Read an Excerpt
Virtually everyone who teaches in our K–12 schools is aware that most school districts today are run the way businesses were forty or fifty years ago: top-down, autocratic, and bureaucratic. Teachers and principals have to use books that are selected by someone in the central office who does not know their school; they must follow a daily schedule set by that central office; and all schools are assigned the same complement of teachers, attendance clerks, and so on—according to the central office formula. Students have to attend the “zone” schools to which they are assigned by that same central office. That educational straitjacket is widely disliked by families and by teachers, but until recently, there has been no alternative.
Now, however, a quiet revolution is under way. That revolution allows each school in some districts the power to make more and more of its own budgetary, personnel, and instructional decisions and permits families to decide which public school their children will attend. Power is being shifted from central office staffs to families, principals, and teachers, and the results so far show that students are doing better, parents and teachers are more satisfied, and school districts may never be the same. It may be too early in this dramatic period of change to declare that top-down management is dead. Most of our nation’s schools still report to strongly centralized staffs. The eight districts that have been pioneering this move to decentralization are inventing and learning as they go. Some of them have made costly missteps. Their story is not one of easy change—it’s a nuanced story of victories and of defeats. This book describes that quiet revolution, what it means, and how every school district can learn from the successes and the failures of the pioneers.
Most people know of at least one excellent public school, perhaps a school that they attended or one they worked in or that a child they know attends now. High-quality public schools offer living proof that public education can rise to the level of excellence, yet they also stand out as exceptions to the rule and thereby raise questions of incalculable consequence for American public education: Is it possible for a school district with dozens or hundreds or even a thousand schools to achieve high quality in virtually every school? Is there something about large school districts that inevitably kills the spirit of the teachers and principals, or is there a way to general excellence through fundamental changes in how we organize public education in this country?
The answers, as I will demonstrate in this book, are in the positive. Ask yourself what you believe is the single most important factor in the success of a school district. Ask anyone you know the same question. Ninety percent of the time, the answer will be “teachers.” Then ask how, as a practical matter, a school district’s student performance can be raised. Now the answer won’t be so clear. This book will demonstrate that the most important element to the success of a school district is management. Good managers will organize schools effectively and will attract and retain good teachers. Without good management, even talented teachers will fail.
I am a professor at the UCLA Anderson School of Management, a graduate school of business. I have spent more than thirty years studying very large companies, such as General Motors, Toyota, Intel, and Macy’s. In the corporate world, it is considered reasonable to expect high quality throughout a company, even a very large company. After all, Toyota built its reputation by making every one of its plants a high-quality operation, not by having some good plants among several poor ones. McDonald’s built its reputation by making every one of its stores deliver quality service to every customer, every time—not by delivering high-quality service sometimes to some customers. If Toyota and McDonald’s can do this, why can’t large public-school systems? As I have suggested, the answer has to do, more than anything else, with the way public education is organized and managed in this country today.
More than a quarter of a century ago, corporations and scholars recognized that large size brings major negative consequences and that the chief antidote to these pathologies is to decentralize decision making down to the operating subunits. A fundamental flaw in public education in the United States is that this lesson has not yet been applied to school districts.
Over time, school districts have become very large. Since 1932, the number of students enrolled in public schools in the United States has more than doubled, from 24 million to about 50 million. Meanwhile, the number of school districts has declined from about 127,000 to about 16,000.1 This means that school districts today on average have fifteen times as many students as their predecessors did seventy-five years ago. Business organizations, functioning in a competitive world, long ago understood that they could not survive that much growth without decentralizing. In a competitive setting, the rule is decentralize or fail. But school districts, not living in a competitive world, have not changed their form. They remain about as centralized today as when they were one-fifteenth their present size. In fact, as a result of growth in the size of school districts, they are actually far more centralized today, in the sense that in each district, one central office staff now makes the decisions that affect fifteen times as many students as before.
Many districts think they’ve already tried decentralization and found that it did not work. For example, the nation’s three largest cities—New York, Los Angeles, and Chicago—have at one time or another grouped their schools into “communities” of several dozen schools or have created subregions of schools and have delegated limited decision powers to the assistant superintendents who run each subregion. Invariably, these districts were disappointed with the results and soon abandoned the experiment. These ineffective approaches to decentralization were bound to fail, because they did not place decision making where it makes a difference—with the principals. If principals are under the thumb of bureaucrats, it does not matter whether the bureaucrats are in a regional office or in the central office. What matters is whether or not the principal has the executive power required to exercise the leadership fundamental to the success of educating children.
