wealth and educational quality? As far as we are able to determine, previous research has not addressed the issue of fiscal neutrality directly. A study of some relevance here is that of Downes (1992), who tracks changes in average standardized test scores across California school districts between 1977, when the Serrano equalization mandate had just been implemented, and 1986. Despite the equalization of per pupil expenditures, educational outcomes, registered by scores of sixth graders on the California Assessment Program (CAP) test, were still highly correlated with school district wealth. A subsequent study of finance reform in Vermont produced similar findings (Downes 2004).

Like many other previous researchers, Downes accepts average test scores as an indicator of school district quality. Because test scores are educational outputs that are heavily influenced by family background characteristics and other factors, we prefer instead to rely upon educational inputs supplied by school districts, that is, teacher experience and teacher education, as measures of quality. If fiscal neutrality can be achieved through funding equalization reforms, low-wealth districts should be able to attract and to retain teachers with as much education and experience as high-wealth districts.

Whether or not such measures of teacher quality translate into higher student test scores is another question. There are large, carefully executed studies that find they do not (Rivkin et al. 2005; Buddin and Zamarro 2009). There are other, large, carefully executed studies that find that more experienced teachers do produce larger gains in student test scores than less experienced teachers (Clotfelter et al. 2007), and other studies that find that teachers with more education also produce positive effects. After collecting and reviewing a large body of data and evidence from all 50 states, Darling-Hammond (1999) concludes that more experienced teachers are more effective, but that the effect of experience appears to level off after about 5 years or so. Many studies she reviews report that teachers with master′s degrees also produce larger gains in test scores, but

Page 7 of 26

The Impact of State Supreme Court Decisions on Public School Finance

that the effect is modest and uneven compared to the effect of experience. According to Rice (2003), teachers with advanced degrees produce large gains in student achievement, but this effect is primarily confined to teachers whose degrees are in the subject area, for example, science or math, that they teach.

As in many areas of social scientific research, then, there does not exist a consensus as to whether or not readily observable measures of teacher quality-years of experience and the possession of an advanced degree-reliably lead to higher student scores on standardized tests. What we do know, however, is that education and experience are attributes that are positively valued by labor markets in general, and that there is every reason to believe that they count for a great deal in the market for teachers as well. For this reason the percentage of teachers in a school district who have attained an advanced degree or who have more years of experience should be a valid measure of the quality of education that the district supplies, and thus useful in determining if educational finance reform has produced significant movement toward fiscal neutrality- even if these measures are not linked to higher test scores.

At this point it is useful to summarize the main findings of previous research. State supreme court decisions that overturn public school funding systems appear to lead to a reduction in the level of inequality in cross-district per pupil expenditures, but estimates of the magnitude of this effect vary considerably. These decisions also appear to produce some leveling up increase in per pupil expenditures, but again, estimates of this effect range from widespread and substantial (Murray et al. 1998) to modest and highly variable (Manwaring and Sheffrin 1997). Whether equity-based decisions have had a more significant impact than adequacy-based decisions upon patterns of public school finance remains an open question. Finally, the few previous studies in this area that have examined the matter of fiscal neutrality, albeit indirectly, tell us that expenditure equalization reforms have so far not led to equalization in educational outcomes.

It is often said that time is money, but more time also means more data. Additional data now available allow us to more accurately estimate the impact of court-ordered education finance reforms, and to reduce the amount of uncertainty that is currently present with respect to many of the key issues. Where previous research has tended to provide more definitive answers, we can either generate more confidence in these findings or call them into question. The first data to be updated are the court decisions themselves.

3. State Supreme Court Decisions and Public School Finance

Most state supreme courts have considered public school finance cases since the 1971 Serrano decision, and, as indicated previously, in 20 states they have found the existing system for funding public schools to

be unconstitutional. In five states they have done so twice. The entries in Table 1 report the state and year in which these decisions occurred. 5 Before 1989 most decisions to reject an existing finance system were due to the court finding cross-district disparities in per pupil expenditures to be unacceptable. These are listed in the second column. Since then most such decisions have been based instead upon a finding that expenditures in some districts were below what was needed to provide students with an adequate education, and these are listed in the third column. Previous studies in educational finance have not differentiated between state supreme court decisions and those made by lower trial courts. We think it is important to do so, mainly because it is reasonable to expect state legislatures to be more responsive to mandates issued by the state supreme court than by a lower trial court. The

5 Our coding scheme naturally differs from that of Murray et al. (1998) in that it incorporates the many decisions state supreme courts have rendered since 1992. Our coding of court decisions also differs somewhat from those of West and Peterson (2007), Hanushek and Lindseth (2009), and Springer et al. (2009) because we distinguish between state Supreme Court decisions and lower court decisions. In the case of West Virginia, Springer et al. base their coding on a 1984 decision, but we use instead the earlier 1979 decision remanding the case to the lower court in the manner of Serrano vs. Priest. Finally, we also differ from some of the previous coding schemes in that we do not include the 1997 decision of the Michigan state supreme court, as it pertained to special education only, nor the 1976 decision of the Wisconsin state supreme court that, instead of requiring more equality in expenditures, overturned an act of the legislature that had been designed to do precisely that. In making our coding decisions we were aided by the Advocacy Center for Children′s Educational Success with Standards (ACCESS) website, which provides excellent information on education finance reform in each state.

Page 8 of 26

The Impact of State Supreme Court Decisions on Public School Finance

fourth column in Table 1 thus reports public school finance decisions issued by lower courts. It is likely that these lower court decisions will eventually be ruled upon by the state supreme court, but that has not happened so far.

4. Court-Ordered Reform and Inequality

A good way to begin our analysis of these court decisions is to build upon Murray et al.′s (1998) study, which compares expenditure patterns before a supreme court has overturned an existing system of public school finance and after they have done so. They use expenditure data, collected from individual school districts at 5-year intervals by the Census of Government surveys, to calculate state-level inequality and average expenditure measures. We have appended to their data, which they generously provided us, three additional panels collected in 1997, 2002, and 2007.

To ensure the validity of our analysis the new data must be comparable to the original Murray et al. data, and so we replicate their procedures in handling the 1997-2007 data. We therefore weight district expenditure data by district enrollment, which in some states varies by many orders of magnitude. In 2007, for example, the smallest school district in

Table 1. Court Decisions that Overturned an Existing Public School Finance System

State Equity Decision Adequacy Lower Court

Decision Decision

Alabama 1993

Arizona 1994

Arkansas 1983 2002

California 1971

Connecticut 1977 1996

Idaho 1998

Kansas 2003

Kentucky 1989

Maryland 2000

Massachusetts 1993

Missouri 1993

Montana 1989 2005

New Hampshire 1993

New Jersey 1973 1990

New Mexico 1999

New York 2003

North Carolina 2004

Ohio 1997

South Carolina 2005

Tennessee 1993

Texas 1989

Vermont 1997

Washington 1978

West Virginia 1979

Wyoming 1980 1995

Page 9 of 26

The Impact of State Supreme Court Decisions on Public School Finance

California had 147 students whereas the largest, Los Angeles Unified, had 707,647. We also drop a small number of extreme outliers that bear little resemblance to conventional school districts. 6 It is also necessary to confine attention to unified (K-12) school districts only, as expenditures made at the high school level are considerably higher than at the elementary level. This results in the exclusion of Montana and Vermont from the analysis as they do not have unified districts. Hawaii is similarly excluded because it has only one statewide school district, as is Alaska due its highly

idiosyncratic public finances. These data, derived from individual unified school districts, are then aggregated to produce state-level Gini indices and other measures of inequality, as well as average per pupil expenditures expressed as real (1992) dollars. The analyses undertaken here are based upon the resultant state-level measures. Eight panels of data (1972-2007) from 46 states yields 368 observations in total.

