William S. Koski
Stanford University School of Law
East Palo Alto Community Law Project

Henry M. Levin
Stanford University School of Education
Center of Educational Research at Standford

Paper presented at the Forum on the Challenges and Opportunities for Funding Public Schools Equitably: Twenty-Five Years after San Antonio Independent School District v. Rodriguez at San Antonio, Texas, October 15 & 16, 1998.

Preface

In 1973, the United States Supreme Court in San Antonio Independent School District v. Rodriguez handed down a landmark decision on public school financing that shaped the policy debate surrounding educational funding in Texas and around the nation. In a 5-4 decision, the Supreme Court Justices found the Texas public school funding system based heavily on local property taxes was constitutional since it provided a minimum level of education to all students. Furthermore, the Court’s actions determined that education was not a fundamental right of the United States Constitution and that the equal protection clause of the 14th Amendment was not applicable since wealth (or lack of it) was not found to be a sufficient justification worthy of protection under the Federal Constitution. As a result of this decision, an overwhelming majority of subsequent school finance cases have been referred to state courts where education under state constitutions is a fundamental right.

As the 21st century approaches, the issue of public school financing is no nearer resolution than in 1968 when Demetrio Rodriguez and a group of parents from the Edgewood Independent School District first filed suit in federal court. While the issues surrounding public school financing have evolved to include school choice/vouchers and the debate has shifted themes from equality to equity to adequacy, the underlying premise remains the same - how should public schools be financed to support high standards and democratic values for all students to experience?

As a result of Rodriguez, four major lawsuits involving the Texas Supreme Court, Edgewood I-IV (1985-1995) have resulted in a restructured school financing system that compensates property poor school districts for their inability to generate funds (i.e., heavy reliance on local property taxes) and redistributes funds gained from property rich school districts. These legal challenges in Texas have certainly narrowed the disparity that existed prior to Rodriguez.

Due to pressure exerted by the aforementioned cases, seven major school finance bills were debated and adopted by the Texas Legislature from the period after Rodriguez to present. The final result being the integration of Senate Bill 7 developed by the 73rd Legislature (1993) and House Bill 4 designed by the 75 Legislature (1997). The current Foundation School Program has equalized public school financing unlike any prior programs. The question remains: Does equalized funding result in adequate education opportunities? Koski and Levin provide an insightful analysis to address this question.

Abstract

This discussion examines three assumptions intricately tied to public school financing litigation and review major developments in fiscal equity that have taken place to affect strategies in addressing the issue of what constitutes adequate public school financing. To facilitate this examination of the issues, Koski and Levin focus on three questions to formulate their analysis of the impact and lasting effects of the Rodriguez case: 1) Do additional dollars spent on education positively affect outcomes?; 2) Does the standard of adequacy lead to a specific solution to fiscal equity and can courts and policymakers develop consensus for standards of an adequate education?; and 3) Will educational finance litigation improve equity? Hence, do court decisions translate into greater fiscal equity? Each of these questions responds to events of the past twenty-five years and focuses on issues of how effective were various strategies in gaining greater educational equity (or at least fiscal equity).

The conclusion is that money matters if it is funneled to the right sources. In addition, judicial recourse has positively affected the redistribution of educational resources among students.

Despite these conclusions, it remains to be determined whether the resulting equalized system of funding public education has produced a more adequate education for all students. As Koski and Levin note, "For reformers and policymakers, these lessons indicate a need to consider carefully how strategies of reform are linked to results. No doubt this is important. But making this link is no easy task and at the end of the day, values and politics will be the final arbiters of educational finance decisions" (p. 39). In so far as Koski and Levin are correct in their assessment regarding policy making as it relates to educational funding, in the end, policymakers must ask themselves when deliberating educational funding issues, what type of education (and resources necessary to achieve it) would they want provided for their own children? This is the ultimate litmus test for educational funding policymakers to address.

Introduction

To a visitor from outer space, democracy in the United States must appear to be puzzling. On the one hand, there are enormous freedoms and constitutional protections for all citizens. Electoral politics extends political participation and representation broadly. And the various arrangements among the executive, legislative, and judicial branches to maintain checks and balances within the government seems to be effective. But, in preparing our young for citizenship and economic participation, there exists a yawning chasm between reality and our democratic ideals. The educational investment to prepare the young to exercise their political efficacy and to meet economic success is highly unequal across the nation and within almost all states. Students fortunate enough to live in wealthy enclaves will have more spent on their education than students situated in poorer geographical entities, hardly a fair situation or one that seems democratic.

This highly unfair system was tacitly accepted until the sixties on the basis that it was simply a by-product of our venerable commitment to local control of schools and the rights of each community to have some discretion over how much it spent on the education of its own children. But in the 1960’s, the passage and enforcement of new federal and state laws began to challenge a large range of discriminatory and unfair practices premised on race and poverty. And within this broad trend, academics such as Wise (1965) and Coons, Clune, and Sugarman (1970) began to question the constitutional basis for existing systems of educational finance and their resultant inequalities. Their seminal ideas stimulated both state and federal challenges to the inequalities in educational funding and spending of which Rodriguez v. San Antonio Independent School District was the first major federal challenge.

In the case of Rodriguez, the U.S. Supreme Court overturned the decision of the Federal District Court that Texas’ method of financing education violated the Equal Protection Clause of the Constitution. The Court was not convinced that students in property-poor school districts were a "suspect classification" since it had been shown that the correlation between district property wealth and family income and race were far from perfect. Among the several other elements leading to rejection of the decision of the lower court were a skepticism that expenditure disparities resulted in damage to students and that state-imposed minimum expenditure levels failed to assure children an adequate level of schooling. Over the last quarter-century these themes have played out more fully in an extensive set of studies that have tried to measure the relation between school spending and school results and in a shift from an emphasis on equity per se to "adequacy" in school funding as a guideline to achieving greater equity. But, in addition, a larger theme has emerged questioning whether litigation has or can have a profound influence on equity in educational finance and educational outcomes.

Using Rodriguez as a backdrop, we have elected to review three assumptions that undergird present litigation for improving educational equity and adequacy, or at least its fiscal foundations: (1) additional dollars spent on education will improve educational outcomes; (2) courts and policymakers can develop universal standards for what is an adequate education; and (3) litigation will lead to improved equity in educational finance. In the next section, we will review the controversy over spending and educational outcomes; this will be followed by a discussion of the slippery issues that are embedded in using adequacy of funding as a criterion for gaining greater equity; and the presentation will conclude with what has been learned about the impacts of court decisions on educational finance.

Do Expenditures Make a Difference?

Virtually all of the challenges to existing funding of schools are predicated on the fact that local, property-tax wealth is an important predictor of what is spent on the education of children in each geographical location. When there is a large tax base of property-tax wealth behind each student, educational expenditures per-student are higher—often considerably higher—than when the property-tax base for each student is low. Presumably, the differences in expenditures translate into differences in educational opportunities and outcomes. So, there is a tacit assumption that differential educational expenditures make a difference in what happens to students. Specifically, those in schools with higher educational expenditures will get a better education. Arithmetically, higher tax bases not only translate into higher expenditures per student, but also lower tax rates for the same expenditure. It is this latter fact that has also generated concerns and legal challenge on fiscal equity in education. But, for the moment we will focus on the inequities in school spending.

