Financing the Schools
To the Editor:
In “‘Serrano’ vs. the People” [September 1972], Chester E. Finn, Jr. and Leslie Lenkowsky attack the judicial decisions (beginning in 1971 with Serrano v. Priest) that would upset the traditional state systems of school finance which today make the level of spending for every child’s public education dependent on the property wealth of his school district. The argument of Messrs. Finn and Lenkowsky against Serrano parallels complaints made by . . . Daniel P. Moynihan. (See Mr. Moynihan’s article, “Equalizing Education—In Whose Benefit?,” Public Interest, Fall 1972.)
Citing each other as authority, the contrapuntal chorus of this anti-Serrano squad is not always clear. But their message certainly includes this much: 1) It would be dangerous and undemocratic for the courts to violate the legislative integrity of education by requiring uniform expenditure for every child in the state and by mandating higher overall levels of expenditure for education; 2) it would be dubious policy to reduce educational spending for the poor in order to increase it for the rich; 3) raising the salaries of today’s teachers would not necessarily improve education; 4) spending more money on schools would be unlikely to raise achievement scores very much. We agree with all four points. These propositions reflect perfectly sound social instincts; it is merely the critics’ understanding of Serrano that lags. Perhaps something useful can be accomplished here with a clarification or two.
The essential facts are these. School districts are dependent upon the local property tax but vary enormously in their taxable resources per pupil; hence, districts with low taxable-property-per-pupil typically have relatively little to spend despite their higher tax effort. The state “equalization” programs praised by Mr. Finn et al. are a joke; in some respects they even operate to give preference to rich districts. This is the case in California; in 1968-69, spending in larger districts (with over 1,000 pupils) ranged from $550 in a poor elementary district (with a 4-per-cent tax rate) to about $1,250 in a rich elementary district (with a 2.5-per-cent rate). In Serrano the California Supreme Court rejected such a “system” as the basis for allocating funds to public education as unconstitutional; other courts have followed, and the issue is now before the Supreme Court in a Texas case argued this past October (San Antonio Independent School District v. Rodriguez) .
The critics fear the scope and impact of such decisions but show limited understanding of the basic concept. Writing in the Public Interest, Mr. Moynihan states the issue to be “. . . simply whether the courts shall order equal educational expenditure.” Anyone who believes this can scarcely have read the decisions. The courts were not even asked to equalize expenditure, and they explicitly rejected any such result. The rule of these cases is merely this: that the level of resources spent upon a child may not be affected by the amount of high-value real estate in his school district: “spending may not be a function of wealth.” This is all the courts have decided.
Nothing in Serrano compels either equality of expenditure or specific levels of expenditures. By rejecting district wealth as a basis for spending, the courts have in no way impinged upon options available to the state legislatures. A state may spend extra money on children who are gifted, disadvantaged, younger, older, play the flute, or hear poorly. The state may even preserve and enhance local control by making spending (in whatever degree) dependent upon the tax rate that local voters are willing to bear, provided each district can raise an equivalent amount per pupil with a given tax rate. With regard to the total amounts spent on education, Serrano leaves the legislatures free to spend less or more than they are currently spending. The question of how much should be spent on education will necessarily, in Mr. Moynihan’s words, be “deliberately chosen and legitimately enacted by the elected representatives of the people”—not by a court.
Because the courts have left to the legislatures the decision of how much should be spent, Messrs. Finn and Lenkowsky’s fear of a “large net increase in educational spending” must necessarily rest on a political prediction that legislatures rather than cut expenditures in rich districts will choose to “level-up,” even though that would mean spending more overall. Although this prediction may turn out to be correct, it is certainly open to question, given the present general annoyance with public education and the numerous recent refusals by local taxpayers to increase educational expenditures. In any case, let us assume the political process in the state legislatures should increase total expenditures. This would only suggest that parents in rich districts are unwilling (the arguments of Christopher Jencks and James Coleman notwithstanding) to have their children carry the risks from lower expenditures which are now borne by children in poorer districts.
Rather than usurping legislative prerogatives, Serrano will leave to the legislatures these issues of allocation and simply require that the decision be based on relevant educational policies, not on district wealth. The “Serranophobes” concede that the system stinks and that the legislatures should act. However, in the face of fifty years of evidence to the contrary they assert that legislatures will act—and wisely—but only if Serrano goes away; if Serrano is the law these same legislators will become damned fools and squander public treasure on the middle class. Perhaps the critic is correct, but is he then entitled to cloak himself in the mantle of democracy? Ironically, it is the Serrano critics, not the courts, who distrust the legislative process.
