John Chubb is a senior fellow at the Brookings Institution. Terry Moe is a professor at Stanford University. This article was condensed and adapted by the editors of NY from the authors’ new book: Politics, Markets, and America’s Schools (Brookings, 1990). Readers who wish to fully assess the methodology and results discussed here should refer to the book and its appendices.

During the Eighties, spurred by the grim statistics on the academic achievement of American students, America came to a remarkable consensus on the need for education reform. An extraordinary amount has been done since then, mostly at the state level, in an attempt to turn things around. Between 1981 and 1986 expenditures per pupil on elementary and secondary education went up 40 percent. We now spend four times as much per child (in real terms) as we did in 1950. In the first half of the 1980s teachers enjoyed larger salary increases than any other occupational group in the country. Reformers have taken steps to reduce class sizes, crack down on teacher incompetence, and boost promotion and graduation requirements. Some states and school districts have launched bold experiments with career ladders and merit pay for teachers, and with decentralized “school-based management,” which New York City School Chancellor Joseph A. Fernandez is pursuing.

The one thing that has not changed much is the wider institutional framework in which schools are run: America’s public schools are still largely controlled by a network of outside authorities, from federal and state agencies to teachers unions, to the central offices of large, consolidated school districts. Our study of American schools, the largest and most thorough ever undertaken, has convinced us that these wider controlling institutions crucially affect school success or failure. Because they ignore these controlling institutions, the education reforms adopted thus far—for all their effort and expense—are destined to fail.

What our study did

Two questions have motivated more than 20 years of “effective schools” research: Do schools make a difference in the academic achievement of their students? And if so, how do effective schools differ from ineffective schools? Scholars here and in Great Britain have come up with some consistent answers that may surprise the uninitiated. First, schools matter enormously. Starting with students of similar aptitude, family background, and peer influence, some schools consistently turn out better-educated students than others. These are called, by educational researchers, effective schools. Second, the most important differences between effective and ineffective schools are organizational and intangible. Effective schools, for instance, have strong educational leadership, particularly from principals. Principals in good schools are dedicated to the classroom and think of themselves as leaders, not administrators. These intangible factors, and there are several others, are far more important than factors such as money, class size, teacher credentials, graduation requirements, and so forth.

Our study built on this earlier effective schools research and took it several crucial steps further. We examined data from 500 randomly selected high schools nationwide, both public and private, including information from 500 principals, 12,000 teachers, and 10,000 students in these schools. The data on the students came from the national High School and Beyond survey of student academic achievement, which tested students in math, science, reading, writing, and vocabulary. The students were examined twice: as sophomores, and again at graduation, allowing us to monitor their progress. The data on principals and teachers came from the Administrator and Teacher Survey, which we participated in, and included detailed accounts of school policy, educational methods, and personnel policy and attitudes. We compared these results against a mass of data about the schools themselves: statistics on such factors as funding, school population, curricula, etc. With the data in hand we began the search for the sources of student, and school, success or failure.

The resulting findings were quite strong. To begin with we confirmed the view of many researchers before us that schools do matter. Over the course of a four-year high school career, being in an effective rather than an ineffective school adds one full year, on average, to a student’s academic achievement, even after controlling for initial aptitude, parental socioeconomic status, and so forth. Schools are the second most important factor—after student aptitude—in student achievement, marginally exceeding the influence of parents and substantially surpassing the influence of peer groups.

What makes an effective school? We found, first of all, that there was no correlation between student achievement and any of the variables on which school reformers have been spending so much time, effort, and money. Neither expenditures, teacher salaries, class size, graduation and promotion requirements, teacher credentials, nor any similar factor made any difference. That is one reason school reform as it is now proceeding will fail.