It seems to follow from both experience and logic that expenditures make a difference in the quality of education. Higher salaries will attract better teachers; smaller classes provide opportunities for more student engagement and teacher attention more counselors and psychologists afford additional guidance for making educational decisions and addressing difficulties; more computers, software, and technology personnel enable more opportunities to use information technologies; longer school days allow more learning time and more course and extracurricular opportunities and so on. And, if we were to ask parents what schools they would want for their children, it is certain that they would prefer those with greater resources to those with less resources, other things equal. So, how could such a premise be questioned? In actuality it was social science that caused the skepticism—some would say the mischief.

The Early Evidence

Just prior to the advent of the equal protection challenges to financing education, James Coleman and Associates (1966) produced a voluminous report for the U.S. government on educational opportunity. That report was based upon the largest social science project ever undertaken (save the decennial U.S. census) with a goal of trying to uncover the determinants of scholastic achievement. To do that, Coleman collected data from several thousand schools, 70,000 teachers, and about 700,000 students and subjected these data to statistical analysis. Statistical measures of student socio-economic status and other family measures, school resources, and school practices that had been collected by questionnaire were used to predict student achievement. Somewhat surprisingly Coleman found that socio-economic measures such as parental education and occupation seemed to explain almost all of the variance in student achievement that could be explained statistically and that differences among schools in their resources were unimportant. The extensive publicity and public debate that followed the release of the report emphasized this finding, and Coleman’s view that socio-economic and racial integration of schools would do more to raise achievement—this was the period of enforcement of the Brown vs. Board of Education decision—than improving school resources. The fact that these findings were not merely based upon opinion, but upon what appeared to be scientific methods and large data sets was very persuasive.

Coleman’s methodology was attacked by economists for some basic flaws that would tend to overstate the role of family background and derogate that of school resources (Bowles and Levin 1968; Hanushek and Kain 1973). Nevertheless, the impressiveness of a massive report sanctioned by the federal government and based upon unparalleled data and the wide dissemination of its results created a new climate of opinion that reached policy-makers and the courts, most of whom did not understand the underlying statistics.¹ The easy accessibility of the findings of the Coleman Report took precedence over any quest to understand the statistical challenges to those findings.

For example, Coleman found that pupil-teacher ratios of schools did not seem to be related to student achievement. This finding was interpreted by many to mean that reducing class size would not make a difference in achievement. But, pupil-teacher ratios are not the same as class sizes for many reasons. First, some schools provide teachers with more preparation periods than others (e.g. five periods of class and two preparation periods vs. six periods of class and one preparation period) necessitating the presence of higher student-teacher ratios for any given class size. Second, some districts place considerable numbers of their teachers in special assignments outside of the classroom, while others do not. Third, some schools have very high concentrations of special education students who historically have been placed in small classes, reducing the pupil-teacher ratio, but not benefiting the regular students who are being tested in these surveys. Fourth, the large urban districts have considerable numbers of teachers who are hired full-time and placed in substitute pools to cover teacher absences. Other districts hire substitutes as needed for day-to-date assignments and do not retain such pools. In all of these cases the teacher-pupil ratio will not be a good indicator of class size. More specifically, inner city schools tend to have larger substitute pools, more special education students, and more teachers on special assignment, so low pupil-teacher ratios particularly understate class size for those schools.

We must bear in mind that almost all studies linking school resources to student achievement, use crude measures such as pupil-teacher ratio (or teacher-pupil ratio), but interpret their results as equivalent to changing class size. This interpretation is clearly invalid. Even when class size is measured directly, there may be serious problems. In general, the students with poorer performance are placed in smaller classes for remedial services. Thus, it would not be surprising to find that even after controlling for family background—students in smaller classes have lower achievement scores. Yet, it is not necessarily true that class size has no effect. The assignment practices themselves can make it appear that students in larger classes have higher achievement.

A similar problem arises when one looks at the relation between teachers’ salaries and student achievement. City schools typically have more problems than suburban schools that place additional demands on teachers and require higher salaries to get teachers of the same caliber. That is, teachers require a higher salary to work in more difficult circumstances than in more amenable ones. But, this means that higher salaries will be associated statistically with teachers who work in schools with large urban populations of "at-risk" students who have poor student achievement. The result is that there is a spurious correlation between low achievement and higher salaries even though the relation is not causal. From a policy perspective, paying higher salaries in either city or suburb should attract better teachers and increase test scores. But, the fact that statistical samples include a spurious correlation between low performance in urban schools and the higher salaries required to get teachers into the city schools diminishes the potentially positive relation between higher salaries and better results.

Sadly, a sophisticated analysis and understanding of what the Coleman results might mean was rarely found among policymakers. And the notion that resources and school expenditures have little or no impact on achievement received an additional boost when one of the original critics of the Coleman Report began to compile the findings of a range of studies aimed at explaining statistically differences in student achievement. In a number of publications, Eric Hanushek (e.g. 1979,1986, 1989) reported a lack of consistency in studies that attempted to link specific measures of educational resources such as teacher salaries, teacher experience, teacher education, and pupil-teacher ratio (averaged across schools or school districts) to average student achievement. He interpreted this lack of consistency to mean that simply spending more on schools or investing in such common practices as reducing teacher-pupil ratios or getting more experienced and educated teachers and paying higher salaries would not be likely to raise student achievement. He buttressed this claim by asserting that although student expenditures had risen massively in recent years, student achievement had not (Hanushek and Rivkin 1997).

Dollars Appear to Make a Difference

In the last five or six years, doubts have been cast on this pessimistic interpretation. Although educational expenditures have risen considerably, it was found that much of the rise was attributed to special education students rather than the typical student who is tested for achievement gains (Lankford and Wyckoff 1986). Further, a substantial part of the rising expenditures was invested in higher salaries to compete for female professionals whose salaries in other occupations had risen precipitously, with market forces requiring higher pay just to retain teachers of existing quality. Both of these sources of expenditure increase could be viewed as productive, even if they would not be expected to have much impact on average student achievement. But, according to a study by the Rand Corporation, there was indeed an increase in achievement and especially among those students who were most disadvantaged. Further, the rise in achievement for disadvantaged or at-risk pupils appeared to coincide with the concentration of increased resources on their education (Grissmer, Flanagan, and Williamson 1998).

Further support for the view that higher expenditures will tend to improve student achievement was found in a reanalysis of the studies employed by Hanushek. Using sophisticated methods from meta-analysis to summarize statistical results, Hedges, Laine, and Greenwald (1994) concluded that the results of the statistical studies were far from random. In fact, they found that per-pupil expenditures showed a consistent positive relation with achievement such that a 10 percent increase in per-pupil expenditures or about $ 500 in 1994 was associated with a 0.7 standard deviation increase in educational achievement, a rather large effect of going from the 50 th to the 76th percentile.