Murray et al. (1998) and other researchers in this area gauge the impact of court decisions ordering educational finance reforms by making simple, before versus after comparisons. The eight panels of data that we have available, however, provide a long enough time frame to investigate both the immediate, short-term response to these court decisions as well as the long-term impact. It is important to be able to do this. It may well be that state legislatures are responsive to court decisions, that they alter or replace existing systems of public school finance to satisfy the court, but that the effects of such innovations dissipate over time. This can happen because of subsequent legislation or court decisions that weaken an initial round of equalization reform, or because of the adaptive behavior of students′ parents. In the decades following the Serrano decision, for example, wealthier school districts in California have had relatively greater success than other districts in gaining approval for local parcel taxes that can be used to augment baseline funding, and have derived more benefit from the financial contributions of local educational foundations (Brunner and Sonstelie 1996; Freelon et al. 2012). Parcel taxes and voluntary contributions raise expenditures on education, but also produce more inequality in the distribution of these expenditures across school districts.

As discussed above, we distinguish between court decisions that reflected concerns about inequality per se-equity cases-from those that determined that poorer districts had been receiving too little funding to provide an adequate education to their students. We also need to be realistic about just how quickly state legislatures can respond to court decisions. Because school years start on July 1 of the previous calendar year, it seems reasonable to assume that if a court decision occurs in year t, any funding reforms made in response to it could not take hold until school year t + 2.

The entries in Table 2 report the number of times cross-district inequality in per pupil expenditures, as measured by the Gini coefficient, increased or decreased, compared to the previous panel year (5 years earlier), following a state supreme court finance reform mandate. 7 The changes that are reported are in the panel year immediately following

Table 2. Change in the Distribution of Per Pupil Expenditures Following Court Decisions

6 We were originally concerned that Murray et al. (1998) were making a mistake in excluding extreme outliers, which they designated as districts with per pupil expenditures greater than 150% of the 95th percentile or less than 50% of the 5th percentile. One of the primary questions of interest in this research, after all, is inequality in expenditures, and so extremely high- expenditure and low-expenditure districts could well be the most strongly affected by education finance reforms. It turns out that only a tiny fraction of districts are excluded by this rule (49 of 10,411 in 2007), and further inspection reveals that these are quite unusual districts. Louisiana, for example, has a Recovery School District (mainly serving students displaced by Hurricane Katrina) that had per pupil expenditures in 2007 that exceeded the 50% higher than the 95th percentile criterion. Another such outlier in the 2007 data is the Bois Blanc Pines School District of Michigan, which enrolled a total of 2 students.

7 Murray et al. (1998) calculate four different measures of inequality: the Gini coefficient, the Theil index, the log of the ratio of spending at the 95th and 5th percentile, and the coefficient of variation. The Gini coefficient, the coefficient of variation, and the Theil index are highly correlated with each other and can usually be used interchangeably. Because the most significant effects of state supreme court reform mandates may well occur in either extremely wealthy or extremely poor districts, the log of the ratio of the 95th percentile and 5th percentile may be problematic in this context.

Page 10 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Variable Inequality Increased

Inequality Decreased

Total

Equity decision, t + 1

0 7 7

Equity decision, t + 2

2 5 7

Equity decision, t + 3 plus

8 6 14

Adequacy decision, t + 1

4 11 15

Adequacy decision, t + 2

5 6 11

Adequacy decision, t + 3 plus

8 8 16

Trial court decision

5 5 10

No reform 127 115 242

the court decision, in the next panel (5 years later), and in the panels after that. In the first panel year following an equity-based decision, inequality in cross-district expenditures fell in all seven of the affected states, and five times out of seven over the course of the next 5 years. In the third and all subsequent panel years following such decisions, however, the degree of inequality in these states increased as often as it decreased (8 of 14 times). Adequacy-based court decisions produce a much weaker pattern. Inequality decreased in 11 of the 15 affected states in the years immediately following an adequacy-based decision, but in all subsequent years there was no tendency for inequality to either increase or decrease. There is similarly no systematic change in cross-district inequality in states where trial courts have ruled against an existing system of public school finance, or in the states that have not experienced court decisions mandating reform.

The findings in Table 2 indicate that state supreme court finance reform decisions have led to decreases in expenditure inequalities-particularly when they were of the first wave of equity-based decisions. But this is just an initial, rough cut. In order to arrive at more reliable and more precise estimates of the effects of state supreme court decisions we replicate Murray et al.′s regression analysis by estimating a (time and entity) fixed effects model, which is of the following form:

Yit = α + βXit + γZit + δi + λt + εit

(1)

where Yit is the Gini coefficient for state i in year t. X is the battery of indicator (dummy) variables, which, following Table 2, take on the value of 1 in the panel year following a state supreme court decision mandating educational finance reform, in the subsequent panel year, and in the third and all subsequent panels. There are thus three indicators each for equity-based and adequacy-based reforms, as well as an indicator for observations following a lower court decision. Our hypotheses are that equity-based court decisions produce less inequality, but that adequacy-based decisions and lower court decisions do not.

Z refers to background political and socioeconomic variables that include a measure of citizen ideology (Fording 2010), the proportion of the state′s population between ages 5 and 17 as well as the proportion over 65, and the extent of income inequality in the state (also measured by the Gini coefficient). Our expectations are that state supreme court decisions mandating less cross-district inequality in educational expenditures are more likely in states that are more liberal, that have more underlying income inequality, and larger proportions of school-age children. δi are state-specific constant terms and λt the year-specific constant terms.

Page 11 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Previous studies in this area by Manwaring and Shefrin (1995) and Springer et al. (2009) also estimate time and entity fixed effects models. This is a standard approach to the analysis of time-series cross-section (TSCS) data, and it has the virtue of introducing pervasive control for omitted variables. But it also requires strong assumptions that, if not satisfied, can worsen biases due to measurement error and misspecification. Out-of-sample predictions are also impossible because it is necessary to have information about year effects which have not yet happened. It may also be preferable to estimate an alternative to the fixed effects model-the random effects model-when the number of panels in the time series is small, and an autoregressive model when serial correlation is present (Beck 2001). Inspection of residuals from the fixed effects model (Equation 1) revealed a small but significant amount of serial correlation, and somewhat higher levels of serial correlation were present in the fixed effects regressions involving per-pupil expenditure levels.