About twenty years ago, I became interested in helping public schools to improve. For many years, I volunteered in a local organization that gave advice to our local school district, and I couldn’t help but notice that more often than not the best schools were outliers and renegades, that they followed a set of principles that were in conflict with the desires and dictates of the school board, the superintendent, and the central office staff. That did not make sense to me. Why would any school district not celebrate its successful schools? Why would it not try to have all of its schools emulate them? I wanted to get to the bottom of these questions, so I’ve spent the past eight years organizing research teams and visiting a total of 665 schools across the United States.
In my first study (of 223 schools in nine districts), I wanted to find out if decentralized decision making—that is, giving power over the school budget to principals—would result in higher student achievement.2 The answer was a clear “yes.” In that study I compared the three largest centralized districts in the country to the three largest decentralized districts and to the three largest Catholic school districts. That study demonstrated that a school district can empower its principals with real, meaningful autonomy. But I did not know at that time what principals did with that freedom to make them successful.
By the time we launched the second round of research, eight large U.S. districts had begun serious decentralization—Boston, Chicago, Houston, New York, Oakland, San Francisco, Seattle, and St. Paul—and we were able to study them all, drawing a total sample of 442 schools from these eight districts. To my knowledge, ours is the only large-scale study of school district decentralization. In all likelihood this reflects the fact that only in the past few years have any U.S. districts implemented true decentralization by giving principals substantial control of their budgets—as much as 87.2 percent of the budget in St. Paul, Minnesota, 85.0 percent in New York City, and 73.7 percent in Houston. Far more typical of the country as a whole are the cases of Los Angeles,3 in which principals control 6.7 percent of their school budgets, and Hawaii, where principals control about 4 percent, according to the Hawaii state auditor. Indeed, as recently as 2001, New York City principals controlled only 6.1 percent of their budgets, before the far-reaching reforms introduced by Chancellor Joel Klein during the administration of Mayor Michael Bloomberg.
None of this is to suggest that success requires only that districts grant a large measure of autonomy to their principals: that step, while necessary, has to be accompanied by four other policies—(1) school choice, (2) development of effective principals, (3) accountability, and (4) weighted student formula budgeting. Together, as we will explore in coming chapters, these Five Pillars of School Empowerment can produce dramatic improvements in achievement for all students, from those in affluent neighborhoods to those in inner-city ghettos. (I will provide information about student achievement in the chapters that follow.)
Let me hasten to add that this book deals primarily with the improvements that can be made in the performance of students if the school district is organized and managed in a decentralized fashion. While organization and management are crucially important and may qualify as the least appreciated aspects of education, there are several other major factors that matter, which are not covered in this book, although they too are of crucial importance. The performance of students is greatly influenced, for example, by the training and skill of their teachers. A great teacher can overcome many bureaucratic obstacles, but even the most dedicated teacher will eventually lose heart in a badly managed school. Also of great importance are a curriculum that is well designed and properly implemented and an approach to standardized testing that makes sense. Also important are the development of very young children even before they begin school and children’s physical and emotional health. These important topics are beyond the scope of this book, which focuses on the management of the school district within which teachers and students interact each day. Based on my research, I can assert with confidence that when a school district grows beyond about three thousand students, the one thing that will most improve the academic performance of students is better management.
When Enron, Countrywide, or General Motors get into trouble, we automatically point the finger of blame at the top corporate management. We assume that the branch managers and the front-line employees are operating within a set of policies that were set at the top and that those policies must be in some way defective. We don’t blame secretaries, customer-service representatives, and auto workers. Nor do we blame the customers, because we know that any large organization has all kinds of customers, some of whom are hard to please, and that it’s the responsibility of the company to learn how to deal with all kinds of people.
For some reason, though, when a school district is in a state of chronic failure, otherwise levelheaded people blame the teachers or the teachers unions or, even more strangely, the students and their parents! Perhaps this is because the public does not typically think of a school district as an organization but rather as a collection of schools, which in turn are viewed as a random collection of students and teachers. As a result, there is a general tendency to explain poor performance as the result of shortcomings of teachers or students, which is really no different from blaming the assembly-line workers or customers at General Motors for that company’s performance problems. A good teacher who is overwhelmed by too many students, confusing schedules, and frustrated parents cannot be successful. An inexperienced new teacher who is assigned to a chaotic school will not develop into a strong classroom performer. Strong teachers are developed, coached, and protected by strong principals.