Apart from the usurpation argument, the critics’ primary thrust is that Serrano will hurt the poor. They imply that the status quo might (by the chance distribution of real property wealth) be better for the poor because the poor live in cities and that the cities are property rich. While they selectively identify certain cities that are, in fact, property rich, numerous contrary examples can be cited. In Washington, D.C., for example, 14 of the 17 districts with more than 10,000 students fall below the norm. Of California’s large cities, San Francisco—with 2 per cent of the state’s pupils—is property rich. Many of the other large cities (like Los Angeles) are about average, and a number are considerably below average, as is the case with San Diego, Fresno, and Modesto. In all events, even if on the basis of comprehensive data not yet available it were someday shown that a majority of the poor do live in rich districts, there would still be all those poor who live in poor districts and are enjoying both high taxes and low spending. They cannot afford to escape to private schools, along with their middle-class neighbors in such districts.
The critics also argue that Serrano will hurt the poor because its effect will be regressive and wasteful. Their primary fear seems to be that any Serrano-induced increase in spending will be eaten up by inflationary salary boosts for middle- and upper-middle income teachers—“the milkman will end up paying more so that his children’s teacher can earn more.” Given the outlook for the teachers’ market over the next decade, this fear seems fanciful.
As any first-year economics student can recite, a price (and teachers’ salaries are a price) is determined by supply and demand. There are market forces at work on both the supply and demand side of the equation that make it very unlikely that teachers’ salaries will rise in the next decade at anywhere near the rate we have experienced since 1950. On the one hand, the number of children who will enter elementary and secondary schools in the next decade will decrease. Therefore, unlike the 50′s and 60′s, school districts will not have ever-increasing needs for additional teachers because of higher enrollments. This decrease in the demand for teachers will be occurring at the same time that the supply of new teachers is at the greatest level in American history.
The net result of these market forces will be, in the words of the Fleischmann report, that “the supply of teachers will almost certainly far exceed demand from the mid-1970′s until the mid-1980′s.” . . . While it can be argued that teachers’ unions will be able to overcome these market forces and extract inflationary salary increases, the work of economists in studying other industries suggests that this is very unlikely. . . .
The critics implicitly acknowledge that the effect of an increase in spending might not be simply to raise teachers’ salaries but instead may be used to hire additional teachers; however, they imply that the only effect of this would be to give additional employment to middle-class teachers. If more teachers are hired, surely these additional inputs will give the children in poorer districts the possibility of new and broader curricular options and experimentation as well as smaller classes. Any calculus of the distributional effects of Serrano must take into account the benefits of such additional inputs.
But additional inputs, say the critics, are of little consequence in the light of what might be called the argument from futility. Giving poor districts more resources is, in any event, purposeless, since more money buys so little additional education. By this they mean that it buys only a modest increase in scores on standardized achievement tests. This conclusion, extracted from Coleman and Jencks, may possibly be right; one day, with reliable and adequate data, they might convince the numerous and often persuasive skeptics. A money-doesn’t-matter argument is currently being promoted in the Supreme Court litigation by certain school districts that are opposing Serrano; these districts are—guess who?—Grosse Pointe, Michigan; Montgomery County, Maryland; Beverly Hills, California; and others all near the top of the wealth and spending scale. Perhaps their implicit contention is that only the rich know how to spend money well.
We don’t know whether more money buys higher achievement scores or significantly raises children’s income as adults. Nor is either conclusion necessary to Serrano. We prefer to believe (with Grosse Pointe?) that the effects of educational spending are real, whether or not they are testable or lead to higher personal income. These effects include the acquisition of countless skills (the cello, drama, foreign language, and the ability to handle complex scientific equipment); they include also the opportunity for the school to broaden the curriculum in an effort to touch capacities for learning that in some children may be released only by access to a special experience or area of study. Perhaps most important of all—as Messrs. Finn and Lenkowsky themselves argue—society must be prepared for “heroic experimentation.” Serrano, of course, invites such experimentation and asks only that the capacity of experiment be fairly and rationally allocated and not be confined to the rich districts.
Christopher Jencks might have added that Serrano is justified simply as a step toward making the long years of compulsory education a decent experience for the child. We would agree with this and suggest further that the quality of the child’s experience may strongly influence his adhesion to and concern with the social order for which his school explicitly strives to prepare him.
Serrano is not, we concede, the gateway to Utopia for the poor or anybody else. It is merely a unique invitation to legislative creativity—a way of telling the responsible agency to turn off the discrimination machine and build us something that has a rational connection with the legitimate aims of government. For those who believe in the political process, such an invitation should come more as an opportunity than a threat.