More generally, we can attain a much higher level of confidence in our results if we are able to determine that they are not significantly affected by the nature of the particular model that is estimated (Angrist and Pischke 2008). We therefore assess the impact of court-ordered finance reform upon cross-district inequality in per pupil expenditures by estimating three other commonly used models. The random effects model, mentioned previously, is shown in Equation 2. It specifies the same substantive variables but also includes a state-specific random element ui:

Yit = α + βXit + γZit + ui + εit (2)

The third model we estimate is an autoregressive model (AR1), as shown in Equation 3:

Yit = α + βXit + γZit + uit (3)

where uit = ρuit-1 εit and -1 ≤ ρ ≤ 1

Finally, we also estimate a first differences model, as shown in Equation 4. This is a variation on the basic fixed effects model in that it also includes indicator variables for the various panel years, but specifies change in inequality from one panel year to the next (i.e., over 5 years) instead of inequality per se.

∆Yit = α + βXit + γ∆Zit + λt + εit

(4)

In all four equations we follow the lead of previous researchers and specify interventions of the court as exogenous indicator variables. But shouldn′t policy interventions by political authorities be seen as endogenous? As Heckman (1978) explains in his classic treatment of this issue, these sorts of variables are both indicators of latent variables as well as changes in behavior. Say we observe, for example, that cross-district expenditures are more equal in states that have had supreme court decisions mandating funding reforms. This could be a consequence of the court decisions, but it could also be due, at least in part, to the population of that state being more supportive of equality in educational funding and the state supreme court following popular sentiment. If so, estimates of the effect of court-ordered finance reforms would be biased upward.

Why, then, do we persist in specifying the decisions of state supreme courts as exogenous variables? We do so because we find it impossible to specify the simultaneous equation system needed to estimate these terms as endogenous. We could find no variable, or set of variables-including the other variables specified in Equation 1-that were at all predictive of educational finance decisions. State supreme court decisions were not a function of the amount of inequality initially present in cross-district expenditures or of state ideology. Nor did they reflect differences in the education clauses or other features of state constitutions (Thro 2010). State supreme court decisions may well be endogenous, but we are simply not privy to whatever the considerations are that influence them to rule one way or the other on education finance cases.

Page 12 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Estimation results are reported in Table 3. The top number in each entry is the regression coefficient, the bottom number below the standard error. Where appropriate, that is, in equations 1, 3, and 4, we report panel-corrected standard errors. Entries in the first column are for the time and entity fixed effects model, those in the second for the random effects model, those in the third for the autoregressive, and those in the fourth column for first differences. Coefficients of the state and year indicators specified in the first and fourth models are not reported.

The results of the regression analyses mirror the pattern observed in Table 2. In the fixed effects model (Column 1), the -1.71 coefficient of the first "equity" term implies that these court decisions produced a significant decline in inequality in the first subsequent panel year. The similar coefficients of the following two terms (-1.86 and -1.91) reveal that the movement toward equality persisted through the duration of the time period covered by the data. Results associated with the random effects model (Column 2) are virtually identical. The coefficients of these variables estimated in the autoregressive model are smaller than those obtained in the first two models, but describe the same pattern: a significant reduction in inequality in the first panel year following an equity-based decision, followed by the persistence of this reduction over time. Coefficients of all three equity-based decision terms in all three of these equations are statistically significant.

Table 3. The Effect of State Supreme Court Finance Reform Decisions upon Cross-District Inequality in Per Pupil Expenditures

Variable Fixed Random AR(1) First

Effects Effects Differences

Equity decision, t + 1

-1.71** -1.61** -1.18** -1.92**

(0.45) (0.51) (0.45) (0.45)

Equity decision, t + 2

-1.86** -1.83** -1.01* -0.19

(0.46) (0.52) (0.54) (0.44)

Equity decision, t + 3 plus

-1.91** -2.03** -1.02* 0.04

(0.42) (0.44) (0.53) (0.37)

Adequacy decision, t + 1

-0.54 -0.64* -0.51 -0.45

(0.35) (0.36) (0.40) (0.28)

Adequacy decision, t + 2

-0.36 -0.47 -0.37 0.02

(0.47) (0.43) (0.58) (0.33)

Adequacy decision, t + 3 plus

-0.55 -0.59 -0.22 0.10

(0.46) (0.43) (0.66) (0.36)

Trial court decision

0.29 -0.03 -0.07 -0.27

(0.37) (0.45) (0.40) (0.33)

State ideology -0.01 0.00 0.01 -0.01

(0.01) (0.01) (0.01) (0.01)

Percent 5-17 0.01 -0.06 -0.04 0.03

(0.05) (0.04) (0.08) (0.09)

Percent over 65

-0.04 -0.28** -0.13 (0.07)

(0.09) (0.10) (0.09) (0.14)

Page 13 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Income inequality

-1.33 -4.81** -3.10 2.65

(3.10) (1.82) (3.70) (3.93)

Constant 9.91** 13.72** 10.26** 0.35

(2.04) (2.11) (3.07) (0.23)

R 2 0.78 0.02 0.43 0.21

Rho 0.62

N 368 368 368 322

*p < 0.05, **p < 0.01.

Coefficients of these variables in the first differences model (Column 4) display a somewhat different pattern, but this is because it is a first differences model. Properly interpreted, the coefficients follow exactly the same pattern evident in the other three equations. The -1.92 coefficient in the first panel after such a court decision shows that when the supreme courts spoke, at least in the context of equity-based school finance decisions, state legislatures listened and quickly acted. The small and insignificant coefficients of the following two equity decision variables imply that after the initial policy response evident in the first panel year following the decision, no additional change in our inequality measure was evident over time.

The impact of equity-based court decisions upon inequality in educational expenditures that we observe is uncannily similar in magnitude to that reported by Murray et al. (1998). This is gratifying, but it also makes sense. Although they did not differentiate between equity-based and

adequacy-based decisions, the court decisions in their dataset, which ends with the 1992 panel, are predominately equity-based. In contrast to the effects attributable to equity-based court decisions, coefficients of the adequacy- based decision indicators in all four regression models register a much smaller decline in inequality, and only one is large enough to reach conventional levels of statistical significance. Coefficients of the trial court decision indicators in all equations are also small and insignificant, as are those of almost all the other variables specified.

To get a sense of how much movement toward equality in cross-district expenditures that effects of this magnitude imply it is useful to compare the two histograms shown in Figure 1. The first histogram portrays deviations from the state mean expenditure level in Texas in 1992, the second in 1997. Between these two panel years the Gini index declines by 1.70, which is very similar to most of the regression coefficients for the equity-based decision variables reported in Table 3. There is a noticeable shrinkage in variance from the first histogram to the second, and additional calculations indicate that the percentage of districts in which per pupil expenditures were less than 90% of the state average fell from 11% to 5%. This amount of change is non-trivial, but it falls well short of being a tectonic shift. California thus remains an unusual case in this respect, as the Gini coefficient for cross-district inequality in per pupil expenditures declined by 5.6 between 1972 and 1976. No other equity-based decision rendered by a state supreme court has come close to having the equalizing impact of Serrano.