Most of the older studies relating school inputs to student achievement relied on average school characteristics such as average class size and average teacher characteristics rather than those that were specific to the student in terms of each student’s actual class size and teachers. Studies using pupil-specific measures emerged in the nineties and found considerable evidence of effectiveness of particular school inputs and district educational spending within states (Ferguson 1991; Ferguson and Ladd 1996).² In general, these studies find larger effects of school spending for at-risk students and low-expenditure districts, and they do not find effects as large as Hedges, Laine, and Greenwald (1994). For example, Ferguson and Ladd (1996) in their study for Alabama in 1990-91 find that a 10 percent increase in per-pupil expenditures was associated with a .36 standard deviation increase among districts, an average rise among districts from the 50th percentile to the 64th percentile.

Further support for the view that school expenditures make a difference were found in a landmark study by Card and Krueger (1992). However, this study did not use student achievement as the criterion of interest, but adult earnings. Using average student expenditure, pupil-teacher ratio, and average teacher salaries for the state of birth at the time of the schooling experience, they found that school expenditure and school resources were consistently related to adult earnings. In short, adults who had gone to schools in states with higher quality schools had higher earnings as a result of two effects of higher quality resources. Better school quality as measured by resources led to greater earnings for any level of educational attainment, but it also increased earnings through stimulating students to complete more school years. Card and Krueger (1996) also found similar results comparing adult earnings of African Americans who had moved to the north from southern states. As impressive as these results are, they have been challenged because although they have been replicated several times with state and district level data, they were not found to be replicated in a study that used school-level data (Betts 1996).

Many of the conclusions that school resources do not seem to be linked to student achievement are based upon the findings for teacher-pupil ratios. As noted above, although many authors tend to interpret teacher-pupil ratio as a surrogate for class size, this interpretation is highly questionable. Direct measures of class size are rarely to be found. One important exception is a study of the State of Tennessee which sponsored a large-scale experiment in which randomly selected elementary school classes were reduced from an average of about 23 students to about 15 students. Achievement of students in smaller classes was compared over several years with students in the larger classes and in larger classes that were assigned classroom aides. The results consistently showed better achievement for the small classes of almost a quarter standard deviation on average and about twice the effect for low socioeconomic students as for higher ones. These results are particularly impressive because they were produced by an experiment where class size was manipulated directly (not statistically), and where the results were reviewed by an eminent statistician (Mosteller 1995) and pronounced to be solid. They were later evaluated econometrically by Alan Krueger (1998) who also found the conclusions of the original study to be valid.

In a paper on the cost-effectiveness of class size reductions, Levin and DeAngelis (forthcoming) submitted the Tennessee results to a benefit-cost analysis. The costs are attributable to the need for more teachers and classrooms to accommodate a given population. Benefits were calculated on the basis of the increase in earnings associated with higher achievement scores. These increases in earnings occur directly because persons in the labor market with higher achievement statistically receive slightly higher earnings; it also occurs indirectly because higher test scores lead to more education which increases earnings. We found that the present value of additional earnings is about twice that of the additional costs associated with reducing class size at a discount rate of 3 percent, and costs and benefits were about equal at a 5 percent rate of discount. Given the fact that the costs are based upon an up-front investment and the benefits are based upon adult earnings received much later, this is a fairly remarkable result suggesting that reductions in class size are probably a worthwhile investment.

Summary of the Effects of Expenditures

What can we conclude from three decades of work on this subject?

• First, it is clear that there are many methodological obstacles to identifying the precise effects of school expenditures on student achievement and adult earnings. Therefore, studies should not be taken at face value, but should be analyzed carefully for their validity. Earlier studies were particularly crude with district and school level variables measured often in inappropriate ways. More recent studies provide student-level data on school resources and achievement and more evidence of school effects.

• More recent studies using student level data show apparent effects of school expenditures and specific school resources on student achievement that have substantial policy implications. This is also verified by major experiments with class size and computer-assisted instruction and (Ragosta, Holland, and Jamison 1982) and peer tutoring (Levin, Glass, and Meister 1987).

• These studies also suggest that expenditures and the educational improvements that expenditures can buy make a greater difference for low spending districts (Ladd and Ferguson 1996) and students from minority or lower socio-economic backgrounds (Grissmer, Flanagan, and Williamson 1998; Finn and Achilles 1990).

• Studies of the impact of school resources on adult earnings show the opposite pattern to those of student achievement. Aggregate studies using state data show much larger apparent effects than studies at the school district and school level.

• All of the participants in the debate including those most skeptical (e.g. Hanushek) agree that money can make a difference in the quality of schools and achievement. However, they differ on whether it makes much of a difference on the average in the present ways that it is being spent. The most recent research suggests that it does make a difference. But, all of the participants believe that it can make a much larger difference by making schools more responsive to incentives and creating schools that are generally more productive with existing resources (Hanushek and others 1994; Levin 1997).

This brings us to the bottom line. At the end of the day will greater equity in educational funding produce greater equity in educational outcomes? On balance the answer seems to be "yes," but with a much larger impact if schools also become more efficient. To check on the veracity of this conclusion one should interview researchers who address the educational expenditure issue to find where they send their own children to school. Without exception one will find that their children are in suburban schools that are relatively high in expenditures with relatively small class sizes or in private schools that are well-endowed with resources. Regardless of what their studies tell us, I would rely more heavily on what their educational judgements are for their own children than their findings for "other peoples’ children."

What are the Criteria for an "Adequate" Education.

Having determined that money does matter on balance, educational finance reform advocates, researchers, and policy-makers are still left with the question of determining how much money? For twenty years this question was largely avoided as the strategies for obtaining greater equity in educational finance attacked the inequalities and their sources directly. Questions were raised under existing state constitutions on how the state could meet its educational obligations when funding was so unequal from locality to locality. It was argued that existing systems of funding education with heavy reliance on local property taxes produced these inequalities, so that states needed to reduce or eliminate dependence on local tax sources when tax bases were so unequal. But, more recently there has been a shift to the notion of assuring that all children receive an adequate education at a high standard. Since 1989 the courts in some nine states (Alabama, Massachusetts, Kentucky, New Hampshire, New York, North Carolina, Ohio, Tennessee, and Wyoming) have embraced this criterion (Minorini and Sugarman 1998). This newer strategy has the advantage of assuring that equity is not achieved at a mediocre level of funding equality, and it eliminates a political source of opposition to educational funding reform by not threatening districts that are spending above this high minimum level.

A leader in the adequacy movement and a pioneer in the movement towards fiscal equity, William Clune (1994b) has defined adequacy as the achievement of high minimal outcomes for all students. He then proceeds to suggest that after defining high minimum goals, it is necessary to identify the forms of school organization, methods, and resources needed to achieve them and to place a cost on these resources and methods. The result of these steps would be to stipulate the educational expenditure level that would provide an "adequate" education for all students and that would integrate school operations with their funding. Presumably the high outcome standards required for an adequate level of education would require much higher expenditures among schools at the lower end of the spending spectrum, and this has certainly been the experience of states that have adopted adequacy standards (Evans, Murray, and Schwab 1997).