John E. Coons
Robert H. Mnookin
School of Law
University of California
To the Editor:
The article by Chester E. Finn, Jr. and Leslie Lenkowsky against school-finance reform can be faulted on three grounds:
- It presents a narrow and selective interpretation of Serrano-type court cases.
- It misapplies recent research on educational achievement.
- It suggests an alternative solution which is logically inconsistent with the arguments presented earlier in the article.
Serrano-type court cases attack the pattern and distribution of educational expenditures, not their level. Hence Serrano is concerned with equity rather than with adequacy of educational achievement. To argue that the gross disparities pointed up by Serrano should not be redressed because equalization implies leveling up rather than leveling down is equivalent to telling poor children that it is just too politically difficult to give them their constitutional rights.
Moreover, the authors argue as though the trial of the pending suits is going to settle the matter. Actually, the reformers behind these cases have asked the courts to rule on the constitutionality of present financing arrangements rather than to design specific solutions. Following the court decisions, I foresee a period of legislative and executive action to devise the necessary remedies, not a period in which judges will be asked “to resolve economic, social, and educational issues of uncommon complexity.” . . .
The authors cite studies by Coleman and Jencks and others as evidence that the commitment of more resources to education will not necessarily result in any improvement. It is not the purpose of this letter to take a position in the current debate over the relationship between spending on schools and educational outcomes. It is my opinion that this debate is very far from being resolved. People who ignore or discount the findings of Coleman and Jencks may have closed their minds, but people who think that these studies prove that schools make no difference are also wrong.
To argue that the finding that “dollars make no difference” justifies the maintenance of a highly regressive and inequitable system is a non sequitur. Even if dollars do not make a difference, poor children should have the same right to the “frills” of modern, bright classrooms, smaller classes, and pleasant learning situations that wealthier children have. Taxpayers in low-wealth districts certainly should have the same right to pay low taxes as those in high-wealth districts.
Finally, the article suggests that additional money would be more wisely spent on income maintenance than on schools. In my view this alternative holds no great promise. Income-maintenance programs, as they now exist and as they have been proposed, offer very little chance of reducing racial or economic segregation. The patterns of housing and peer-group associations are not going to be changed by providing families with little better than subsistence-level incomes.
The authors do suggest one remedy with which I can agree—their call for “heroic experimentation.” But how can the low-wealth school district afford it under today’s financing arrangements?
H. Reed Saunders
Task Force on School Finance
Department of Health, Education, and Welfare
To the Editor:
Several of the important questions with respect to educational finance are at the moment far from answered. It is interesting, however, that writers such as Chester E. Finn, Jr. and Leslie Lenkowsky find it easy to accept questionable results which support their own predispositions on various subjects. They contend that fiscal neutrality would redistribute money in incorrect directions (i.e., toward teachers and not strongly enough toward central cities), and that added money for the schools would add little in terms of pupil achievement. Their conclusions are based mainly on an interpretation of the research results of a few social scientists.
As is the case with most authors who want to “prove” that educational expenditures are not productive, Messrs. Finn and Lenkowsky cite the Coleman study and the more recent Jencks study. I find it highly illuminating that they imply that only a handful of social scientists comprehend the statistical intricacies of these studies. Perhaps this is a tacit admission of their own inability to comprehend the sharp criticisms of Coleman’s methodology which apply almost equally strongly to many of Jencks’s results. (See Samuel Bowles and Henry M. Levin, “The Determinants of Scholastic Achievement—An Appraisal of Some Recent Evidence,” Journal of Human Resources 3, Winter 1968; and Glen G. Cain and Harold W. Watts, “Problems in Making Policy Inferences from the Coleman Report,” American Sociological Review 35, April 1970.) Even for the statistically unsophisticated it should not be difficult to comprehend that high levels of spending on education generally occur in districts for which typical socioeconomic status indicators will also be high. To attempt to determine whether it is the differences in educational expenditures or the differences in status which explain learning differentials is simply impossible given the data we have to work with. A simple truth is that if statistical techniques are used which feed in variables in order to explain some phenomenon, the first variable fed in will appear to explain most of the variance actually due not only to itself but to other variables which are systematically related to it. Since both Jencks and Coleman enter socioeconomic variables first, most variance in achievement is “explained” by them, and little is left to be explained by educational expenditures. As is demonstrated very ably by Hanushek and Kain (in “On the Value of ‘Equality of Educational Opportunity’ as a Guide to Public Policy,” in On Equality of Educational Opportunity, edited by Frederick Mosteller and Daniel P. Moynihan), the set of educational-expenditure variables used by Coleman can explain more variance in achievement when entered into the analysis first than can either the set of peer-group or family variables when they are entered into the estimating equation first. Any statements made about the small amount of variance in performance that can be attributed to school-input variables can therefore be made equally strongly about the other types of variables. Their relative strength is simply a function of the authors’ essentially arbitrary choice of an ordering of variables. Even though Jencks supposedly understood this problem in Coleman’s work, it is noticeable that he uses the same methodology in his book Inequality.