Figure 1 also helps us understand why equity-based funding reform decisions did not result in more dramatic changes in school funding patterns. Even before Serrano and the equity-based funding reform decisions that followed, states were providing large amounts of revenue to school districts to supplement local property tax revenues. This assistance had already reduced expenditure inequalities to much lower levels than would have existed if school districts had been funded solely by local property taxes.

2 In San Antonio Independent School v. Rodriguez (1973), the US Supreme Court ruled that public school finance was a state- level issue and did not fall under the purview of federal constitutional law. As a result of this ruling, state supreme courts became the key judicial actors in this policy arena. Because there is much more inequality in average per pupil expenditures across states than within states, this decision also significantly limited the scope of finance equalization reform.

Page 14 of 26

The Impact of State Supreme Court Decisions on Public School Finance

A potential problem with our assessment of the impact of state supreme court decisions is that we do not account for "spillover" effects. Legislatures in other states may have viewed this sort of litigation as troublesome, controversial, and to be avoided, and thus adopted educational finance reforms in order to head it off. They might have concluded that by preemptively adopting certain modest measures, for example, some additional aid to poor districts, they could reduce the likelihood of the state supreme court ordering them to make more expensive and far- reaching reforms. If so, our analysis underestimates the influence of the courts. Alternatively, we do not and cannot observe the counterfactual situation of what would have happened in states where the supreme court overturned an existing system of finance if the courts had not acted as they did. The courts′ decisions may have simply made educational finance a higher legislative priority, and led the legislature to address equity or

Figure 1. Deviation of District per-Pupil Expenditures from Weighted Mean.

adequacy concerns sooner rather than later. This would be especially plausible if finance reform was flowing with the "current of history," as Rosenberg (1991) puts it, and enjoyed significant public support. If this is the case we would have over-estimated the consequences of these court decisions.

For a number of reasons we do not believe our findings are compromised by the problems suggested by either scenario. First of all, state legislatures need not have viewed court rejection of an existing finance system as inevitable, or even likely. In the decisions listed by Murray et al. (1998), plaintiffs won only 16 of 54 times. Secondly, as discussed previously, state legislatures can choose to resist, or at least indefinitely delay, complying with the court′s mandate. Finally, when we confine attention to the subset of states in which the supreme court has not ruled against an existing system of public school finance, we see no evidence of the current of history flowing in one direction or the other. Between 1972 and 1997 inequality in cross-district expenditures tended to decline, but thereafter tended to increase. In 2007 the average Gini coefficient in these states was no different from what it had been in 1972.

5. Court-Ordered Reform and Average Expenditures

We next assess the impact of court-ordered finance reform upon educational expenditure levels by estimating regression equations that mirror those of the inequality analysis. Here the dependent variables are average per pupil expenditures in state i at time t in the fixed effects, random effects, and autoregressive models, and change in average per pupil expenditures in state i since the previous panel year in the first-differences model. All expenditure and income figures are expressed as real (1992)

dollars. As before, we measure the impact of supreme court education finance decisions by tracking expenditures in the panel year immediately following the court decision, in the subsequent panel year, and in the years after that. We differentiate, as previously, between equity-based and adequacy-based decisions, as well as those issued by

Page 15 of 26

The Impact of State Supreme Court Decisions on Public School Finance

lower trial courts. We also substitute (state-level) average per capita income for the income inequality terms specified previously. Results are reported in Table 4.

In the analysis of inequality the various models we estimated yielded very similar results. They were also quite consistent with those obtained previously by Murray et al. Here there is more diversity in the results, as well as noticeable differences between Murray et al.′s findings and our own. Although they conclude that state legislatures responded to equity-based court decisions by increasing overall per pupil expenditures, there is little evidence of leveling up here. The only coefficient of the equity-based terms that is large enough to be statistically significant is the t + 2 panel in the first differences equation, and across all the equations there are nearly as many coefficients that have a negative sign as positive. Our findings thus indicate that equity-based court decisions have not systematically affected per pupil expenditures one way or the other. Litigation that has not advanced beyond the trial court level also has no discernible impact upon expenditures levels.

In contrast, there is considerable evidence indicating that adequacy-based decisions, which were almost all handed down after the time frame of Murray et al.′s study, did induce the increases in average expenditures predicted by the leveling-up hypothesis. All but one of the coefficients of the adequacy-based decision terms in the first three equations, as well as the crucial t + 1 coefficient in the first differences equation, are large, positive, and statistically significant. Ranging from 203 to 519, these coefficients point to increases in per pupil expenditures in the states covered by adequacy-base decisions that range from 4% to 10% of average expenditures during this time period.

Another difference between these results and those obtained in the inequality analysis is that many of the control variables register strong effects upon average per pupil expenditures. The large negative coefficients of the 5-17 cohort term in all four equations imply that per pupil expenditures are higher (or grow more rapidly) when there are relatively fewer pupils to spend money on. This finding also suggests that there may be considerable sluggishness in the response to enrollment increases or decreases. The state ideology terms in the random effects and AR(1) equations are also large and significant, indicating that more liberal states provide higher average levels per pupil expenditures. The strongest effects, however, are associated with real personal income, which loom large relative to the effects associated with court decisions. Figure 2, which plots state average per pupil expenditures during this time period against state average personal income, shows how strongly linked these variables are in the cross- section (Pearson′s r = 0.89). The temporal association is just as

Table 4. The Effect of State Supreme Court Reform Mandates upon Per Pupil Expenditures, 5-Year Panels

Variable Fixed Random AR(1) First

Effects Effects Differences

Equity decision, t + 1

-42.18 -104.65 -1.72 223.77

(171.88) (175.50) (168.04) (154.37)

Equity decision, t + 2

253.50 177.57 111.42 293.90*

(176.18) (176.85) (219.73) (156.32)

Equity decision, t + 3 plus

124.11 81.45 -30.92 -60.76

(137.48) (149.64) (209.80) (92.09)

Adequacy decision, t + 1

330.45** 354.81** 239.73* 202.71*

(123.28) (124.29) (113.86) (110.93)

Adequacy decision, t + 2

249.98* 273.32* 256.51 24.24

(136.17) (146.36) (159.29) (141.24)

Adequacy decision, t + 3

372.47** 519.05** 444.27* 45.09

Page 16 of 26

The Impact of State Supreme Court Decisions on Public School Finance

plus (112.68) (144.44) (221.39) (110.03)

Trial court decision

15.95 39.67 60.16 119.71

(87.26) (155.66) (124.56) (81.78)

State ideology -2.33 8.82** 14.04** 1.67

(3.38) (2.96) (3.54) (2.71)

Percent 5-17 -102.51* -82.65** -98.61** -136.42**

(48.98) (17.08) (32.65) (49.84)

Percent over 65

93.22* 69.06* 55.49 -28.67

(46.43) (31.39) (38.12) (73.42)

Personal income

0.22** 0.25** 0.23** 0.15**

(0.03) (0.01) (0.03) (0.04)

Constant 2616 586 1178 101

(1694) (728) (1290) (112)

R 2 0.95 0.84 0.53 0.4

Rho 0.72

N 368 368 368 322

*p < 0.05, ** p < 0.01.

strong. Multiplying the coefficient of the real personal income term, (which averages 0.22 across the four equations) by the growth in real personal income experienced in states over the 1972-2007 time period (67%) implies that over 75% of the growth in per pupil expenditures can be attributed to growth in real incomes. Supreme Court adequacy- based decisions, as indicated above, boosted per pupil expenditures in the states subject to these decisions by an additional 10% at most.