We also found, as had previous researchers, that intangible and organizational factors are the most important. The influence of principals has been mentioned, we confirmed it. We found, as had others before us, that effective schools have clear academic goals; they make academic achievement their highest priority and let everyone in the school know it. Effective schools treat teachers like professionals; the teachers share the school’s goals and feel effective in pursuit of those goals. They participate extensively in school decisions; within their classrooms they are free to tailor their practices to the needs of their students. In short, effective schools are organized effectively. They are organized like teams, held together by consensus, cooperation, and shared goals, not primarily by rules and regulations. The unsuccessful school is bound primarily by rules and regulations; it is a classic hierarchical bureaucracy, rather than a team.

As we say, researchers have been pointing to these organizational and intangible factors for years as the key to school success. But none of the effective schools research had gone on to the next question: Why do some schools organize themselves in effective ways, while others fall short? All schools obviously want to be successful organizations, to make children learn. Why do some succeed and some fail? We did ask that question and our research uncovered a powerful explanation: Effective schools enjoy a great deal of autonomy. Principals and teachers are relatively free of outside control and free to pursue educational goals. Dysfunctional schools, on the other hand, are to a much greater extent ruled by outside authorities, and have much less autonomy. School superintendents, and central office administrators exert direct bureaucratic control; teachers’ unions impose bureaucratic control indirectly through union contracts. Intensive control of school policy by any of these forces produced equally bad results.

In some ways this is a common sense result: Schools that act like bureaucracies are controlled by some larger bureaucracy. But to study systematically the relationship between a school’s effectiveness and its autonomy, we had to solve two difficulties. First, we needed a way to measure autonomy. Second, we needed to disentangle cause from effect. If outside interference and ineffectiveness went together, was the interference causing the bad performance or was the bad performance provoking the interference?

To pin down the relationship between performance and autonomy, we set out first to devise a way to measure external control. In our survey of principals, we asked them to assess their own degree of autonomy from superintendents, central administrators, and unions in several areas, including setting curriculum, choosing instructional methods, hiring new teachers, dismissing or transferring teachers, and setting disciplinary policy. We then rolled together these five measures of outside interference to form one general overall index of “administrative constraint.”

There is reason to believe that control over personnel is especially important to a school. So we asked principals a number of separate questions to gauge in depth their control over hiring and firing. We used the results to create another index called “personnel constraint.”

Thus we created two measures or indexes of bureaucratic control: a broad, comprehensive index that we called “administrative constraint,” and a deeper, more specific index that we called “personnel constraint.” Armed with a way to measure outside interference, we compared the level of such interference in effective and ineffective schools.

As shown in Figure 1, the differences are quite striking. Overall, 68 percent of ineffective schools have above average levels of administrative constraint, compared with 39 percent of effective schools. On each of the five separate measures of administrative constraint (curriculum, instruction, hiring, firing, and discipline), effective schools come under less control from superintendents and central office administrators than dysfunctional schools. In four out of the five areas, the differences are large.

Losing control over hiring and firing of teachers appears to be especially debilitating to schools. Ninety-two percent of ineffectively organized schools had above-average administrative constraints on firing, and the vast majority were also subject to above average constraints on hiring. Among effective schools, only about 40 percent reported above average personnel constraints. This makes sense. If an effective school is one that functions as a team, with shared goals, then the freedom to build a cohesive team should be a great advantage.

A look at our second bureaucracy index, the deeper index of personnel constraint, confirms this finding. (See Figure 2.) Ineffective schools are almost six times more likely to face above average personnel constraints than effective schools. Forty-five percent of ineffective schools face above average interference in personnel decisions, compared with about 8 percent of effective schools.

In preparing this deeper index of personnel constraint, we asked principals about “barriers” they faced in “obtaining teachers with excellent qualifications” and in “firing or refusing to renew the contracts of poor teachers.” Only a small fraction of principals of effective schools say union constraints keep them from hiring top-notch teachers or firing incompetent ones. A majority of the principals of ineffective schools, on the other hand, say unions are a very large obstacle in this regard. Only a quarter of the principals of effective schools say the central office makes it difficult to get excellent teachers; 52 percent of the principals of ineffective schools say so.