The use of adequacy as a criterion for improving educational finance is both intuitively and politically appealing, although not without considerable controversy on feasibility and details (Clune 1994a). It is attractive because it links funding to the production of high educational outcomes for all students. Indeed, it is possible to reform simultaneously the accountability systems for measuring what schools are producing, the organization and operations of schools to improve their effectiveness in reaching the high level outcomes for which they are being held responsible, and the funding of schools to enable schools to obtain the resources that are necessary to succeed. And, at a time when there is concern about the quality of the labor force and the competitive position of the U.S. economy in international competition, adequacy for high minimum standards is an attractive aspect. Further, by not threatening the funding of local educational agencies that meet adequacy standards, a powerful source of opposition to more traditional equity strategies is muted.

It should be noted that an adequacy standard was used traditionally as a basis for equalization funding by the states, although in a rather different manner than in its current advocacy. In the early twentieth century it was recognized that some local educational agencies had too limited a tax base to provide even a minimal educational provision for their students. The response was to create a system of equalization aid for local educational agencies based on a guaranteed amount of revenues below which expenditure could not fall. This was widely known as the "foundation plan" (Garms, Guthrie, and Pierce, 1978: 190-93). Districts that could not raise enough revenues at a mandatory tax rate to meet the foundation level would receive assistance from their state governments to reach that level. A major difference between the adequacy standard embedded in foundation plan funding and the current use of equity in educational funding is that the former was merely an attempt to meet a minimum level of schooling while the latter is designed to provide a high level of educational opportunity and outcome.

The first reference to using the criterion of adequacy as a basis for meeting the educational needs of particular students was established by Chambers and Hartman (1983). These researchers were concerned with developing a model to assist school districts and states to estimate the costs of meeting the needs under state and federal law of their special education students. They developed a resource cost model that begins by specifying the handicapping conditions of the population of special education students. This is followed by setting out the educational requirements and associated resource needs to meet the needs of these students as judged by experts on educational programs. Finally, the necessary resources are evaluated for their costs in order to set out an overall cost for meeting student needs. Chambers and Parrish also created the capabilities of looking at different approaches to meeting educational needs to ascertain if similar results might be obtained at lower cost. By placing all of this in an interactive computer model, they were able to incorporate the input of decision-makers directly in determining the resource needs for adequate and equitable programs for handicapped populations.

What Chambers and Parrish proposed for special education is essentially the format for proposals to fund schools on the basis of adequacy; that is, a method by which funding would be adequate to get all students to high levels of learning. This means that first a state must (1) define what is meant by educational adequacy; (2) determine what it takes in resources to meet adequacy standards as defined, and (3) translate resources into costs which will be funded to meet adequacy. Each of these appears to be amenable to research which might yield results to guide courts and legislatures. In fact, the knowledge base is highly inadequate to answer the questions posed by each of these criteria. Let us review each in turn.

The Meaning of Educational Adequacy

Certainly the term "educational adequacy" has a strong appeal to it. We can think of educational adequacy in relation to the competencies that adults require to be productive thinkers, workers, citizens, and parents (Inkeles 1966). This perspective on schools has a long tradition in sociology where schools are viewed as agencies of socialization for preparing competent adults. At this level of abstraction there can be considerable agreement. Minorini and Sugarman (1998) in a review of state cases using adequacy criteria point out especially the wide agreement by the courts on preparation of students for citizenship and the labor market. All students should be prepared to be competent adults. However, translating what it means to be competent and how to measure competencies creates a level of complexity and controversy that immediately makes definitions of educational adequacy far more problematic.

What specific competencies must be included in the high minimal outcomes for all students, and how should they be measured? It is appealing to view educational adequacy standards being set for many criteria and assessed by test scores. But, which subjects should be required in the curriculum and what level of competency is adequate? It is clear that we do not know precisely what educational competencies are strong predictors of adult performance (e.g. Levin 1998), and there is almost no data base of validation studies to assist us. Thus, the areas that are chosen and what constitutes adequate performance levels will be largely a matter of judgement. At the very least we might expect to look for competencies in communication such as reading, writing, speaking, and listening (interpretation of what is being said). Beyond this, there are a variety of subjects in mathematics, sciences, social studies, foreign languages, critical thinking, and so on. Bear in mind that even social studies includes history, government, geography, economics, psychology, sociology, and anthropology. Sciences include general aspects of scientific thinking and technological applications as well as biology, physics, chemistry, geology, and a host of other disciplines. Then there is knowledge of the arts, artistic expression, literature, and culture. Even these do not include the interpersonal skills necessary for effective functioning.

So, out of a large constellation of competencies it will be necessary to choose which particular ones have priority in the adequacy formulation and how we will measure adequacy for competence. Choices will have to depend heavily upon values in terms of what is important as well as pragmatic issues of what can be tested and costs. It will not be practical or even feasible in terms of the costs and time requirements for testing to provide 25 different tests of adequacy at each grade level, so much of the adequacy determination will have to be made through more limited testing and insuring curriculum coverage without a direct measure of many outcomes. It should also be noted that existing school testing programs are far more limited than even the most modest versions suggested by educational adequacy standards (Sternberg 1997).

Producing Adequacy

Even if we can get agreement on what are adequate educational outcomes, we need to specify how they are produced. At this moment education is more an art than a science in that we cannot predict precisely what conditions will produce which outcomes for a wide range of different populations. Studies that have attempted to do this even for a single subject and grade level have not shown consistent results, even if there is some evidence that some resources have shown positive results for some subjects as emphasized in the previous section. It is one thing to say that class size and better teachers and additional funding are associated with higher student achievement in reading or mathematics. It is quite another to ask what specific programs and educational processes produce adequate levels of educational outcome on a predictable basis for all students in all subjects and areas of competency. Clune (1994b) has suggested that the ambiguity in understanding how to produce adequacy can be resolved by giving schools leeway in selecting different approaches and assisting schools to adopt approaches that have shown promise according to the evidence of success. Presumably, these types of programs could be evaluated for costs, and those costs could be used as guidelines for making fiscal provisions.

This is not an unreasonable approach, but it is far from the original goal of specifying the resources required for approaches that have been shown to meet the adequacy standards. This is especially noteworthy given that virtually all evaluations of existing educational programs tend to be limited to one or two subjects at given grade levels rather than the multiplicity of competencies and their developments at all schooling levels. Other observers have simply concluded that adequacy should be determined less by educational outcomes than by "opportunities to learn." That is, experts might agree on which conditions are necessary in schools to provide learning opportunities for all students such as appropriate class sizes, teacher qualifications, curriculum, materials, and so on. And, these can be used as a basis for determining costs. While this has much to offer as a practical solution, it relinquishes the goal of high educational outcomes as adequacy standards.