The data we have analyzed so far are state-level measures derived from individual school district data collected by the Census of Governments at 5-year intervals. Fortunately, we also have available state-level expenditure data collected annually by the National Center for Education Statistics (NCES). Re-estimating the Table 4 equations with annual data allows us to conduct another important check on the robustness of the results we have so far obtained. The NCES annual time series data run from 1969 to 2009. In using these data we must necessarily be more

Page 17 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Figure 2. State per Capita Personal Income and per Pupil Expenditures.

concerned with the problem of serial correlation, but it also means that with 41 years of data from 46 states we now have 1886 observations and a potentially important increase in statistical power.

A modification that we make in switching over to the annual data is to change the way in which we estimate the temporal impact of supreme court finance reform decisions. Results of both the inequality and expenditure analyses indicate that to the extent such decisions produce a policy response, that response is evident in the first subsequent panel of data. We thus focus on the years immediately following equity-based and adequacy-based court decisions, and so the time indicators we specify are for the first post-decision year, the second post-decision year, and the third and all subsequent post-decision years. As before, because school years start on July 1 of the previous calendar year, we assume that if a court decision occurs in calendar year t, any funding reforms made in response to it could not take hold until school year t + 2 at the earliest. The year indicators are thus denoted as t+ 2, t + 3, and t + 4 plus, respectively.

As in Table 4, coefficients associated with equity-based decisions are small and generally cannot be distinguished from zero. The two coefficients that are statistically significant, that is, the t + 4 plus terms in the fixed and random effects models, achieve this status primarily because of the much larger number of observations upon which they are based. They are actually only slightly larger in magnitude than the corresponding coefficients in Table 4. As was also the case in the 5-year panel analysis, all coefficients of the adequacy-based decision indicators in the first three equations are large and statistically significant. They are smaller than

the corresponding coefficients in Table 4, but only slightly so. In the fourth equation, the first differences model, the effects of adequacy-based decisions are muted. In Table 4 the coefficient of the t + 1 indicator was large enough to be significant, but here it is considerably smaller and falls short of conventional levels of statistical significance.

On the other hand, there is a suggestion in this equation of a cumulative effect over time, as the t + 2, t + 3, and t + 4 plus coefficients sum to 144.6. This is smaller than the effects registered in Table 4, but not negligible. The size of the coefficients also increase from one year to the next in the first three equations-particularly in the AR(1) model. The pattern we observe in the more finely grained annual data, then, suggests a slightly different inference than we had reached previously. The policy response to adequacy-based court decisions is clear, but it is not necessarily instantaneous. This could be because it takes new funding formulae a few years to take hold, or because some state legislatures react to court decisions more rapidly than others. In any case, the impact of adequacy-based supreme court decisions upon average per pupil expenditures appears to take at least a few years to be fully felt.

Page 18 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Results in Table 5 also closely track those in Table 4 in terms of the effects associated with the demographic cohort variables and personal income terms. In all equations the results indicate that per pupil expenditures increase with gains in personal income and with shrinkage in the relative size of the school age cohort. On the other hand, coefficients of the state ideology measure in Table 5 do not suggest, as did two of the coefficients in Table 4, that more liberal states provide higher levels of educational expenditures.

Our analyses thus indicate that the effects of finance reforms undertaken in response to equity-based supreme court decisions differ from those associated with reforms made in response to adequacy-based decisions. These differences, moreover, are in line with what we had hypothesized. State policymakers responded to equity-based decisions by implementing reforms that reduce cross-district inequalities. These reforms did not lead to either overall increases or overall decreases in average per pupil expenditures. The second wave of court decisions based upon considerations of adequacy, in contrast, has not produced measurable changes in the extent of cross- district inequality in expenditures, but has led to higher overall levels of funding for public education. The large secular increase in per pupil expenditures over the past several decades, however, is much more the product of growth in personal incomes and a decline in the relative size of the cohort of school-age children than it is of court- ordered finance reforms.

6. Educational Finance and Educational Quality

Although the California Supreme Court ultimately mandated that per pupil expenditures be almost completely equalized across school districts,

Table 5. The Effect of State Supreme Court Reform Mandates Upon Per Pupil Expenditures: Annual Panels, 1969-2009

Variable Fixed Random AR(1) First

Effects Effects Differences

Equity decision, t + 2

-83.32 -117.81 26.23 90.17

(161.76) (152.51) (78.37) (69.25)

Equity decision, t + 3

53.23 -15.76 46.43 111.78

(161.17) (152.59) (93.37) (69.21)

Equity Decision, t + 4 plus

163.34** 152.31** 127.54 14.83

(62.62) (53.64) (91.21) (15.75)

Adequacy decision, t + 2

279.35** 299.17** 155.10** 67.78

(113.09) (104.69) (57.51) (43.99)

Adequacy decision, t + 3

316.91** 313.40** 264.77** 58.29

(112.47) (104.86) (77.01) (44.22)

Adequacy decision, t + 4 plus

308.68** 402.94** 308.80** 18.58

(33.45) (45.76) (87.69) (14.58)

Trial court decision

34.52 93.44 95.85 41.88*

(38.58) (70.27) (69.60) (19.93)

State ideology -7.04** -0.06 1.35 -0.14

(1.51) (1.17) (1.30) (0.69)

Page 19 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Percent 5-17 -94.45** -102.26** -140.80** -69.24**

(22.41) (7.14) (25.20) (21.82)

Percent over 65

87.71** 62.07** 47.90 -108.66**

(20.48) (15.43) (26.70) (37.62)

Personal income

0.22** 0.25** 0.19** 0.08**

(0.02) (0.00) (0.02) (0.01)

Constant 2396** 1529** 3516** 243.7**

(774) (330) (893) (5.6)

R 2 0.95 0.82 0.42 0.31

Rho 0.90

N 1886 1886 1886 1840

*p < 0.05, **p < 0.01.

this was not necessarily the outcome that Coons et al. (1970) and other reformers sought. Their objective was instead to achieve fiscal neutrality, which is to say that the quality of education that students receive in public schools is not to depend upon the wealth of the district in which they resided. According to Downes (1992), this goal was not achieved. District wealth and quality of education, as measured by student performance on standardized tests, were as highly correlated in California a decade after the Serrano equalization reforms as they were before. It could be, of course, that the effects of finance reform on student performance take longer than a decade to manifest themselves. Once more revenue starts flowing to previously poor, low-expenditure school districts, it might take many more years to enhance the reputation of the school district, to hire and retain better teachers, to increase student morale, and to raise the expectations of students, teachers, and parents. Downes′ negative findings might be a consequence of not allowing the funding equalization

experiment to run for a long enough period of time. It behooves us to check for this possibility.