For their part, strong principals have to be decision makers who can make a myriad of choices each day on the spot—for example, to modify curricula and schedules as needed or to add new reading specialists in place of less important staff positions in order to meet the needs of students and teachers. Principals who are capable of these kinds of decisions aren’t born; they have to be selected and trained and then furnished with a set of district-wide policies that give them autonomy and hold them accountable for student performance as well as for teacher and parent satisfaction. At the same time, the central office staff must be organized to serve the unique needs of each school rather than to enforce a one-size-fits-all policy.
How can school districts be organized for success?
First, by achieving the right balance between staff positions and classroom positions. The tendency in school districts today is to have far too many staff jobs, so that, according to one study, only 43 percent of school district employees nationwide are “regularly engaged in classroom teaching.”4
This isn’t the same as having an insufficient number of credentialed teachers on district payrolls, because as many as one-third of a district’s credentialed teachers may be in staff jobs, not in the classroom. This maldistribution of human resources comes about because state school funds are sent to the district headquarters, which then decides how much of the money to keep for district staff and how much to allocate to schools. Research has shown that, over time and economic cycles, school districts tend to lay off teachers disproportionately (instead of staff) when budgets are tight and then to hire both teachers and staff when times are good.5 As a result, with each economic cycle the staff become a larger and larger fraction of a district’s total employees.
One goal of decentralization is to reverse this trend by sending the money directly to the principals. That way, the central office cannot lay off teachers in a down cycle, and principals are unlikely to do this unless those cuts are unavoidable. The result, we have found, is vastly fewer staff and many more teachers in decentralized districts.
This vicious cycle of centralized districts begets a second organizational pathology. When a district has too few classroom teachers, student loads per teacher rise to the point where teachers can no longer know their students well enough to establish a bond of trust with them. Without this trust, a teacher can neither establish an orderly classroom nor push a student to do his or her best, and the teacher’s job often becomes frustrating and constantly stressful. Under these conditions, many veteran teachers understandably want to escape the classroom and get into staff jobs, and the result is that they create pressures within the district to create yet more staff jobs to which they can migrate. We’ve found that when a district is run in a decentralized way, staff positions decline, teaching loads decline dramatically, and most teachers no longer have the desire to escape the classroom, because now they know their students and enjoy their teaching.
Perhaps most important, successful school districts tend to have low TSLs—low Total Student Loads per teacher. Total Student Load is little known to most people, yet it is crucial to classroom success.
Total Student Load may be the single most important fact to know about a school, particularly a middle school or high school. But by and large, TSL is still largely a secret, and few people know why it is important.
Total student load is not the same as class size, although the two are related. To put it simply, TSL is the number of students that a teacher has to get to know each term and the number of papers that a teacher has to grade each time. For example, in most New York City middle or high schools, the union contract specifies a maximum of 170 students in five classes of 34 students per class. That means that if a teacher assigns a paper, he or she must read, grade, and write constructive comments on 170 papers. Ask any teacher if he or she can do this for two or three ten-to fifteen-page papers each semester, and they’ll tell you that it’s impossible. Such a student load also means that each semester teachers must get to know 170 new adolescents and form a sufficient bond with each one to be able to reach those students, support them, and push them to do their best. Teachers will tell you that’s impossible, too.
Our research shows that Total Student Load has a very powerful effect on student performance, and that the decentralized schools of New York City have achieved particular success by reducing TSL to levels previously unheard of in public education.6 A principal who has control over staffing, curriculum, and schedule can strategically manage these variables to greatly reduce Total Student Loads. In most districts, though, principals have little or no control over these variables, and we have found that most every school in those districts has a high TSL as a result.
It would be useful to be able to compare these Total Student Loads to national data on all public schools, but, since TSL is still largely a secret, no such data exist. We do know that the TSL of 87.7 in New York City’s first Empowerment Schools is low compared to what prevails in most schools across the country. In most school districts, the contract between the teachers union and the district specifies the maximum class size and number of classes per teacher that is permitted. From these figures, we can calculate the maximum TSL. In Los Angeles, for example, the maximum high school TSL permitted by the teachers union contract is 225 (five classes of forty-five students), although many teachers are above that figure. In Clark County, Nevada, where school construction has not kept up with the explosive population growth in Las Vegas, TSL at some high schools reaches 260, as teachers are encouraged to “sell back” their free or prep period and teach an extra class for additional pay. In doing so, teachers are giving up their allocated time each day for grading papers and for preparing for the next day’s classes, which must negatively impact the quality of instruction in their classes.