We also asked principals to agree or disagree with the following proposition: “When two teachers with the same credentials apply for permanent transfer to this school, we decide which one to accept on the basis of their competence, not on the basis of their seniority.” A majority in both categories agreed with this proposition. But over 90 percent of the principals of effective schools agreed that they hired the most competent applicant, while only 60 percent of the principals of ineffective schools did so.

Thus in five key areas of education policy—and in personnel decisions especially —external, bureaucratic control is much stronger in ineffective schools than in effective ones. Effective schools have much more control over staff, curriculum, and educational policies. Ineffective schools are much more likely to have these decisions made by external authorities. In the first rough cut of the data, bureaucracy appears to be a major negative influence on schools.

Controlling the variables

The obvious response to this discovery is that not every correlation implies a cause and effect. Perhaps schools that are failing for other reasons get “taken over” by outside forces. Perhaps schools blessed with good students tend to be let alone, but would do just as well even if they were not. To untangle cause and effect, we constructed a new model controlling the analysis for other factors that our data found correlated with schools’ effectiveness, including student ability and behavior, parent socioeconomic status, and school funding. We found that even after controlling for these other factors, bureaucracy is the single most powerful factor in school effectiveness.

When parents, students, and the level of bureaucracy are taken into account, school funding has no significant influence on school effectiveness. Both students and parents on the other hand are a major influence on school effectiveness. All other factors being equal, effective schools are much more likely to emerge when schools serve well-behaved, well-prepared students who come from families with high socioeconomic status.

Student behavior is a particularly powerful factor. Taking two schools, average in all other respects, a very orderly school would likely be ranked around percentile 80 in effectiveness; a very disorderly school would score down around percentile 28.

But the strongest determinant of school effectiveness, more important than family background or any other factor, was autonomy from bureaucratic control. Both administrative constraint and personnel constraint are important independently; together they add up to the biggest single influence on school effectiveness. Taking two schools average in all other respects except that one comes under high bureaucratic influence, and the other low, the highly bureaucratic school will rank only in the 24th percentile of effectiveness, meaning that 76 percent of all schools will be more effective. The highly autonomous school, however, would be expected to rank in the 81st percentile in effectiveness, making it more effective than all but 19 percent of schools. Thus the difference between high and low levels of bureaucracy can amount to almost one full year in student achievement over a four-year high school career. Bureaucratic influence can make or break school performance all by itself.

In other words, of schools with similar student and family “endowments,” some are much more bureaucratically controlled than others, and the more bureaucratic ones are on average much less effective. Bureaucracy is a prime cause, not the result of, school failure. A school that is free to chart its own course is much more likely to develop an effective organization, and thereby boost academic achievement. Schools tend to prosper most when outsiders trust them and leave them alone. Which brings us to the last and most crucial question answered by our study: What causes schools to come under bureaucratic control? Is it something beyond all control, or are there ways to make a given school much less bureaucratic.

We suspected that the degree of bureaucratic control over schools might correlate with several variables including family background, student body characteristics, (initial achievement, additional achievement, and behavior) the amount of contact between parent and school, school “sector,” and school location. By “sector” we mean whether a school is public or private. By “location” we mean whether it is located in an urban, suburban, or rural area.

We were interested in family background because some observers have argued that more affluent parents might be more capable of fending off outside interference and poorer parents might be less capable of overseeing their schools, thus encouraging more oversight by the central authorities. Student characteristics seemed a possible factor because achievement and behavior stand for “good” or “bad” schools and outside authorities might be more likely to take control of “bad” schools. As for school sector and school location, they are measures of the type of institutional control to which schools are subject. Public schools are, in almost every case, part of larger administrative systems. In urban areas, public schools are usually embedded in extensive, politically controlled hierarchies sprawling over an entire city or county. In rural or suburban areas, on the other hand, a public school district may contain only a few schools, and nearly everyone in town may share the same rough vision of what the schools should be doing, which allows for minimal layers of administrative control. Most private schools are relatively free from outside control: Independent private schools are mostly self-governing, and parochial and church schools are mostly run by the parish or congregation. On the other hand, some private schools are also part of large systems—such as the parochial schools in large urban dioceses. That is why we use both sector and location to measure institutional control.