As in the other analyses we have undertaken, the passage of time should work in our favor. In order to determine if there is a long lag in the effect of education finance reform upon educational outcomes we need only repeat Downes′ work with data collected many years later. For a variety of reasons, however, we cannot build upon his data and analysis in the same nearly seamless way we were able to do in replicating Murray et al.′s (1998) research. Downes bases his analysis upon the performance of sixth-graders on the now defunct CAP test. The more recent data we have are Academic Performance Index (API) scores for the year 2000. These scores are based upon the performance of all students in the district. 8 District CAP scores ranged from about 100 to 400 with a mean initially set at 250, whereas district API scores in 2000 ranged from 420 to 900, and averaged 670. Secondly, Downes confined his analyses to 170 unified school districts in California, but we have 324 such entities. Thirdly, Downes′ regressions include only district wealth and income measures. We are able to specify other variables that significantly affect test scores, most notably the ethnic composition of the district′s students and the percentage of students with greater academic need. This means that whereas our analysis is similar to Downes′, it cannot be characterized as a replication and we cannot directly compare the regression coefficients that we obtain with his. The regression equation that we estimate takes the following form:

8 District API scores for 2000 were calculated by taking the weighted average, based upon enrollment, of all schools within a district.

Page 20 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Yi = α + βXi + γZi + εi (5)

where the dependent variable Yi is the school district API score for all unified school districts in California in 2000. The Serrano reforms have largely severed the link between local property values and school district expenditures, and because of Proposition 13 real-estate values are not reassessed in a regular, systematic way. We therefore use income instead of housing prices to measure district wealth, and so specify Xi as the median household income in school district i, in thousands of real (1992) dollars, at the time of the 2000 Census. If district wealth and educational outcomes are no longer linked, β = 0. Zi is a battery of control variables that include the following: (1) the percentage of students in the school district who are from ethnic minority populations, that is, black or Hispanic; (2) the percentage of students in the district who have an Individualized Education Program (IEP), which is provided for students who qualify for special (and more costly) education services; and (3) indicator variables for urban and for rural school districts. Districts with higher percentages of minority and IEP students are expected to have lower API scores, and both urban and rural districts are expected to

have lower scores than districts located in suburban areas. 9 Results are presented in the first column of Table 6.

The large, positive, and significant coefficient of the median income term mirrors Downes′ finding of a strong association between district wealth and average test scores. The 3.05 coefficient implies that an additional $10,000 of household income adds 30 additional points to the district′s API score. District median incomes ranged from $17,000 to $125,000 at this time, which means that the difference in income between the richest and the poorest districts yields a difference of 325 points in API scores. This is over two-thirds of the range we observe in these scores. Fiscal neutrality, by this metric, remains far out of reach. Coefficients of the minority and IEP percentage measures are in the expected negative direction, and are also statistically significant. Those of the urban and rural location indicators do not differ significantly from zero.

As indicated above, we cannot directly compare the regression coefficients in this equation with those obtained by Downes. We can, however, compare correlational measures. Downes reports that in 1986 the Pearson r correlation between his district income measure and district CAP scores was 0.52. The correlation we observe in the 2000 data between district income and test scores is considerably higher-0.76. Some caution must be exercised in making inferences on the basis of correlation coefficients (see Achen 1977), but Downes′ data from 1986 and ours from 2000 are too similar for the Pearson r values to lead us too far astray. In our view it is safe to conclude that in the 15 years following Downes there was no weakening in the linkage between district income levels and student performance, and this linkage may well have grown stronger.

As we argued previously, however, using students′ test scores as a measure of educational quality is problematic. How students perform on tests is an educational output measure, and is thus a product of their academic endowments, family and community environments, peer effects, and other factors that are not supplied by the school district. District wealth and test scores could thus be highly correlated even if the educational offerings of each district are identical. A potentially better way to investigate the matter of fiscal neutrality is to analyze quality indicators that are educational inputs, and which should therefore be a function of the financial resources available to school districts. There are two measures of educational quality that we can employ, and they both pertain to the teaching staff: the average number of years that the teachers in the district have been teaching, and the percentage of teachers that have attained an

Table 6. Academic Performance and District Wealth in California

9 There are other variables that are also highly predictive of district average API scores, including the percentage of students who qualify for free or reduced-price lunches and the percentage of students classified as English Language Learners. These variables, however, are so highly correlated with the racial/ethnic composition measure that they cannot be included in the same regression analysis.

Page 21 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Variable District District Percent of Average

API Score API Score Teachers with

Years of

a Master′s Teacher

Degree Service

or Higher

Median income (Thousands of

3.05** 2.83** 0.31** -0.01

1992 dollars) (0.17) (0.19) (0.04) (0.01)

Percent minority -2.16** -2.24** 0.03 -0.04**

(0.11) (0.12) (0.03) (0.00)

Percent IEP -1.40 -1.57* 0.35 0.04

(0.79) (0.78) (0.20) (0.04)

Urban district 3.35 1.33 3.02 -0.02

(6.24) (6.21) (1.55) (0.28)

Rural district -7.51 -7.41 -0.61 -0.23

(6.33) (6.27) (1.58) (0.29)

Percent of teachers with a 0.66**

master′s degree or higher

(0.23)

Average years of teacher service

-1.35

(1.25)

Constant 631** 645** 8.94* 14.91**

(14) (23) (3.47) (0.63)

N 324 324 324 324

R 2 0.82 0.83 0.21 0.21

*p < 0.05, **p < 0.01

advanced (post-baccalaureate) degree. As indicated above, previous research on education production functions indicates that teachers with master′s degrees produce gains in student achievement only when the degree is in the subject area they teach. The impact of experience, similarly, seems to be fully realized within the first 5 years of teaching. Unfortunately, we do not have these more fine-grained data available, and must proceed with what are admittedly fairly coarse measures.

The second column in Table 6 reports the results of a regression in which these two indicators are added to the right-hand side of the equation we previously estimated. These results indicate that school districts with higher percentages of teachers with post-baccalaureate degrees have higher API scores. This effect is not a powerful one, but impressive given that this equation also includes the median income, percent minority, and percent IEP variables that are so strongly associated with API scores. Having more experienced teachers is not predictive of student performance in this equation. This null finding may be the consequence of aggregation. Although the

Page 22 of 26

The Impact of State Supreme Court Decisions on Public School Finance

percentage of teachers with post-baccalaureate degrees ranges broadly from zero to 64%, the average years of teacher service has much less variance. In most school districts this figure is within 2 years of the mean of 13 years.

The next step we take in the analysis of fiscal neutrality is to determine the extent to which teacher quality-a key educational input supplied by school districts-is reflective of district wealth. The coefficients reported in the third and fourth columns of Table 6 are thus obtained from regressing the two teacher quality measures upon median household income and the other control variables specified previously in the first two equations. A positive coefficient of the income term would indicate that equalizing per pupil expenditures, as the Serrano reforms have done, has not severed the link between district wealth and educational quality.

The large (0.31) and significant coefficient of this term in the regression pertaining to the percent of teachers with advanced degrees (Column 3) is precisely what we observe. Wealthier school districts have teaching staffs with considerably higher levels of education. The regression on average years of teacher experience (Column 4) shows that wealthier districts are not, on average, staffed by teachers with more years of experience. Districts with larger percentages of minority students, however, are served by less experienced teachers. Richer districts are more successful in attracting teachers with advanced degrees, and districts with larger minority populations end up with less experienced teachers.