As we’ve indicated earlier, Total Student Load is a different concept from class size. Consider a teacher who teaches five classes of thirty-five students for a TSL of 175. We can reduce her TSL to a much more manageable 105 either by reducing the number of her classes to three or by reducing her class size to twenty-one. Both solutions would be far too expensive, however, to be implemented on a large scale in our public schools.
The challenge for management (that is, principals and superintendents) is how to get TSL down to the neighborhood of 80 students per teacher in a way that is affordable for a public-school district. What decentralized districts have done—New York City most consistently—is to empower principals to find creative ways to drive down TSL without increasing costs. This cannot be done by top-down, central office mandates, but, as we will see in the following chapters, it can be done by empowering principals to make their own budget decisions. With budget autonomy, principals can work with teachers and staff to find ways unique to their school to reduce TSL dramatically.
One school, for example, may not need or want security guards or professional development staff, while another may not need attendance clerks or registrars. Allowing principals to make these staffing decisions and to use the money saved for more classroom teachers is part of the solution to reducing TSL.
Principals have also used their freedom to shape curricula and schedules to reduce TSL. In several schools we found that principals had combined the English class and the social studies class into a single humanities course or had combined math and science into a Math-Science Integrated (MSI) course. At least one principal invented a way to schedule classes so that all students could have a one-on-one “office hour” visit with any teacher they desired. We will describe these initiatives in greater detail in the chapters to come. The result of these kinds of adjustments is often a dramatically lower level of TSL, but without any increase in school budgets.
As part of our research we analyzed three years of data on student performance at each of the eight decentralizing urban districts. In analyzing student performance, we tested the other variables that might affect student performance, including class size, teachers’ years experience, percent of teachers who have full or partial credentials, use of professional development coaches, minutes devoted per week to math and to literacy instruction, and a few dozen additional factors. Among these, only TSL had a noticeable effect on student performance in every district, and that effect was large.7
In New York City, for example, we found a virtuous relationship among three factors—the principal’s control over budget, Total Student Load, and student performance. An increase of ten percentage points in the principal’s control of budget was associated, all else being equal, with a reduction in TSL of twenty-five students per teacher and an increase of eleven percentage points in the percentage of students who reach the federal government standard of “proficient or above” in standardized math tests. Further studies will be required to tell us whether, as seems likely, there is a ceiling on this effect as a principal’s autonomy rises to even higher levels. Even so, the magnitude of this improvement in performance is far larger than that of any reform strategy tested in other research. Studies of school reform typically reveal gains that are a fraction of what we found.
I’ve already noted that we analyzed several different kinds of innovations that schools often use, and we found that the only change that has a noticeable effect on student performance is reduction in Total Student Load. While those other variables are surely of great importance, I am now drawn to the conclusion that focusing on these individual factors is like pouring new wine into old wineskins. It’s the container—in this case the school district—that must be radically redesigned, if changes in curriculum and professional development are to be truly effective.
Total Student Load is a cold statistic that represents a set of human relationships between teachers and their students. A low middle or high school TSL, in the range of 80, means that teachers can know their students and that students are aware that teachers understand them and care about them to the point that, if they slack off or miss school, the teacher will call home. A low TSL may also mean that teachers have enough time out of the classroom to make themselves available to their students and to encourage students to come to see them during office hours. One-on-one encouragement that a student is too embarrassed to seek in front of a class can make the difference between a student’s giving up and continuing to work hard in school.
School districts do not measure and report to parents or the public the degree of trust between students and their teachers (except in New York City, as we shall see later). Nor, they feel, is there a practical way for districts to measure or report on student motivation, which is a direct consequence of the student-teacher bond (New York City now measures this, too). Every teacher will tell you, though, that until students trust a teacher, they will not open up to the teacher’s efforts. I can affirm this myself as a classroom business school teacher of thirty-eight years.
How can a teacher establish a personal bond of trust with 140 or 225 or more new students each term? Experience shows that it cannot be done. When TSL approaches 80 students per teacher, however, this bond does develop and student motivation and performance rise, sometimes dramatically, on a district-wide basis.
School officials often complain to me that decentralization or school empowerment means trying to turn principals into CEOs of schools; principals, they will say, are not meant to be business managers but instructional leaders. Many principals, to my dismay, echo this sentiment.