We found that bureaucratic schools tend to be public rather than private, and urban rather than non-urban. They also tend to serve students with more behavioral and academic problems, from families of lower socioeconomic status. The next step was to distinguish cause and effect from mere correlation, by measuring the significance of each individual factor, while controlling for the others.

Having done so, we found that neither parents nor students had a powerful effect on bureaucracy. Neither initial student achievement nor student behavior had any significant effect on administrative constraint. Additional student achievement had only a small effect. Personnel constraint varied in the opposite way: in response to initial achievement and student behavior, but not in response to gains in student achievement. Schools with less able, or misbehaving student bodies had less control over hiring and firing than others. But schools in which students learned a great deal were no more likely to be allowed autonomy than schools in which they learn little. Overall, the influence of students on the level of bureaucracy was inconsistent and not large.

Parents also had surprisingly little effect on bureaucracy. Parent socioeconomic status had no significant effect and parent-school contacts only a slight effect and that only on administrative constraint.

For a better idea of why some schools are bureaucratic while others retain their autonomy, we must look to the final two variables: school sector and school location. Both of these, it will be recalled, are indirect measures of the influence of the wider institutions that control schools. Most American public schools operate under political hierarchies, controlling many schools serving many students from different backgrounds. They must heed rules and regulations issued by superintendents, community school boards, central office education bureaucrats, and state legislatures.

Private schools respond to market incentives rather than to political rules. They can set distinctive educational goals, make them known to families, and admit (or retain) only parents and students committed to those goals. They can gather an educational community that agrees on basic priorities. On the other hand some public schools in rural or suburban areas can do the same thing.

As these considerations suggest, we found that both a school’s sector and its location have a major effect on the level of bureaucratic control it will encounter. Moving a school from an urban to a nonurban location reduces administrative constraint by nearly three quartiles; a shift from public to private sector reduces it by more than two quartiles. No parent or student variable, by contrast, lowers administrative constraint even one quartile.

Similar results emerge for personnel constraint. Vast improvements in student achievement or student behavior can reduce personnel constraints by a quartile to a quartile and a half. Changes in school sector or school location, on the other hand, reduce personnel constraint by about twice as much. Our findings were strong and clear: School sector and school location, our proxies for the mode of institutional control, were overwhelmingly important in determining whether a school would be bureaucratic or autonomous, while other factors mattered much less if at all.

It will be recalled that this analysis controls for the kinds of students and parents in each school. In other words our results are not merely a function of the tendency of urban schools to have more difficult students and less interested parents. Urban schools may have more troublesome student bodies, but—on our evidence—that is not the major reason they are more bureaucratic. The crucial difficulty of urban schools is that they are embedded in large, heterogeneous school systems.

The same unambiguous results apply to the public-private split. When all else is equal—when schools serve the same kinds of students from the same kinds of families, situated in the same general area—private schools are far less bureaucratically controlled than public schools.

Bureaucracy and school failure are not historical accidents, but inevitable consequences of the way we govern our public schools. The Federal Government now administers some 70 elementary and secondary education programs; state governments now provide more money to schools than local governments. These monies come with rules, regulations, and other strings attached, leading inevitably to more bureaucracy and centralized control. Some public schools are regulated down to the finest detail. Outside authorities decide which students attend which schools, allocate school budgets, hire and fire teachers, design the curriculum, choose the textbooks, prescribe teaching methods, impose disciplinary policies, and furnish the classrooms. Moreover as school districts have grown much bigger on average (their total number has shrunk from over 100,000 in 1945 to less than 16,000 today), rules and regulations have come to play a larger part in their operations. We believe schools controlled by large heterogeneous political authorities tend to be much more bureaucratic than schools that are not. When the communities they serve are sprawlingly diverse, school goals are likely to become politicized. The schools become an arena for a struggle over whose values, priorities, and programs will be enshrined in all the schools in the district, a struggle that undermines the kind of educational leadership and teamwork found in effective schools.