It should be noted that the regression equations that focus on educational inputs (teacher characteristics), like the equations that focus on outputs (test scores), are also made problematic by the presence of reciprocal causation (Hoxby 2001). More district wealth can produce better schools that do a better job educating children, who thus do better on standardized tests. But perceptions of district quality are in turn capitalized into housing prices, and districts associated with higher test scores thus attract higher-income families who can afford to live there. The perception that a school district has higher quality teachers may additionally attract higher-income families, which in turn may attract better teachers. Parents and their children-and so, too, teachers-thus sort themselves into school districts along income and ethnic lines. The strong positive associations between district wealth, teacher characteristics, and student educational attainment are equilibrium results, not causal relationships. Nevertheless, the results of these analyses indicate that a full generation after the Serrano equalization reforms, this self- reinforcing equilibrium had not been disturbed. Indeed, as we noted earlier, if we choose to dispense with regression results and look only at correlational measures, it appears that over time the goal of fiscal neutrality has slipped even farther away.

7. Conclusion and Prognosis

Beginning with Serrano v. Priest in 1971, equity-based finance reform decisions issued by state supreme courts have led to a decrease in cross-district inequality in per pupil expenditures. They have not produced

overall increases (or decreases) in average expenditure levels in the states affected by these decisions. By the end of the 1980s litigation of this nature had run its course, but in subsequent years many state supreme courts overturned existing systems of public school finance for failing to provide adequate education to students living in poor school districts. Adequacy-based decisions have not produced measurable changes in the extent of cross- district inequality in per pupil expenditures, but have led to higher overall levels of funding for public education. The dramatic nationwide increase in per pupil expenditures over the past several decades, however, is largely the product of growth in personal incomes and a decline in the relative size of the cohort of school-age children, and not of court-ordered finance reforms.

Advocates of public school finance reform have achieved some victories in state supreme courts. State legislatures have implemented changes in funding regimes to comply with court mandates. But if the ultimate goal of public school finance reform is fiscal neutrality in the provision of education, this litigation has achieved little or nothing. In California, decades after Serrano and the most far-reaching equalization reforms implemented anywhere in the country, the association between the wealth of a school district and educational quality remains strong and persistent. If one′s concern is the quality of education that students receive and not the amount of money spent on them, the victories that reformers have won in the courts have been, to paraphrase Rosenberg, hollow victories.

Page 23 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Is there a way forward from here? As long as public schools are embedded in the matrix of geographically defined school districts, it seems to us that the answer is no, and that the goal of fiscal neutrality in public education remains a pipe dream. But it is possible today, and increasingly more common, to provide elementary and secondary education online in "virtual" school districts that transcend school district boundaries. Enrollment in virtual school districts is growing rapidly, and today nearly two million K-12 students take at least one class online. In principle, advances in online educational methodologies allow teachers to develop customized curricula for their students, students to quickly and frequently receive feedback from their teachers, and parents to more effectively monitor the performance of their children as well as their children′s teachers. The internet also allows students to take honors classes, AP classes, and language classes not otherwise offered by their local school district. Moe and Chubb (2009) characterize these rapidly developing internet-based technologies as "liberating learning."

It would be a mistake to see online learning as a panacea for what ails primary and secondary education in the United States today. Several states have encountered problems with the implementation of online classes, and at this point it is hard to tell whether these are the growing pains inevitably experienced with developing technology, variations on the

same problems that trouble brick and mortar schools, or a new class of problems (NBCNews.com 2012). In any case, a major feature of online learning in virtual school districts is that it doesn′t matter where students or parents or teachers are physically located. Fiscal neutrality in public elementary and secondary education might best be achieved, then, when school district boundaries no longer matter. If equalization of educational opportunities is to occur, it appears that it may well be through the Cloud and not through the courts.

Appendix: Data Sources

Inequality measures for the 1972-92 panels were provided to us by the authors of Murray et al. (1998). We supplemented their data with school district data gathered in 1997, 2002, and 2007 by the Census of Governments. These data are available at the US Census Bureau′s website on State and Local Government Finances. As indicated above, we followed Murray et al.′s procedures in calculating state-level inequality and average per pupil expenditure measures for the 5-year panels, and used STATA routines (Jenkins 2004) to estimate all equations.

Annual data on education expenditures and average daily attendance (ADA) in each state are from the Digest of Education Statistics, published by the National Center for Education Statistics. The deflator used to convert expenditure figures into constant (1992) dollars was the state and local government price index, reported in the Bureau of Economic Analysis National Income and Product Accounts Tables. The BEA is also the source of per capita income data, which can be found at their website under the section on Annual State Personal Income.

The dynamic ideology measure for each state is that developed by Fording (2010). Population and age cohort data for each state are from US Census Bureau′s website on Population Estimates. Gini coefficients for state income up through 2005 are those reported by Guetzkow et al. (2006) and in the US Census Bureau publication "Income, Earnings, and Poverty Data from the American Community Survey," for all subsequent years.

California′s API data for 2000 were not yet reported at the school district level, but they were reported for individual schools. Thus the school district API data for 2000 were calculated as the weighted average of the API scores for all of the schools in a unified district, using the number of students in each school as the weight. API scores are available from the California Department of Education. Data on the demographics of students in California school districts were taken from the National Center for Education Statistics Common Core of Data. School district wealth measures are from the 2000 Census. Teacher quality measures are from the California Department of Education.

References

Achen, Christopher. 1977. "Measuring Representation: Perils of the Correlation Coefficient," 21 American Journal of Political Science 805-15.

Angrist, Joshua, and Jörn-Steffen Pischke. 2008. Mostly Harmless Econometrics: An Empiricist′s Companion. Princeton: Princeton University Press.

Page 24 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Baicker, Katherine, and Nora Gordon. 2006. "The Effect of Mandated State Educational Spending on Total Local Resources," 90 Journal of Public Economics 1519-35.

Beck, Nathaniel. 2001. "Time Series-Cross-Section Data: What Have We Learned in the Past Few Years?" 4 Annual Review of Political Science 271-93

Brace, Paul, Melinda Hall, and Laura Langer. 2001. "Placing State Supreme Courts in State Politics," 1 State Politics and Policy Quarterly 81-108.

Brunner, Eric, and Jon Sonstelie. 1996. "Coping with Serrano: Voluntary Contributions to California′s Local Public Schools," Proceedings of the Eighty-Ninth Annual Conference of the National Tax Association 372-81.

Buddin, Richard, and Gema Zamarro. 2009. "Teacher Qualifications and Student Achievement in Urban Elementary Schools," 66 Journal of Urban Economics 103-15.

Clotfelter, Charles, Helen Ladd, and Jacob Vigdor. 2007. "Teacher Credentials and Student Achievement: Longitudinal Analysis with Student Fixed Effects," 26 Economics of Education Review 673-82.

Coons, John, William Clune III, and Stephen Sugarman. 1970. Private Wealth and Public Education. Cambridge: Harvard University Press.