The moral of the above story is that we do a very poor job of statistically confirming what helps children achieve. The reasons for this are various, and include seriously inadequate data sources and exceptionally naive attempts to build mathematical models of the educational process. The attempts so far have had little information on expenditures over time on various students, almost no useful information regarding quality of teaching, and little sophistication in the attempt to determine which inputs at which point in a child’s life are important for academic achievement. But to state that our work on explaining educational achievement is poor is very different from stating that . . . none of the variables used in this analysis leads to achievement. . . . To interpret not finding a statistically significant relationship as meaning that no relationship exists is simply stretching the usefulness of statistical methodology beyond its limited capacity. . . .
It seems then that the strongest case that can be made on the basis of such studies as Coleman and Jencks is that statistical methods combined with simple models of how educational output is produced have not been able to indicate that a significant relationship exists between added school inputs and increased pupil test scores. This is certainly not nearly as strong a statement as the often implied finding that “school inputs have been found to have little or no effect on school results.” . . .
The other point stressed in the article is that fiscal neutrality would lead to redistribution away from “needy” pupils. If the ability-to-pay for education is defined as property wealth per pupil, then this contention might be correct; but the problem is with the measure of ability. . . . A measure of ability must be derived which takes into account all sources of tax revenues in a locality as well as competing claims for this revenue for other public purposes. Thus, if central cities have little ability to finance education because other “necessary” public services demand so many tax dollars, an accurate “ability measure” would reflect this condition. If such a measure of ability is then equalized among communities, it is doubtful that the results implied by Finn and Lenkowsky would follow. . . .
Perhaps the use of property values as the measure of ability is to be avoided; but to avoid the provision of equal opportunity may have strongly negative consequences for the poor. It should be noted that the preceding statement is equally true even if the only thing purchased by school expenditures is a pleasant experience for pupils. It is difficult to find precedents for suggesting that states should set up systems through which, in a publicly-provided service, desirable goods are provided to citizens on the basis of wealth. . . .
Messrs. Finn and Lenkowsky imply that educational expenditures may increase achievement significantly in districts in which less than some necessary minimum is being spent. . . . The fact that added educational expenditures have little effect on test scores on the average over the nation does not mean that added expenditures would not have a strong positive effect on test scores in those districts which spend little because of the lack of property wealth. It is to just such districts that the Serrano decision, correctly implemented, could provide much financial aid.
John S. Akin
Institute for Research on Poverty
University of Wisconsin
Chester E. Finn, Jr. and Leslie Lenkowsky write:
Our correspondents have thoughtfully provided COMMENTARY readers with excellent examples of the logic—rather, the illogic—animating those reformers afflicted with what might be termed “Serranophilia.” Thus, the letters by H. Reed Saunders and by John E. Coons and Robert H. Mnookin contend that Serrano and its kin are nothing more than tax cases when, in fact, their legal raison d’être rests upon an affirmation of the fundamental importance of schooling, the belief that the fiscal affairs of school systems have much to do with the quality of service available, and the conviction that such quality affects the educational experience of children. Then, too, we learn from these letters that the courts will intervene to prod legislatures and executives to take appropriate action; conveniently forgotten is the certainty that such remedies as are chosen will last only until the next round of litigation. The Serranophiles also say that the legislatures will save us from the excesses of Serrano through all manner of corrective acts, including limiting the power of the teachers’ unions to bargain for salary increases and providing specially earmarked grants. But if that is the case, critics like ourselves may be forgiven our bewilderment as to why Serrano was necessary in the first instance.
A certain blindness is much in evidence among our correspondents. Despite a consensus of results from at least three major studies, one comprehensive review of research, and countless smaller reports, the new reformers refuse to concede the validity of the social-science findings on education that began with the Coleman Report and continue through the work of Christopher Jencks (although Jencks’s study in fact draws a policy implication congenial to their own). We do not for a moment suggest that reading scores or adult earnings are the only school “outputs” worthy of note, but we do suggest that the burden of coming up with more persuasive evidence now rests on those who would reject the findings of Coleman, Jencks, Moynihan, Mosteller, the RAND Corporation, et al.