We devised one final test of our central contention: that American schools perform poorly because they are imbedded in political hierarchies. We have established the influence school sector has on bureaucracy. But we can also relate school sector directly to our measure of a school’s “effectiveness.” If we are correct that control by large political hierarchies undermines schools, then when all else is held constant, private schools should be more effectively organized than public schools.

The idea that private schools, as such, outperform public schools, has been controversial. The superior performance of private schools is widely recognized, but many skeptics view this as a statistical artifact. Private schools perform better, they contend, because they attract abler, and expel more troublesome, students. If they served the same population of students as the public schools, their apparent superiority, the skeptics argue, would disappear.

To settle this question we tested the relationship between school sector and school effectiveness, holding other factors constant. The results were striking. School sector clearly emerged as the most important influence on school effectiveness. When student achievement and behavior, family background, parent-school contact, school size, and even school location are all taken into account, private schools still tend to be more effective organizations than public schools. The standardized influence of school sector is at least three times the magnitude of the standardized influence of any other variable, including school location. All by itself a change from public to private status vaults a school from the middle of the second quartile to the middle of the top quartile of effectiveness.

In summary our study, based on the largest data base ever assembled for such research, showed three things:

By itself, effective school organization (leadership, ambitious academic goals, professionalism among teachers, teamwork, etc.) is capable of adding more than one full year in student achievement over a four-year high school career.

By itself, bureaucracy (excessive control on administration and personnel by outside forces) can destroy effective school organization, and autonomy (the absence of bureaucracy) can restore it.

By itself, changing the institutions of school control to liberate schools from large heterogeneous political hierarchies can create the autonomy necessary for effective schools, thus adding a full year to student achievement.

The autonomous, flourishing public school is not quite extinct today. Under the most favorable conditions—in non-urban areas with able students and interested parents—some public schools work well. Under any other circumstances, market control does much better. The qualities that public schools need to develop, private schools naturally possess. Qualities such as leadership and teamwork cannot be ordered into existence by new instructions from superintendents or new directives from a chancellor’s office. They emerge from the institutional structure of private education. Public schools cannot behave as independent entities because they are not independent. Various political bodies create the incentives they respond to, the standards they must live up to, and the limits on what they are allowed to do. Private schools, on the other hand, are controlled not by reams of regulations but by the need to please parents and students. They focus on the goal most families are concerned with—academic achievement—and they are free to pursue that goal without meddlesome interference.

If public schools are ever to improve, the institutions that control them must first be changed. To really fix our schools we must look beyond the classroom into the school system—the wider set of institutions that determine how schools are run.

It is clear, then, why the reforms of the last decade were doomed to fail.

The First wave of reform, beginning in the late 1970s, tried to improve schools the old fashioned way: by spending a lot of money and by imposing tough new controls, ranging from higher graduation and certification requirements, to new tests of teacher competence, standardized student tests for promotion, and so forth. The formula was simple: better courses, better teachers, higher standards, more money.

The reforms sounded impressive. The results have been less so. State legislatures can do only so much. They can force students to spend so many minutes sitting in classrooms where courses with certain titles are taught, by teachers who have particular paper credentials. Teachers and students can be made to go through the motions, following state-imposed rules to the letter. But laws cannot make children learn or teachers teach.