Corcoran, Sean, William Evans, Jennifer Godwin, Sheila Murray, and Robert Schwab. 2003. "The Changing Distribution of Education Finance: 1972-1997," Russell Sage Foundation. Available at www.russellsage.org/programs/proj_reviews/si/revcorcoran01.pdf.

Darling-Hammond, Linda. 1999. Teacher Quality and Student Achievement. Seattle, WA: Center for the Study of Teaching and Policy, University of Washington.

Downes, Thomas. 1992. "Evaluating the Impact of School Finance Reform on the Provision of Public Education: The California Case," 45 National Tax Journal 405-19.

---. 2004. "School Finance Reform and School Quality: Lessons from Vermont," in John Yinger, ed., Helping Children Left Behind: State Aid and the Pursuit of Educational Equity. Cambridge: MIT Press.

Evans, William, Sheila Murray, and Robert Schwab. 1997. "Schoolhouses, Courthouses, and Statehouses after Serrano," 16 Journal of Policy Analysis and Management 10-31.

Fernández, Raquel, and Richard Rogerson. 1999. "Education Finance Reform and Investment in Human Capital: Lessons from California," 74 Journal of Public Economics 327-50.

---. 2003. "Equity and Resources: An Analysis of Education Finance Systems," 111 Journal of Political Economy 858-97.

Fording, Richard. 2010. "State Citizen and Government Ideology," Available at: www.uky.edu/_rford/stateideology.html.

Freelon, Rhoda, Melanie Bertrand, and John Rogers. 2012. "Overburdened and Underfunded: California Public Schools Amidst the Great Recession," 2 Multidisciplinary Journal of Educational Research 152-76.

Gerber, Elisabeth, Arthur Lupia, Mathew McCubbins, and D. Roderick Kiewiet. 2000. Stealing the Initiative: How State Government Responds to Direct Democracy. Upper Saddle River, NJ: Prentice Hall.

Guetzkow, Joshua, Bruce Western, and Jake Rosenfeld. 2006. "State-Level Data on Income Inequality 1963-2003," Russell Sage Foundation, program on the Social Dimensions of Inequality. File income_smooth.txt. www.inequalitydata.org.

http://www.russellsage.org/programs/proj_reviews/si/revcorcoran01.pdf

http://www.uky.edu/_rford/stateideology.html

http://www.inequalitydata.org/

Page 25 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Hanushek, Eric. 2008. "Education Production Functions," in Durlauf, Steven, and Lawrence Blume, eds., The New Palgrave Dictionary of Economics. Basingstoke: Palgrave Macmillan.

Hanushek, Eric, and Alfred Lindseth. 2009. Schoolhouses, Courthouses, and Statehouses. Princeton: Princeton University Press.

Hanushek, Eric, and Steven Rivkin. 2010. "Generalizations about Using Value-Added Measures of Teacher Quality," 100 American Economic Review 267-71.

Heckman, J. 1978. "Dummy Endogenous Variables in a Simultaneous Equation System," 46 Econometrica 931-59.

Hoxby, Caroline M. 2001. "All School Finance Equalizations Are Not Created Equal," 116 Quarterly Journal of Economics 1189-231.

Hoxby, Caroline M., and Ilyana Kuziemko. 2004. "Robin Hood and His Not-So-Merry Plan: Capitalization and the Self-Destruction of Texas′ School Finance Equalization Plan". No. w10722. National Bureau of Economic Research.

Jenkins, Stephen. 2004. "Analysis of Income Distributions," Stata Technical Bulletin. STB-48 sg104. http://www.stat.com/stb/stb48/.

Kiewiet, D. Roderick, and Kristin Szakaly. 1996. "Constitutional Restrictions on Borrowing: An Analysis of State Bonded Indebtedness," 12 Journal of Law, Economics and Organization 62-97.

Kollars, Deb. February 2, 2003. "A Labyrinth of Spending," Sacramento Bee. Available at www.sacbee.com.

Kremer, Michael. 1995. "Research on Schooling: What We Know and What We Don′t," 10 The World Bank Research Observer 247-54.

Leyden, Dennis. 1992. "Court-Mandated Changes in Educational Grant Structure," 47 Public Finance 229-47.

Long, David. 1973. "The Property Tax and the Courts: School Finance after Rodriguez," in Peterson, George, ed., Property Tax Reform. Washington: The Urban Institute.

Manwaring, Robert, and Steven Sheffrin. 1997. "Litigation, School Finance Reform, and Aggregate Educational Spending," 4 International Tax and Public Finance 107-27.

Moe, Terry, and John Chubb. 2009. Liberating Learning: Technology, Politics, and the Future of American Education. San Francisco, CA: John Wiley & Sons.

Murray, Sheila, William Evans, and Robert Schwab. 1998. "Education-Finance Reform and the Distribution of Education Resources," 88 American Economic Review 789-812.

NBCNews.com. 2012. "Online Schools Face Backlash as States Question Results," October 3.

Rice, Jennifer. 2003. Teacher Quality: Understanding the Effectiveness of Teacher Attributes. Washington, DC: Economic Policy Institute.

Rivkin, Steven, Eric Hanushek, and John Kain. 2005. "Teachers, Schools, and Academic Achievement," 73 Econometrica 417-58.

Roelke, Christopher, Preston Green, and Erica Zielewski. 2004. "School Finance Litigation: The Promises and Limitations of the Third Wave," 79 Peabody Journal of Education 104-33.

Rosenberg, Gerald N. 1991. The Hollow Hope: Can Courts Bring About Social Change? Chicago: University of Chicago Press.

http://www.stat.com/stb/stb48/

http://www.sacbee.com/

Page 26 of 26

The Impact of State Supreme Court Decisions on Public School Finance

Rubinfeld, Daniel L. 1995. "California Fiscal Federalism: A School Finance Perspective," Constitutional Reform in California: Making State Government More Effective and Responsive. Berkeley: UC Berkeley Institute of Governmental Studies.

Silva, Fabio, and Jon Sonstelie. 1995. "Did Serrano Cause a Decline in School Spending?," 47 National Tax Journal 199-216.

Springer, Matthew G., Keke Liu, and James W. Guthrie. 2009. "The Impact of School Finance Litigation on Resource Distribution: A Comparison of Court-Mandated Equity and Adequacy Reforms," 17 Education Economics 421-44.

Thro, William, and Craig Wood. 2010. "The Constitution Text Matters: Reflections on Recent School Finance Cases," 251 Education Law Reporter 510-32.

Verstegen, Deborah, and Richard King. 1998. "The Relationship Between School Spending and Student Achievement: A Review and Analysis of 35 Years of Production Function Research," 24 Journal of Education Finance 243-62.

Verstegen, Deborah, and Teresa Jordan. 2009. "A Fifty-State Survey of School Finance Policies and Programs: An Overview," 34 Journal of Education Finance 213-30.

West, Martin, and Paul Peterson. 2007. School Money Trials: The Legal Pursuit of Educational Adequacy. Brookings: Washington, DC.

End of Document

The Impact of State Supreme Court Decisions on Public School Finance

Reporter