I confess that I’m always chagrined at this complaint, even though I understand what motivates it. Most school districts, after all, have largely taken their principals out of decision making and so have often selected as principals people who don’t want responsibility, authority, or accountability. Almost all school districts have standard staffing formulas—the number of teachers of each kind, of assistant principals, attendance clerks, assistant registrars, counselors, and so on, depending on school enrollment. Principals are generally not involved in deciding either how many or which teachers, assistant registrars, or attendance clerks to hire; it is as if these individuals are widgets. As we shall see, though, a math teacher who would also like to teach a gym class is not a substitute for a math teacher who only wants to teach math, and the decision of which one to hire should be left to the principal. Empowered principals in the decentralized districts of our study overwhelmingly involve teachers in these decisions, sometimes to the point of having the entire teaching staff vote on the decision, which is then binding.
Unsurprisingly, principals in traditional, centralized districts have become accustomed to having little power or authority. As a tradeoff, they also have little accountability, which explains why weak principals can manage to hang on to their jobs for what seems forever. In this arrangement the central staff are happy because they enjoy virtually unquestioned power; principals are happy because they have total job security and are not accountable for anything; but teachers, students, and parents suffer the consequences.
To principals who object that they should be instructional leaders instead of executives, I have some pertinent questions: What does it mean to be the instructional leader of your school? Do you control the staffing? How about the use of teacher’s aides? How about the choice of the teachers you hire? Do you get to set the curriculum that would make the most sense in your school? What about choosing the books and materials? How about selecting the bell schedule (number of periods a day)? How about the professional development budget?
If the answer to all those questions is no, as it almost invariably is, then what does it mean to be called the instructional leader? If the school district decides on how to staff each school, which personnel to send to each school, and what books the schools must use; if principals do not decide what professional development courses need to be provided to teachers but instead have to send teachers to courses selected or designed by the central staff; if principals do not control the curriculum, schedule, or budget, then in what way is the principal the instructional leader of the school? What else is there to control that is instructionally important?
If there is any vital job for a principal, it is precisely to control the major instructional variables of budget, staffing, curriculum, schedule, and professional development. Those are the levers through which the instructional leader of a school, the principal, can build a great school by attracting, developing, and retaining strong teachers.
Large school districts that follow the traditional centralized approach to management have a chronic shortage of qualified principals. That should come as no surprise. No one wants a job that has great responsibility and no power and in which he will be publicly criticized for all failures. New York City’s school district used to be like that. New York City had a chronic shortage of credentialed principals back when it was a centralized, top-down, enrollment-formula district. Today, with its new decentralized approach that empowers principals, New York City makes sure that new principals are trained and ready. Some of the new principals are being trained in a fifteen-month program run by the Leadership Academy, and all new principals receive at least one year of coaching support through the academy. Instead of a shortage of principals, New York City now has approximately six hundred qualified applicants each year for the sixty-five openings in each new class at the Leadership Academy, about which we will learn more in a later chapter.
The decentralizing districts that we have studied are moving to replace their large schools with small schools of about 350 to 600 students each. A few large schools remain and are successful, and there is no reason not to keep these schools as they are. The trend, though, is to replace large, 3,000-student schools with small schools. Small schools are advantageous in large part because they are easier to manage. A small school will have about twenty to twenty-five teachers and only three to five non-teaching staff. Inexperienced but bright neophyte principals who have proven themselves as excellent teachers can handle that assignment. The experience in our eight cities shows that there are large numbers of capable teachers who are eager to become principals in these small schools in districts that give principals autonomy. All of a sudden, the job of principal can become more attractive, just as reducing Total Student Load can make being a teacher attractive.
Many district officials will agree that this seems to be an appealing idea but will then plead that switching to small schools is just not possible, because the capital funds are lacking to build all of the new small schools. To these people I would say, visit New York City. I have visited school after school in that city that has made the conversion from one large school to several small schools with a capital budget of fifty dollars.
How is that possible? In one case, I visited an old building of four floors that housed a school of 3,200 students in the South Bronx. The district went to this school with a bucket of yellow paint and a large paint brush (thus the fifty dollars). On each of the four floors, they painted a bright yellow stripe down the middle of the floor and declared that henceforth the right side of the hallway would be one school, and the left side of the hallway would be home to another school. In this way, they turned the four floors into eight schools of 400 students each. On average, though, the conversion of an old large school into several small schools in New York City does necessitate some reconfiguration, and the average cost for converting each small school is about $100,000, according to the district.