The Costs of Adequacy

All of the foregoing is designed to derive the costs of reaching adequate outcomes for all students. But, it is well-known that all measurable educational outcomes depend not only on schools but on the experiences that students have in their homes, families, and communities. It is safe to say that there exists no set of school interventions that can fully compensate for differences in socioeconomic background. Less than 10 percent of a child’s waking hours between the ages of 0-18 years of age is spent in school, meaning that a child’s knowledge and proficiencies will be heavily dependent upon the 90 percent of time spent outside of the school. For schools to get all students to a high level means that considerable resources would have to be invested in those students who are exposed to less-enriched home and family environments. Of course, there is a long tradition of compensatory education, and fiscal weighting formulas for education often take into account the higher costs of education for children from economically disadvantaged backgrounds.

But, without a clear picture of what resources it takes to obtain adequate educational results for particular types of children, it is not possible to provide even roughly precise estimates of what additional funding is necessary for educationally needy children. Indeed, what the costs are for meeting adequacy standards for children without special needs cannot be readily calculated from our existing knowledge base for reasons set out above. But, even as we contemplate extra resources for educationally at-risk students, we need to confront the fact that depending completely on schools to meet adequacy criteria may be inappropriate. Children from economically disadvantaged families face greater obstacles to meeting needs for health care, housing, nutrition, pre-school enrichment, and material needs than other students, and this surely explains some of their educational deficiencies. Simply reducing class size or getting more qualified teachers is unlikely to be as efficient a response to meeting these needs as direct investment in them. At the same time, government investment in training and employment of unemployed or underemployed parents may have a stronger effect on educational performance of students than an equivalent effort in schools.

That is the costs of improving student performance and raising it to levels of adequacy may be only partially met through schools and may require other interventions which are complementary to school improvement or even superior ways of using some of the available resources to raise achievement. Some of the fiscal requirements for meeting adequacy in student proficiencies will require greater investments in schools, but good fiscal policy should also consider other types of investments in family support and well-being that will be necessary to enhance student performance. As all of the above suggest, the costs of meeting educational adequacy standards for all students is still more a matter of guesswork than empirical science.

Whither Adequacy?

As we began this section, we wish to reiterate that the concept of adequacy expressed as high minimum educational outcomes for all students is a worthy policy goal to pursue, and it seems to comport well with what many of the state courts believe that state constitutions embrace. What we have tried to emphasize is that the knowledge base for implementing adequacy leaves much to be desired. Although it clearly contains insights that might be a guide to policy, it does not contain solutions. Gaps in defining and measuring adequacy, specifying educational approaches to meeting adequacy standards, and stipulating costs that can be embedded in fiscal formulas are substantial. Minorini and Sugarman (1998:18) express it best in their succinct summary: "As we see it, what is ‘adequate’ will come down to a matter of judgment, probably informed by expertise."

Does Litigation Make A Difference?

The Challenge of Educational Finance Reform Through The Courts

But educational finance reform-minded courts face arguably greater barriers than these oft-cited capacity constraints because courts in educational finance cases most often do not get to design the remedial finance policy. Rather, educational finance reform plaintiffs often seek and courts order declaratory relief, not remedial relief. In a declaratory relief action, the court simply declares the rights of the parties in the suit and decides whether those rights have been infringed. The court does not order prospective, coercive relief as it does in issuing a remedial decree. Rather, the court declares the extant funding scheme unconstitutional, often outlines in vague and general terms what would make a constitutional funding scheme, and relies upon the legislature and executive to come up with a policy that will honor the principles and rights announced in the declaratory relief order.⁵ This, of course, is the rub. Courts do not necessarily face remedy-design limitations; rather, they face significant remedy implementation limitations that pit the judicial branch against the legislative and executive branches. This institutional tug-of-war raises questions of democratic control, separation of powers, and the role of the court as a countermajoritarian check on the political branches.⁶

In a judicially ideal world, the translation of court order to public policy would be seamless and accurate. In reality, however, because courts must rely on the political branches to implement educational finance reform, the courts in educational finance cases face a two-step implementation problem: (1) they must persuade the legislature to adopt a funding scheme that is consistent with the rights declaration, and (2) that policy must in fact affect the distribution of educational resources in the manner required by the rights declaration. To overcome the first of these implementation problems, the court must somehow get the legislature to act the way it wants it to act. This is no mean feat for an institution with almost no ability to provide incentives to or coerce the other institutions to act, excepting the largely symbolic threat of civil contempt. This enforcement problem is exacerbated by subtle forms of non-compliance. Although outright defiance of educational finance decisions is not unheard of,⁷ legislatures almost always "comply," to some degree, with the order by crafting a remedial educational finance scheme and avoiding the contempt sanction. The question is whether this "compliance" is so watered-down by the political process that it does not bring about meaningful change in school funding. Many political interests, including wealthy school districts and other groups with prior claim to the state treasury, will oppose shifts of funds to poor districts. Constituency and re-election pressures will weaken the resolve of even compliance-minded legislators. And what looked like a clear legal right and standard stated by the court will become fuzzy and unworkable on the floor of the state assembly. In short, Law in the judicial opinion will not look exactly like Law as passed by the legislature.

Of course the court always has the opportunity to disagree with the legislative remedy, strike down the remedial funding scheme, and send it back to the statehouse for another try. Some have viewed this "dialogue" between the judiciary and legislature as a logical, if not healthy, approach to reforming educational finance systems (Jaffe and Kersch 1991; Brown 1994). But this dialogic process is cumbersome, expensive, and possibly frustrating for all parties involved. As the former University of Texas at Austin President Mark Yudof eloquently put it, "school finance reform is like a Russian novel: it’s long, tedious, and everybody dies in the end" (Yudof 1991, p. 499). We need look no further than the iterative and sometimes disappointing process of reform here in Texas or the Dickensian school finance litigation in New Jersey. Ultimately, the dialogue must end and the legislature and the judiciary must reach agreement on what is a constitutional school funding scheme.

Getting the policy passed, however, is only half the implementation battle. The actual distribution of educational resources must reflect the judicial declaration of constitutional principles, i.e., the legislature and educational bureaucracy must implement the remedial policy in a fashion faithful to the judicial decree. The linkages between policy design and social outcome are myriad and provide many opportunities for slippage. For instance, technical or informational limitations may hamper the design and implementation of policy, or fiscal crises may dash the best laid plans, even when the policy-makers are well-intentioned. Indeed, state legislatures, such as Wyoming, have retained expert consultants to help them in designing remedial educational finance schemes. Such reliance on outside experts will become even more common as reformers, courts, and legislatures turn toward theories of finance adequacy, rather than finance equity, because designing an "adequate" educational system is viewed as a "technical" exercise. We have already discussed the difficulty of this exercise. The lesson there was that, due to technological and other limitations, ex post measures of equity and adequacy may not gibe with carefully designed funding schemes. Equally problematic is the long-run deterioration of any equity-minded educational finance plan (Brown and Elmore 1982). Such deterioration may result from shifting political priorities and policies, as well as localized flouting of the equity principle by those who informally bolster their own schools’ coffers with private support. In sum, even seemingly well-crafted legislative remedial policy may not result in compliance with the letter or spirit of the court’s decision.