So, the logical next step is to impose standardized tests to make sure children are learning what the legislature has told them they must learn. These tests then become part of the bureaucratic burden schools must carry. Tests have several problems: They are limited one-shot measures of performance. Moreover, if schools are rewarded or penalized according to these imperfect measures, teachers will inevitably begin “teaching to the test,” which may undermine the quality of instruction. But most importantly these imperfect tests will be used as an excuse to impose new bureaucratic rules on failing schools, precisely the way to make the problem worse. As for money, the relationship between it and effective schools has been studied to death. The unanimous conclusion is that there is no connection between school funding and school performance, except perhaps at the stark extremes of deprivation or gross abundance. In our own study, we found effective schools did have some 20 percent more resources than dysfunctional schools. However, once controls were introduced for factors such as family status and student aptitude, the correlation between money and success disappeared completely. Schools with more affluent students do tend to spend more money and they get better educational performance. But it is the higher socioeconomic status of the families, not the additional money in the education budget, that explains the better performance.

All the first-wave reforms sound nice, and some may help in small ways, but at best they divert people and resources from real education reform and at worst they unleash new bureaucratic pathologies that on balance, make schools worse. After all the much publicized reform, the condition of American education remains grim.

By the late Eighties, a new consensus emerged on the need to liberate public schools from their bureaucratic straitjacket. The second wave of school reformers clearly understood that bureaucracy is bad and autonomy is good; but they did not look very deeply into how or why American schools came to be bureaucratized. With no clear analytical foundation to guide them, reformers reached for a grab bag of ideas that seemed to promise a reduction in bureaucracy. Of these, the most influential is school-based management. School-based management has been tried in a number of school districts from Chicago to Dade County, Florida, and at least one state (Hawaii). New York City is about to embark on a major experiment with it.

Under school-based management, central authorities delegate policy and budget decisions to principals, who are in turn often required to consult school councils composed of teachers, parents, and sometimes students. This is a nod in the right direction. But it leaves the traditional institutions-of political control intact. Schools remain embedded in politically controlled hierarchies. Education officials, elected or politically appointed, may delegate more but they remain accountable to decisions, the public for school performance; their necks are on the line, and they know it.

Authorities are jealous of their powers and as soon as schools make mistakes the authorities reclaim the authority they have delegated. And so a host of new rules, regulations, and requirements arises. Parts of the old bureaucracy are jettisoned, only to be replaced by new ones. The central office giveth, and the central office taketh away. As long as bureaucratic authority exists, it will eventually be used.

All the major players running our schools—including the unions—complain that the schools are too bureaucratic. They mean what they say. But what they do—quite rationally, in response to the institutional incentives—is bureaucratize.

Our reform efforts are doomed to failure unless and until we transform the institutions that control the schools. The only way to give schools autonomy is to make them accountable from the bottom up, to parents, rather than from the top down, to bureaucrats. The only alternative to bureaucracy is choice.

Public school choice has become quite controversial in education circles, and only a handful of school districts around the country have experimented with it. Yet surveys reveal that the vast majority of parents want to choose the schools their children attend; once choice plans are implemented, they receive even more popular support.

Choice takes diverse forms, not all of them good for education. The most common choice plans are built around “magnet” schools, often as part of desegregation schemes. Magnet schools generally offer special or advanced programs and are open to children from throughout the district, space allowing and provided they meet the qualifications. For the few children who are offered a choice, magnet schools have worked well. Racial balance is maintained. Dropout and absenteeism rates go down; achievement scores go up. But magnet school systems are only half-magnetized; most children are still trapped by geographic zoning in deteriorating public schools. Indeed the non-magnet schools may decline if they become dumping grounds for the district’s mediocre teachers and troublesome students.

Thus in most “choice” plans a majority of students remain trapped in schools of assignment. Even many of the more ambitious experiments with choice suffer from a deeper drawback: They are entirely focused on demand to the exclusion of supply. Parents may choose only among a fixed set of existing schools, all of which have their financial support guaranteed; authorities take action to ensure that no schools are “under enrolled,” a bureaucratic euphemism for what happens when schools are so bad no one wants to attend them. Indeed, schools that do badly are often implicitly rewarded, being singled out for bigger budgets and more staff. Rarely do reformers take steps to free up the “supply” side of the school system. But for choice to be effective, new and different types of schools must be encouraged, so that people have an attractive (and dynamically responsive) set of alternatives from which to choose.