It’s true that the new principals did not have nicely designed offices and that all of the schools have to share a library, lunch-room, gymnasium, and other facilities. The principals have a council that meets regularly to coordinate use of the shared resources. There is no question, however, that small schools are more manageable, that the number of candidates who can effectively manage small schools is vastly larger than the pool of qualified principals for very large schools, and that inadequate capital funds is a weak excuse for failure to take this step.
Here a cautionary note is in order. In some school districts, small schools have been created en masse through large outside grants. In some cases, the districts that accepted this money did not understand the concept of empowering schools. As a result, they created lots of small schools but did not empower their principals. What they demonstrated is that if you convert a large, failing urban school in a centralized district into several small schools that are also run in a centralized, top-down fashion, you will now have several small, failing urban schools! Unfortunately, these districts may think that they have tried decentralization, when in fact they have not.
Many school districts think that they’ve already tried decentralization and have found that it did not work. As I indicated earlier, they most likely tried superficial approaches that were bound to fail, because they did not place decision making where it makes a difference—in the hands of principals.
Many districts have implemented site-based management, commonly known as SBM, and have found it wanting, too. That may seem pretty strong evidence that decentralization does not work, because SBM involves the creation of a school site council at each school. Yet the mere fact that a school council has been created does not mean that either the school or the principal has decision-making power. Moreover, even if school councils did have real power, that fact would vitiate the accountability system, thus creating another serious problem.
The new approach to school empowerment typically grants control over the budget to the principals, not to school committees. Some districts, such as San Francisco and St. Paul, do require that every school have a site council made up of teachers, parents, and community members and that the council have the power to approve the school plan and budget, but other empowered districts, such as Houston and Oakland, have councils that are advisory and do not have final decision powers. New York City vests the decision power in its principals but requires that each principal consult with a school site council.
It is prudent to grant budget autonomy to a principal? Not only is it prudent, it is absolutely necessary if there is to be any accountability. In 1988, Chicago implemented a state law that provided that every school in the city had to have an elected council of parents, teachers, and community members. Councils had the power to hire and fire principals and to approve a large part of each school’s budget. What followed was infighting and accusations of graft, patronage, and sometimes of outright stealing by members of the school councils. In the end, a new mayor, Richard Daley, persuaded the state to pass another law that gave the mayor the power to appoint the city’s school board and through it to control school budgets.
A related example comes from New York City. In 1966, the state legislature passed a law that created elected school boards for each of the thirty-two community subdistricts in New York City. Only about 5 percent of eligible voters turned out for local school board elections, though, and in some cases a board member was elected with as few as 238 votes.8 The result, as in Chicago, was chaos. After more than two decades of grand jury investigations and exposÉs by watchdog groups and newspapers, the citizens of New York were fed up with the widespread impression that these local school boards had used their vast budgets to become centers of patronage and political power. A new state law, passed in 1996, stripped the local boards of their budget control, giving all financial control to the chancellor of the New York City Department of Education, where it remains today.
Accountability has to mean that an identifiable person or persons bear direct responsibility for the use of budgets and that, if they use the money badly, they will either lose their jobs or suffer some other major consequence. Where school site committees with decision powers are used, accountability is compromised, because parents and community volunteers cannot be fired or sued for poor school performance, and no teacher can be held accountable for anything that is not included in the union contract. This is not to say that a principal should exercise totalitarian control over the school budget. To the contrary, every district that has empowered principals with budget control has also created an elaborate accountability system to see that teachers, parents, and others are part of the decision process, as we shall see in the section on accountability.
Some people argue that school district decentralization is nothing but a weak form of what they consider to be “true” choice—vouchers and charter schools. They may be right, in the sense that decentralization, as we shall see, has strong features of choice, because all of our eight districts permit families to pick any public school within the district (with preference given to neighborhood children) without seeking a waiver from their local or “zone” school. Why not go all the way, these people argue, and permit families to receive a voucher that they can spend at any public or private school?
My answer is based on practical experience. California has had two initiatives over the past fifteen years that would have created statewide vouchers. The first initiative was backed with a campaign of some $50 million, the second with a campaign said to have cost $70 million. Both initiatives lost by very large margins of about 70 percent to 30 percent. Anyone who has been involved in politics will tell you that when you spend that much money on a campaign and still lose that big, it’s time to move on to the next issue.
Nonetheless, there are successful voucher systems in some states, and the United States Supreme Court has ruled that vouchers may even be used to attend Catholic schools under certain conditions. Perhaps a time will come when a system of vouchers that will provide equality of access will be invented. Until then, vouchers will remain a very small part of our educational system.