In the two-step process of implementing educational finance decisions — policy design in the legislature and policy implementation by the state bureaucracy — there is much room for inarticulation. Given this inarticulation, one might suspect that judicial efficacy in the educational finance arena has been limited. Yet reformers continue to seek redress through the courts. Social science researchers must then ask the question: has litigation and court intervention brought about meaningful school funding reform in the past two and one-half decades? Here we explore the effects of litigation induced educational finance reform on three specific areas: per-pupil spending equity, taxpayer equity, and overall school spending.

The Effects of Judicial Intervention in Educational Finance Policy

In the wake of the first and second "waves" of educational finance cases in the early to mid 1970’s, researchers focused their attention on the effects of those litigations on various aspects of educational finance. Put most simply, these studies analyzed the effects of a simple independent variable, the legal command, on a dependent variable, some measure of the effect of or the compliance with the legal command. Perhaps no other effect of educational finance litigation has been as often studied as the effect of litigation on the "equity" of the distribution of school funds. This makes sense in light of the fact that the very object of most early reformers was the provision of an equitable educational finance system.

Assessing the impact of litigation on educational finance equity has proven difficult, however, for a number of reasons. First, the very measurement of equity is a value-laden and politically-charged yardstick (Berne and Stiefel 1984). In their seminal work on the measurement of equity in school finance, Berne and Stiefel (1984) identify a number questions that reflect the complexity of equity of concepts. Equity for whom — taxpayers or children? Although some school finance plans such as "tax base equalizing" may promote taxpayer equity, inter-district, per-student equity may remain compromised. Equity of what — dollars/resources, processes, achievement, or outcomes? Although recent policy attention has been focused on raising student achievement, traditional school finance reform and most school finance research has been directed at the seemingly more manageable input measures. What is the equity principle — horizontal equity (equal treatment of similarly situated persons), vertical equity (unequal treatment of differently situated persons), or equal opportunity (the absence of educational input differences among students as a result of factors beyond the control of the student such as race, gender, school district wealth, or geographic location)?⁸ And finally, what statistical measures best capture and quantify the chosen definition of equity? By their very nature, these questions cannot be answered with certainty or objectivity, as the choice among equity targets, objects, measures, and concepts is driven by values and cannot be dictated with vague admonitions such as "equal protection of the laws."

A second limitation of the research on the impact of judicial educational finance decisions is the small scale of many of the studies (many focus on only a handful of states) and the concomitant inability to generalize the findings of those studies beyond the highly contextual effects in those states. Third, although recent studies designed sophisticated controls for internal and external threats to validity, many early studies were plagued by the inability to isolate causality in the court’s decision. Did equity improve due to judges’ opinion or was it a booming economy, for instance (Dayton 1996)? Recognizing the many measures of equity and the limitations of judicial impact research in this area, we review the highlights and primary conclusions of the research that has analyzed the impact of litigation on finance equity.

Much of the early work on the impact of educational finance cases were non-comparative analyses of the distribution of school resources in those few states that had their finance systems overturned by court order. Moreover, much of that work suggested the limited efficacy of judicial involvement. Analyzing the effects of the Serrano and Robinson decisions in California and New Jersey respectively, William Clune found that the per-pupil spending gap between property poor and property rich districts narrowed, but did not close, after the legislatures in those states adopted remedial legislation (Clune and Lindquist 1979). He also found, however, that state aid to property poor districts increased at a much faster rate than state aid to property rich districts, suggesting that the primary effect of the litigation was local property tax relief (a subject matter to which we will return). Similarly, Richard Rosmiller (1986, p. 10), analyzing the educational finance cases to date, found that "[e]ven in states where litigation has been successful, the amount of reform achieved has not been startling" (Rossmiller 1986). These limited findings of efficacy were coupled with cautious pronouncements of the limits of court-ordered reform in the politicized arena of educational finance (Brown and Elmore 1982).

The tentative early inquiries were soon followed in the 1990’s by bolder findings on the ability of the court to influence funding equity. Richard Salmon and David Alexander, who undertook a qualitative and quantitative analysis of the state legislative responses to educational finance decisions (both those upholding and overturning educational finance plans), found that in those states in which the court found the educational finance plan unconstitutional, major structural changes were undertaken by the legislature, while only minor changes were undertaken in those states where the plan was upheld (Salmon and Alexander 1990). Moreover, inferring from the data on the percentage of state aid allocated to districts, Salmon and Alexander concluded that "it appears that substantial movement toward greater fiscal equity has occurred" in those states that had their finance plan overturned, while "movement toward greater fiscal equity has been considerably more modest" in those states whose finance plans were upheld (p. 269). Bradley Joondeph (1995) and Douglas Reed (1995) similarly found equity gains in the wake of judicial intervention in the states each analyzed. Joondeph found that in five states — Arkansas, California, Connecticut, Washington, and Wyoming — that had their school financing schemes ruled unconstitutional prior to 1984, spending gaps narrowed between poor and affluent districts (Joondeph 1995). (This "good" effect of judicial intervention was countervailed, however, by Joondeph’s "bad" and "ugly" findings that we discuss below). Similarly, Reed, using time-series data, also found that in each of Connecticut, New Jersey, and Texas equity increased in per-pupil spending after the court handed down its decision. He noted, however, that the size and the permanence of the equalizing trend varied significantly from state-to-state. Thus, Joondeph’s and Reed’s case studies suggest that judicial intervention can enhance equity, at least in the short term.

These case study findings were further bolstered by two sophisticated analyses of data from school districts in nearly all of the 50 states (Hickrod, Hines et al. 1992; Evans, Murray et al. 1997). What makes these latter studies persuasive is that they are better able to isolate the effects of the judicial intervention by comparing those states where courts have overturned school funding schemes to those states in which the scheme was upheld or no litigation has ever been present. Hickrod, Hines, Anthony, Dively, and Pruyne (1992) (hereinafter Hickrod) provide several independent variables for control purposes and a measure of finance equity — the coefficient of variation. The equity measures, however, are only provided for two time points, 1980 and 1987. The eight independent variables are based on litigation-related factors: (1) plaintiffs won at supreme court level; (2) plaintiffs won at supreme court level, but further compliance litigation was filed; (3) plaintiffs lost at supreme court and no further complaints filed; (4) plaintiffs lost, but new complaints have been filed; (5) litigation is present and no decision is rendered; (6) no litigation is present or case is dormant; (7) states in which the supreme court declared education a "fundamental right;" and (8) states in which supreme court declared education not a fundamental right. Each of these variables has a theoretical relationship (or non-relationship as it were) with one or more of the dependent variables. For instance, those cases the plaintiffs won and in which no further compliance litigation has been filed should affect the dependent variable in the direction the court intended, but the same cannot be said for those cases in which further compliance litigation was filed. Most important, the sixth category — no litigation is present or case is dormant — is a reasonable control group in that active litigation has never been present in the state and could never have directly affected the distribution of educational resources whether through a judicial opinion or response to a filed complaint.