The most ambitious experiments in public school choice have been carried out in Cambridge, Massachusetts and in East Harlem’s School District No. 4 in New York City. Cambridge’s eight-year experiment with a choice-based elementary school system has been a huge success. In 1980, Cambridge made all its elementary schools magnet schools, with parents and students applying to the school of their choice. Student achievement scores are up, and the gap between achievement scores in the worst and best schools is dwindling. Perhaps the best evidence of success is that Cambridge public schools are winning back students they lost to private schools. By 1987, 89 percent of the district’s kindergarten students chose to enroll in public schools, compared to 78 percent in 1979.

The most radical—and most promising—exercise in public school choice is Manhattan’s District No. 4, serving some 14,000 East Harlem students from prekindergarten through the ninth grade. In 1973, District No. 4 ranked last of New York City’s 32 school districts in reading and mathematics. The demographics of the district made this seem inevitable: More than half of all families are headed by single females; almost 80 percent of all students are poor enough to qualify for free-lunch programs; almost all the students are minorities: 60 percent are Hispanic, 35 percent are black.

But District No. 4 was also lucky. It had dynamic leaders who were willing to take risks. Beginning in 1974, they oversaw the creation of an expanding number of alternative schools built around distinctive themes, philosophies, and programs. Schools were henceforth to be identified with programs, not with buildings. A given building might house a number of very different schools, each with its own “director” (a teacher with the responsibilities of principal), staff, and student body.

Schools control their own admissions decisions. They are largely free to make their own decisions about programs, methods, and structure. Many of the preexisting formal rules imposed through collective bargaining or dictated by the central office have been waived or simply ignored.

The names of some of the district’s Junior high schools help illustrate just how spectacular the variety can be when the supply side is liberated: the Academy of Environmental Science, the Creative Learning Community, the East Harlem Career Academy, the East Harlem Maritime School, the East Harlem School for Health and Bio-Medical Studies, the José Feliciano Performing Arts School, Music 13, the Isaac Newton School for Math and Science, Northview Tech for Communication Arts and Computer Science, the Rafael Cordero Bilingual School, the School of Science and Humanities.

Freeing up the supply and governance of schools has not led to the kind of chaos or unfairness that critics of choice commonly predict. Because schools go out of business if they fail to attract enough clients, the matchup between schools and students is surprisingly close. In recent years, 60 percent of the students have received their first choices, 30 percent their second choices, and 5 percent their third choices.

The results have been dramatic. While only 15.9 percent of the district’s students were reading at or above grade level in 1973, 62.6 percent were doing so by 1987. Once the worst district in the city for reading and math achievement, by 1987 District No. 4 ranked around the middle of New York’s 32 districts.

Nonetheless, the reform in East Harlem still suffers from what may turn out to be a fatal flaw. Its flowering was nurtured by a small group of visionaries who used—and often exceeded—their authority to liberate schools from district control. But the institutions of political control are still in place. The same public authority that let East Harlem’s schools go their own way could one day reclaim them.

The future of public education

Effective schools, as we have seen, exhibit certain characteristics: strong leadership, clear and ambitious goals, strong academic programs, teacher professionalism, and staff harmony. These are not independent forces, each of which adds a little bit extra to a student’s learning, and which together add up to an extra year. Rather they are integral parts of a coherent syndrome, signs of a functioning school organization. They go together. The key to effective education is not to implement a miscellany of reforms but to unleash the productive potential of schools and their personnel.

Reformers have learned to revile bureaucracy and praise autonomy. But most also believe autonomy can be “granted” to schools within the existing framework of political control. They believe that the very same superintendents, central office administrators, school boards, city councils, and state governments that created the education bureaucracy can be persuaded to dismantle it, and nurture a new population of autonomous schools. They are wrong; it is precisely these political institutions that breed bureaucracy.