Charter schools, in contrast, have grown briskly since 1991, when they were first permitted in the state of Minnesota, and they now enroll about 3 percent of all U.S. students in the forty states that have passed charter school laws. Charter schools enroll between 25 percent and 55 percent of the students in Dayton, Detroit, New Orleans, and Washington, D.C., and about 7 percent of all students in Los Angeles.
Charter schools bear a strong resemblance to strongly decentralized school districts. For one thing, they are public schools: they charge no tuition, and they must accept all comers, though in some cases they may require an admissions test (as some district public schools do, for that matter). Charter school principals typically enjoy full autonomy in much the same manner as principals of empowered or autonomous schools in St. Paul, Boston, or New York City. In fact, one could argue that a fully decentralized school district operates as though it had converted all of its schools into charter schools.
It is also notable that every one of the eight decentralizing urban districts in our study except Seattle is competing against charter schools (the state of Washington does not yet permit charter schools). In many cases, our research team concluded that it is the success of charter schools in raising student achievement that has encouraged superintendents to decentralize, since charter schools have proved that public schools can be run with a high degree of autonomy and be successful. The popularity of charter schools with teachers and parents has also made the case that small schools are desirable, and the freedom of choice that has drawn many families to charter schools has also had an impact on school district managers.
Charter schools may well have an even more profound benefit for public education. Charter schools are always schools of choice. No student is assigned to a charter school. Charter schools can survive only if they can continue to attract enough students to make their budgets. In other words, they compete against other charter schools and the school district for students, even though they sometimes partner with that same district by participating in joint planning and other activities.
We do not know whether the decentralization of our eight large school districts will survive the test of time. Even if decentralization produces superior student achievement, the world of political decisions is such that a new president of the United States, a new mayor of Boston, Chicago, or New York, or a new governor of California, Minnesota, Texas, or Washington could sweep away even the most successful school district reforms. One force that will push back against any potential backtracking in decentralized districts is the continued success of charter schools. A large and successful population of charter schools in any city serves as a constant, visible reminder of the efficacy of local school autonomy and of school choice. In the face of that daily reminder of the strength of decentralized decision making, it will be far more difficult than it would be otherwise to reverse the progress that is taking place in these eight districts.
That being the case, why don’t we simply convert all school districts to charter schools, rather than attempt the arduous task of reforming those districts? My answer is that these decentralizing districts are, in effect, doing just that. If Boston or Chicago spreads its pilot-school model to all of its schools as Houston, New York City, Oakland, Seattle, St. Paul, and San Francisco have done, then these districts will be indistinguishable from very large charter school systems, with the one difference that they will have traditional union contracts, while charter schools tend to be nonunion. The Boston Pilot Schools have an abbreviated union contract, but all of the other seven districts have found ways to bring about decentralization while maintaining more traditional union contracts (other than Houston, because Texas does not permit collective bargaining, but the state does have associations of teachers). As long as collective bargaining is not incompatible with successful school empowerment (and apparently it is not), there is no reason to doubt that decentralized districts can be as successful, or even more successful, than charter schools.
There is another element in the argument about replacing existing school districts with charter schools. New York City has 1,467 schools and Los Angeles 878. Many of these are large schools. To replace these schools with small charter schools (most charter schools to date are small) would mean having perhaps 3,000 charter schools in New York City and 1,700 or more in Los Angeles. As a practical matter, who is going to open and run that many schools? Will it be thousands of mom-and-pop operators, each with one or two charter schools? Who, then, will oversee the quality of education and the proper use of public funds at these schools? No county or state is equipped to do so, and if those education agencies were to build new bureaucracies to take on the task, the result twenty years from now might well be an even larger and more cumbersome set of educational bureaucracies.
Clearly, any practical plan for large-scale charter school growth would require that a small number of very large charter school operators rise up. Each of these would operate perhaps five hundred or a thousand charter schools. Each would then be a large school district and prone to the familiar pathologies of large size. Perhaps with time and good management, these new charter school behemoths could be created. I would argue, though, that it is far more practical, given the success of the eight districts described in this book, to work the problem from the other side, by decentralizing the districts that we have now.