Hickrod then divided the states among these independent variables and measures the change in the coefficient of variation (1980 and 1987). A major finding is that a slight though uneven reduction in inequity is related to plaintiffs’ winning cases, while losing does not affect inequity one way or the other and the control group showed a growing disparity in funding. Thus, it appears from the Hickrod study that the absence of litigation is related to growing inequity, but a plaintiff win may decrease inequity. These findings are somewhat favorable for those who seek reform through litigation.

Though an ambitious effort to eliminate alternative explanations of changes in the distribution of educational resources, this study (like many of the case studies) still fails to account for explanations internal to each state. For instance, the narrowing of inequity may have been effected by changes in the size of the student population, localized economic fluctuations, or per capita income in the state. In other words, although the no-litigation-present "control group" may address certain external validity threats such as national economic trends, there remain many threats to internal validity because the growth rates are all attributed to legal action, rather than other within-state causes. Similarly, because the study only measured the dependent variables at two points in time, one can never be sure that the judicial decision, as opposed to some other event, caused the change in the outcome variable. As Hickrod acknowledges, a time series of fiscal measurements is needed to reveal the connections between litigation and the outcome.

Evans, Murray, and Schwab (1997) (hereinafter Evans) address with time series data and econometric modeling a number of questions left open by Hickrod’s study. Specifically, Evans addresses the question of whether within-state inequality is different in those states in which there was successful education finance reform litigation from those states in which there was no reform or there was legislative reform without preceding court-ordered reform. Using data on revenues from 16,000 K-12 school districts in 47 states (Hawaii was excluded because it was fully state funded, while Montana and Vermont were dropped because they have almost no unified school districts) over the 1972-92 time period, the authors employed several methods to determine the effects of litigation and reform on inequality.

First, using a decomposed Theil index, the authors determined that roughly two-thirds of the variation in spending is a between-states effect. Thus, litigation directed solely at equity within-state can address only a third of the inequity nationwide. (Unfortunately, this is another legacy of Rodriguez: the inability to rely on federal law as a means of bringing about greater inter-state finance equity). Second, the authors diagrammed the log of the ratio of the 95th and 5th percentile of per pupil expenditures for two time points, 1972 and 1992. In doing so, they compared the states where there was no litigation or litigation was unsuccessful (the nonreform states) to those states in which the plaintiffs won (the reform states). They found that inequality fell in all 11 reform states compared to only 46% of the nonreform states. They then plotted the growth rate in revenues over that 20-year period at the 5th and 95th percentiles and found that revenues rose in all states at both ends of the spectrum, but, in the reform states, growth among the low revenue districts was typically much greater than that in the high revenue districts. Thus, litigation induced inequality reduction may have been generated by aggressive funding increases to the poorer districts. Third, the authors sought to confirm this finding by determining the mechanism by which inequality was reduced in all states. To do so, they first compared the reform states with the nonreform states in terms of the difference in actual Theil indices for state funding vs. the Theil indices for a counterfactual educational finance plan where states provide districts with flat grants equal to actual state revenues per student. From this analysis they found that reform states had targeted funding to the poorest districts at an amount greater than nonreform states. Thus, inequality was reduced primarily by an increase in state spending at the bottom of the distribution.

At bottom, although early studies were tentative about the impact of judicial intervention on inequality in school spending, later and somewhat more sophisticated studies seem to suggest that court decisions have reduced the revenue and spending disparity among property wealthy and poor districts, despite the difficulties courts face in implementing equity decrees. We remind the reader, however, that the conclusion of court efficacy in reducing spending and revenue disparities is a narrow one. This only tells us that one input measure of horizontal equity — money — has been improved. This tells us nothing about output and outcome measures, nor does it address issues of vertical equity or equality of opportunity as defined by Berne and Stiefel. Studies have considered, however, the effect of educational finance litigation on an equity target other than students — studies have considered the impact of litigation on taxpayer equity.

The findings here seem consistent. Successful educational finance litigation has resulted in a greater centralization of funding at the state level and a reduction in local property tax burden for school spending. The Clune (1979), Hickrod (1992), and Evans (1997) studies all support this conclusion. Hickrod specifically found that, compared to those states in which plaintiffs lost or no litigation is present, states in which plaintiffs have won their case experience a greater centralization of funding at the state level.

The Evans study took this finding a step further. There, the authors employed a fixed-effects econometric model to isolate the effects of litigation on education revenues. By fixing the "state effects," their model accounts for all factors that are constant across time, but vary across states (such as the degree of "liberalism" in the state). Similarly, the authors employ a fixed-year variable to hold constant all those factors that impact all states equally (such as recessions) but vary across time. The model does not allow for specific state-year interaction terms, however (the authors example is the possibility that Massachusetts — but not Texas — was more liberal in 1977 than in 1972). For each independent variable, two models were then estimated: one using court intervention as the covariate and another using years after court intervention as the covariate. The authors findings can be summarized by saying that court-ordered reform has the effect of raising state revenues and leaving local revenues unchanged.

Additionally, Evans employed the fixed-effects model, but included as an additional covariate a factor that represents legislative reforms undertaken without court intervention. Their findings are that the legislative reform variable does not alter the impact of litigation variable and that legislative reforms that were not in response to litigation had no perceptible impact on any of the the revenue measures. They further found by using descriptive statistics and comparing legislative to litigation-induced reform that the legislative reforms did not target the poorer districts as aggressively as the litigation-induced reform states. Thus, the authors concluded that "[r]eform without successful litigation is typically ineffective. States that have attempted reform on their own did not change the overall level of education revenues or the distribution of those revenues" (p. 28). This is part of the general theme of these studies. Successful litigation is related to a reduction in inequality and a centralization at the state level in educational funding.

Apart from educational finance reformers’ stated objective of reducing inequity in school funding, reformers may hope that overall state and local spending on education will increase as a result of their efforts. The mechanism by which such increase may be effected is the "leveling up" of property poor school districts to the spending levels of their more affluent peers. To the extent that reformers seek some "adequate" level of funding, logic suggests that the current level of funding is inadequate and must be increased. Yet judicial intervention in educational finance policy may not have the effect of increasing educational spending in a state. If the goal of litigation is equity, state and local funds may simply be redistributed to poorer districts to achieve that goal without changing overall spending (though folks here in Texas understand the political difficulties of such "Robin Hood" proposals). If the goal is adequacy for poor, low-performing districts, the same type of redistribution may occur (though we recognize that New Jersey’s solution has been to raise the bottom without restricting the top spenders). More insidious, successful educational finance litigation might have the perverse effect of decreasing middle and upper middle class support for public schools in general, causing the flight of such parents’ children to private schools, and resulting in the concomitant pressure to reduce public school spending. Stated more benignly, it is possible that a plaintiff victory in court will be viewed as a precursor to a tax hike thereby causing a taxpayer revolt against school spending. Indeed, some have debated whether California’s infamous, school-choking Proposition 13 was caused by the successful Serrano litigation (Silva and Sonstelie 1995; Fischel 1996). In theory, it is not at all clear what effect successful litigation has on overall school spending. Let’s turn to the empirical evidence on the question.