Messrs. Coons and Mnookin recognize that there are differences among cities. So did we; that, in part, led us to conclude that the courts are an unsuitable instrument for the reform of school financing. Yet one would be mistaken to look for much rejoicing among the new reformers if the Supreme Court should decide Rodriguez in such a fashion as to limit its ruling to the particular case in dispute. It is national principles that our correspondents seek, and it is national principles that we consder least helpful in reforming fifty disparate state systems of school finance, each of which contains hundreds of special cases and unique conditions.
Our correspondents accuse us of anti-democratic bias. But we did not contend that only the rich know how to spend money well, or that the various state legislatures never squander the taxpayers’ money. Indeed, we thought the absurdity of those propositions was clear enough so as not to require comment. Our argument was, simply, that educational reform has very little to do with issues of school financing; moreover, that changes in school financing have perhaps even less to do with tax reform or progressive income redistribution.
Let us, however, even assume for the sake of argument that Serrano is simply a case for equitable taxation: is it then to be applauded? Our purpose was to show that nothing in the field of school finance is any longer as simple as one might wish, that ending one form of inequality usually induces another, and that the courts and those who make use of them would do well to consider the second- and third-order consequences of their actions. As we read the complex data in this area, certain propositions seem basic: First, low property wealth is very imperfectly related to the income levels of a community. Second, a community with high personal income—typically a suburb—generally also has a high tax rate for educational purposes, but a relatively low overall tax burden. It spends much of what it raises on schooling, less on other services. Third, a community with high property wealth—notably a big city—is likely to have an extremely high overall tax burden, but to spend a much smaller portion of it on schooling. The amount spent per pupil may be quite high. And finally, voter preference for educational expenditures varies directly with personal income.
From these, we conclude that Serrano’s rule of “fiscal neutrality” will provide tax relief primarily to middle-class property owners living in suburbs (and to the small population of rural folk), while doing little good and much possible harm in the cities. Suggested formulas that comply with the “Serrano rule”—such as the “power equalization” advocated by Mr. Coons—do not correct this problem and may make matters worse. The demonstration is complicated, but has been set forth recently in a note in the Yale Law Journal (Vol. 81: at 1303, 1972) and in an article by Arthur Solomon and George Peterson in the Public Interest (Winter 1973). The essential points are that property wealth and community income are not the same and that districts with large concentrations of low-income residents tend to spend a smaller proportion of their tax revenues on education.
John S. Akin recognizes these points in warning against the use of “ability-to-pay for education” criteria. He would prefer a more general measure of fiscal ability, which takes into account “competing claims” for public revenues. This is a valid argument if one is interested in restructuring the fiscal arrangement of American federalism, although there is no reason why education should be the only service upon which reform focuses. But it is not the argument of Serrano, which is rather exclusively concerned with relieving the alleged inequities of community wealth, i.e., property valuation. This approach is primarily of tactical merit; presumably not even the most activist courts are yet ready to adjudicate the rectitude of different state-local patterns of providing public services. Yet the decision to let smaller cities pay for fire, police, and sanitation on a county-wide basis, while large cities assume their own costs, is clearly at least as important in shaping “competing claims” as are indices of potential tax revenues.
There is little in Mr. Akin’s argument about the Coleman Report, either, that would lead us to revise our assertion that few social scientists understand its statistical intricacies. While the technical point Mr. Akin makes about the order of entry of statistical variables is absolutely correct, he fails to emphasize that the Coleman analysis has been replicated, taking this issue into account. The most thorough of the reanalyses aimed at settling this dispute was performed by Marshall S. Smith (in Mosteller and Moynihan, On Equality of Educational Opportunity, pp. 230-342). After assessing the merits of the criticisms of the Coleman Report, and agreeing with Mr. Akin and others that there are good reasons for suspecting statistical biases in the results, Smith nonethless finds himself in basic agreement with the original findings. We might reiterate that Smith’s (and Coleman’s original) position seems to us to be increasingly adopted by scholars who have studied the issue.
The point of this lengthy reply is not to settle the question of whether the Coleman findings are correct. Rather, we are simply asserting that in recent years the conventional wisdom about schooling has been seriously questioned. Virtually every commentator, including Mr. Akin, agrees that the issues raised by On Equality of Educational. Opportunity ought to be settled, either by disproof or by developing new models of schooling that are effective. The Serranophiles offer a third alternative—ignorance—pretending that the virtue of the conventional wisdom has never lost its loveliness. We feel that the course of action suggested by Serrano is undesirable and will only attach to the banner of educational reform issues that are no longer germane and have the potential only for creating endless discord and confusion.