Moreover, the other factors that produce large differences in school organization—student ability and behavior, and parental socioeconomic status—are largely out of the control of schools or school reformers. Institutional structure is not only the major factor shaping schools, it is the only factor that reformers can do anything about. We obviously cannot legislate bad students out of existence, or give all American schoolchildren interested, well-educated, affluent parents. But we can give schools the freedom they need to do a good job.

In New York City especially, choice seems the obvious answer. District 4 is already here as a model. Chancellor Fernandez is committed to school-based management, which could help free up the supply side of the system, provided the bureaucracy does not reassert control the first time a school does something innovative enough to make it nervous. But the only way to ensure that the bureaucracy does not crack down, and that the schools really are free to serve the best interests of students, is to liberate the demand side as well; to make schools accountable not to the bureaucrats but to parents.

The reforms we propose have nothing to do with vouchers, or with “privatizing” the nation’s schools. The choice system outlined would truly be a public school system. (See “A Blueprint for Choice,” p. 50.) There is nothing especially “public” or democratic about shackling schools with layers of authority remote from the persons most concerned with learning: parents, children, teachers, and principals.

It is fashionable these days to say that “choice is not a panacea.” We think reformers would do well to entertain the notion that choice is a panacea. Choice should not be placed in the same grab bag as other piecemeal reforms. It has the capacity all by itself to bring about the kind of transformation reformers seek. The other reforms, including school-based management, cannot be implemented without the cooperation of a bureaucracy. They perpetuate the very structure that is suffocating American education. Choice transforms that structure.

Only choice can free public schools from the dead hand of bureaucracy, only choice can free schools from conflicting political agendas to concentrate on their primary goal: academic achievement. It is time to change the way we think about, and govern, our public schools. Our children deserve no less.


What would a choice system look like? Taking into account all that our research has shown, we think it ought to look something like this:


  • The state would set minimal graduation, health, and teacher certification requirements for chartering all “public schools.” Any school, currently public or private, that successfully applies for a public charter would be deemed a public school, eligible to receive public funds. Local school districts would retain authority over the schools they run, but not over any other state-chartered schools within their boundaries.
  • The state would also maintain a “choice office” in each district with a few simple duties: to maintain a record of all school-age children and the level of funding (or “scholarship” amounts) associated with each child; to forward that money to schools; and to inform families about the school choices available.


  • Federal, state, and local education funds will take the form of “scholarships,” which will flow to the public school of the parents’ choice. Parents may not “add on” money for their own children’s tuition, but school districts may decide to tax themselves at a higher rate in order to provide their children higher scholarships. The state contribution may be larger for children from poor districts than for children from wealthy ones, to help equalize scholarship amounts. Local, state, and federal governments may also add to the scholarships of underprivileged, troubled, or special students, stimulating the emergence of new programs and new schools.


  • Each student will be free to apply to any public school in the state. Schools will make their own admissions decisions, subject only to non-discrimination requirements. This is absolutely crucial. Schools must be able to define their own missions and build their own programs; they cannot do this if their student population is thrust on them by outsiders.
  • Students not accepted anywhere will make a second application to schools with unfilled places. After this second round, some students may still be without schools. Parent liaisons will take informal action to match up these students with appropriate schools.


  • Each school will determine its own governing structure. The state may not impose requirements for career ladders, advisory committees, textbook selection, in-service training, preparation time, homework, or indeed anything else. Local school districts may retain school boards, supervisors, and other governing apparatus, but they must pay for them entirely out of the revenue they derive from the scholarships of those students attending district-run schools.
  • The state’s responsibilities include making sure that schools meet their charter criteria, abide by non-discrimination laws, and inform the public about staff and course offerings, parent and student satisfaction levels, and standardized test scores. State governments will not, on the other hand, police school quality or student achievement. For performance, schools are held accountable from below, by parents and students who directly experience the results and are free to choose.

City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next