Among the early studies that examined this question, the Hickrod (1992) study found that, comparing the level of per pupil revenue in all 50 states in the two years of 1970 and 1990, there was an increased growth in the rate of combined state and local per-pupil revenues in those states in which the court found the funding system unconstitutional. Recall that Hickrod’s study compared those states in which litigation was successful to those in which litigation was unsuccessful or had not been pursued. Thus, Hickrod further concluded that even in those states in which plaintiffs lost, the rate of growth was greater than that in which plaintiffs won. This finding suggests that the mere filing of a lawsuit is related to education spending.

Hickrod’s study, however, was later criticized by Michael Heise on a number of grounds (some of which the Hickrod team itself recognized) (Heise 1995). For instance, Hickrod’s data and analysis could not eliminate internal threats to validity such as the effects of the state budget or student enrollment on student spending. Moreover, because Hickrod used only two data points, it is extremely difficult to link the independent variable — the court’s decision — to the school spending variable. Any number of intervening factors may have coincidentally affected school funding levels. Thus, time series data would be helpful. Heise sought to correct those flaws by using time series data and econometric modeling of state school spending.

Heise asked the question of whether "court decisions that declare education as a fundamental right and invalidate educational finance systems lead to increases in total state education spending" (p. 198). To address this narrow question, Heise developed a general model of state education spending that isolated the effect of the judicial decision and controlled the effects of the prior year’s education spending, the total spending of the state, and student enrollment changes. He also controlled for inflation.

Heise compiled data from Connecticut and Wyoming, each of which was subject to judicial intervention on equity grounds in 1977 and 1980 respectively. A visual inspection of total education spending in each state revealed no apparent change in spending associated with the judicial decision. Heise then used an ordinary least squares regression to estimate the parameters for the above-stated independent variables. He found that the parameters for prior year state education spending were positive and significant for both states; that prior year total spending was significant for neither state, that prior year enrollment was significant and negative in Connecticut; and that judicial intervention was not significant in Wyoming and was significant and negative in Connecticut. The positive relationship between the prior year’s education spending and the current year education spending is unsurprising. The other findings are more puzzling. Total state spending had no impact, i.e., additional state dollars in these states went to spending priorities other than education — educational spending did not grow as rapidly total state expenditures. The negative parameter for student enrollment in Connecticut is surprising, but, according to Heise, reflects a sharp enrollment decline in Connecticut without a concomitant drop in total spending. Finally, Heise considers the judicial intervention variable and concludes that "the picture that emerges does not support the general assumption that successful equity lawsuits result in increased state education spending" (p. 213).⁹ Heise’s study goes a long way toward addressing internal factors other than litigation that might be related to educational spending. In fact, such a time-series design with controls for threats to internal validity may be the strongest design for single-state case studies. The drawback to Heise’s design is that it does not compare states to each other and thus does not control for external threats to validity and cannot be easily generalized. The Evans study (1997) addressed those issues.

As we already mentioned, using time-series data from 47 states and a fixed-effects regression model, Evans found that court-ordered reform had the effect of raising state revenues while leaving local revenues unchanged. This finding seems at odds with Heise’s and Joondeph’s findings of decreased state funding. That said, there is sufficient ambiguity as to the effects of educational finance litigation on overall school spending that we view it as an open question.

Summary of the Effects of Litigation

Twenty-five years after Rodriguez and educational finance litigation in some 43 states, what have we learned about the impact of such litigation? Despite the many obstacles to judicial efficacy, the quantitative studies we have reviewed here suggest that educational finance equity litigation has resulted in greater, though far from perfect, per-pupil spending equality among districts and students; greater centralization of school funding at the state level and concomitant local property tax relief; and an ambiguous effect on the overall level of school funding in the state. That litigation has reportedly had such direct effects is noteworthy given the likely indirect effects it has had. Surely every state legislature is aware of the possibility of educational finance litigation and many have likely taken prophylactic measures, such as adopting more equitable finance schemes, to avert such litigation. Thus, even in those states without direct judicial intervention, the impact of law and litigation may be present.

We also emphasize the preliminary nature of these studies. In the past eight years there has been a great deal of judicial activity in educational finance and it is too early to determine what has been the effect of that activity. We particularly note that this latest round of educational finance litigation has been aimed at questions of adequacy as much as equity. Whether adequacy litigation affects the distribution and level of education spending is a question that should receive research attention. For now, the research suggests that the reformers’ and litigators’ time has not been wasted in court.

Conclusion

The Rodriguez decision was a massive setback for those who sought and continue to seek educational equity and economic justice through the courts. In an opinion joined by five Justices, the Supreme Court not only declared that "poverty" was not a suspect classification deserving of heightened protection under the Equal Protection Clause, but also determined that education, at least above some minimal level, is not a fundamental right under the Constitution. Thus, educational finance reform was effectively foreclosed in the federal courts. Undaunted, reformers turned to state courts and constitutions for relief.

Twenty-five years and dozens of state high court opinions later, we have learned a number of lessons from the laboratory-states. We have learned that money matters, at least on balance and in the right places. We have learned that, despite barriers to judicial efficacy, litigation has affected the distribution of educational resources among students. And we are now learning that adequacy as a touchstone for finance reform, though intuitively and politically appealing, poses its own difficulties. For reformers and policy-makers, these lessons indicate a need to consider carefully how strategies of reform are linked to results. No doubt this is important. But making this link is no easy task and at the end of the day values and politics will be the final arbiters of educational finance decisions. That being the case, it is worth recalling Justice Thurgood Marshall’s dissent in Rodriguez. In response to the claim that variations in educational spending do not necessarily result in discriminatory consequences for the children of disadvantaged school districts, Justice Marshall opined:

It is an inescapable fact that if one district has more funds available per pupil than another district, the former will have greater choice in educational planning than will the later. In this regard, I believe the question of discrimination in educational quality must be deemed to be an objective one that looks to what the State provides its children, not to what the children are able to do with what they receive.

* * *

[I]t is difficult to believe that if the children of Texas had a free choice, they would choose to be educated in districts with fewer resources, and hence with more antiquated plants, less-experienced teachers, and a less-diversified curriculum. In fact, if financing variations are so insignificant to educational quality, it is difficult to understand why a number of our country’s wealthiest school districts, which have no legal obligation to argue in support of the constitutionality of the Texas [school funding] legislation, have nevertheless zealously pursued its cause before this Court...

At the very least, in view of the substantial inter-district disparities in funding and in resulting educational inputs shown by [the Rodriguez plaintiffs] to exist under the Texas financing scheme, the burden of proving that these disparities do not in fact affect the quality of children’s education must fall upon the [State].