The decentralization of entire school districts is new to the United States, where it dates only to the mid-1990s, but it has been practiced for nearly three decades in Canada, Great Britain, New Zealand, and Tokyo (though not elsewhere in Japan). All of these districts roughly follow the pattern first invented in the approximately eighty-thousand-student Edmonton Public Schools more than thirty years ago.9
Edmonton is a blue-collar city and has the largest percentage of low-income students in the province of Alberta. Edmonton Public Schools operates 197 schools for its approximately eighty thousand students, more than Boston, Oakland, San Francisco, St. Paul, or Seattle. The district must compete directly against local Catholic schools, which are funded on an equal basis by the government and enroll about 30 percent of the city’s students.
In 1973, former principal and hog farmer Mike Strembitsky became superintendent. Applying his common sense and his dislike of central bureaucracy, he invented the system of decentralization that is now being implemented in all eight of our study districts. Strembitsky served for twenty-two years as superintendent, long enough to cement this approach in place. Today, several superintendents later, Edmonton still adheres to this system. The district website includes this statement: “Our Philosophy: … Choice is the foundation of our district’s approach to education … We believe that the ‘one size fits all’ model of education is no longer appropriate.”
More than thirty years after it began decentralization, Edmonton now has a four-year high school graduation rate that is 68 percent and rising, above the 55.8 percent of New York City. Edmonton students take the provincial Diploma Examination, roughly comparable to New York State’s Regents Examinations, or can elect to receive a “school awarded” diploma, as in New York City. Students must pass two of the diploma examinations to receive the provincial diploma. In 2007, 86.5 percent of students passed the English exam (compared to the provincial passing rate of 87.7 percent), and 82.2 percent passed in math (compared to 81.1 percent for the province). These results are impressive, given that Edmonton has a higher proportion of low-income students than any other district in the province. Data on the racial breakdown of scores are not available; Canadian school districts are not allowed to collect data on the race of their students.
Edmonton also pioneered annual satisfaction surveys of students, parents, school employees, and principals. In 2007, 1,057 Edmonton residents who have no children in the public schools were surveyed, and more than 80 percent responded that they were satisfied with the overall quality of education provided by the public schools. Nearly 60 percent of that sample said that they felt the education provided by the district was equal to or better than that provided by private schools, and nearly 95 percent said that they felt Edmonton Public Schools provided a quality of education that is better than elsewhere in Canada. Among parents of Edmonton students, about 97 percent were satisfied with the overall quality of education, and more than 90 percent were satisfied with their child’s school principal. More than 85 percent expressed satisfaction with the district’s Board of Trustees. Among high school students, 80 percent reported that their schoolwork is “interesting,” and 90 percent said that they get help from their teachers when needed. Among elementary students, nearly 90 percent found their schoolwork interesting, and more than 95 percent said that they get help when they need it. In addition, every school administers these surveys to parents, students, principals, and school employees every year and posts the results publicly.
Edmonton Public Schools provides a benchmark by which we can evaluate the progress of the eight U.S. districts of this book. The basic elements of decentralization that were developed and refined over thirty years of experience there have stood the test of time. They make up the framework that has guided the reforms in all eight of our study districts. Of equal importance is that Edmonton has maintained its decentralized organization through several decades, through the normal turnover of elected school board members, and through several superintendents.
As we will see in the chapters to follow, decentralization is more an ongoing process than a fix-it-once-and-leave-it-alone construction project. Once a district has successfully decentralized, its task has just begun. Someday, school districts all over the country might be decentralized. If and when that day arrives, it will be much easier to maintain a decentralized form, because every new superintendent, school board member, principal, and teacher will accept the idea as natural. Now, however, a decentralized district is pretty much alone, surrounded by traditional, centralized, top-down districts and their people. Under those circumstances, it’s a constant struggle to keep decentralization going.
No superintendent stays in office forever, and neither does a school board member or a mayor. Turnover in the governance system is inevitable, and whenever new people arrive at the top of the decentralized district, the autonomy of the schools will face a new peril. Until decentralization someday becomes the national norm, new superintendents and school board members will come in with the traditional, top-down, centralized idea of how a district should be managed. In the stories of the school districts that follow, we’ll see the struggle that sometimes ensues after a major succession in the apparatus of governance. As a result of this constant process of change, many decentralizing districts find themselves living in a nuanced world rather than one that is all one way or another. There is reason for optimism even in the face of these challenges, though, because the heart of a decentralizing district lies in the entrepreneurial, independent spirit of the principals and teachers. Once they have tasted the freedom that goes with decentralization, they never entirely lose their desire to keep it.
Now that we’ve laid out some basics, we’re ready to dive into the subtleties and the practicalities of empowering schools in a decentralized district. Let’s start our coast-to-coast tour of school district decentralization.
© 2009 William G. Ouchi