Find more Kappan articles in the searchable PDK Publications Archive

What General Motors Can Teach U.S. Schools About the Proper Role of Markets in Education Reform

By Richard J. Murnane and Frank Levy

It is human nature to search for "magic bullets" that promise to improve performance without painstaking effort -- but choice plans, charter school programs, and school-based management initiatives are not magic bullets, Messrs. Murnane and Levy warn. These approaches will contribute to better schooling only if they stimulate change based on the Five Principles discussed below.

MARKET-BASED initiatives have been among the most durable types of reforms to be proposed for improving American education. Over the last 25 years public school choice, tuition tax credits, and vouchers that families could use to pay for education at private schools have all been advanced as ways to raise student achievement. More recently, school-based management and charter schools have emerged as strategies both to stimulate initiative and to hold schools accountable.

All these proposals are based on a single underlying theory: if schools were free to design their programs and to market these programs to families, U.S. education would improve. In this article, we argue that the theory is neither right nor wrong, but incomplete. Well-designed market-based initiatives can be one-half of effective education reform. But without the other half of reform, even well-designed initiatives will accomplish little. And under any circumstances, badly designed initiatives can do great damage. The argument begins with a 15-year-old episode from a surprising source, General Motors.

The Lessons of Hamtramck

In the early 1980s General Motors was in trouble. Production costs were high. Product quality was low. Consumers increasingly were choosing competitors' products, and General Motors' market share was declining. The market was sending a clear signal: GM had to change the way it was doing business. But what kind of change? Much of GM's labor force worked with dated equipment and had only minimal skills. The workers were further limited by the web of rules written into union contracts. Many General Motors plants had more than 75 job classifications, each with strictly defined duties.

Confronting these problems was like confronting the Gordian Knot: renegotiating work rules, upgrading training, reorganizing management, unraveling a history of weak productivity. But if bureaucracy and work rules were GM's weaknesses, money was its strength. And so GM evolved a strategy to spend its way out of problems by investing in robotics. Robots -- robots that Ford and Chrysler could not afford -- would be the sword to sever the Knot. Problems in specific areas would not be solved; they would simply be bypassed by computerized machines.

The showcase for these ideas was to be GM's Hamtramck plant, a newly built Cadillac plant outside Detroit.1 At Hamtramck 260 robots would handle assembly, welding, and painting operations. Parts would be moved from place to place by 50 automatic vehicles. Laser beams and cameras would be used to monitor quality. Since the point of the equipment was to bypass organizational problems, most of those problems were not addressed. Frontline workers received minimal new training. Management structures were left unchanged. Production plans were based on the expectation that the new equipment would work right the first time.

Hamtramck was a disaster. Robots failed to install parts or apply paints properly. Occasionally they destroyed the car on which they were working. Because expectations had been so high, backup systems were weak. Getting the right people to fix the robots often required several hours, during which the assembly line was idle. Employees were both frustrated and scared. The cars themselves were awful. Hamtramck was giving the rest of the world lessons in what not to do.

For every manager -- a CEO, a school principal -- the Hamtramck experience holds two lessons. First, consumer choice can signal an organization when change is needed -- no small accomplishment. But consumer choice cannot tell an organization how to change. Second, it is human nature to search for magic bullets -- the new technology or set of rules that will painlessly solve problems -- but this search distracts attention and saps energy from the real work that successful change requires.

GM and the Schools

General Motors' problem in the early 1980s was not that its cars were worse than those the company had produced a decade earlier. Rather, GM's cars had improved slowly while the competition had improved more quickly, and so GM cars of the early 1980s were not good enough. American public schools today face a similar problem. The education American schools provide to students today is not worse than it was in the early 1980s. Indeed, average math scores from the National Assessment of Educational Progress were a little higher in 1992 than they were a decade earlier, and reading scores were about the same.2 The problem is that education that was good enough to obtain a middle-class job in 1979 is not good enough in the economy of today.

During the past 15 years, the skills required to succeed in the American economy have changed radically, but the skills taught in most schools have changed very little. As a result of a growing mismatch between the skills of most graduates and the skills required by high-wage employers, a U.S. high school diploma is no longer a ticket to the U.S. middle class.

As late as 1979, a 30-year-old man with a U.S. high school diploma earned a yearly average of about $28,300 (all income figures are in 1994 dollars). That income, combined with a wife's earnings from a part-time job, secured the family a solid place in the middle class. Then, almost without warning, the economy changed. By 1983 U.S. manufacturing, threatened by imports, was rapidly downsizing, and a 30-year-old man with a high school diploma earned an average of $23,300 a year. By 1993, with computers transforming both U.S. manufacturing and U.S. services, a 30-year-old man with a high school diploma earned an average of $20,500. The significance of this decline in earnings becomes all the greater when we realize that in 1994 nearly half of all 30-year-old men had not gone beyond high school.

Improving a school's performance is at least as hard as improving the performance of a GM plant or of most other organizations. It requires enlisting the energies and knowledge of frontline workers -- the people who do the work each day. This is hard enough in a business where all the frontline workers are on the payroll. In a school, frontline workers include the teachers, who are on the payroll, but also students and parents, who are not.

In such organizations, simple top-down commands don't work. Rather, best-practice organizations (both schools and private sector firms) have learned to manage through a set of five basic ideas -- Five Principles -- to elicit the frontline effort that improved performance requires. At first glance, the Five Principles appear to be simple common sense:

Ensure that all frontline workers understand the problem.

Design jobs so that all frontline workers have both incentives and opportunities to contribute to solutions.

Provide all frontline workers with the training needed to pursue solutions effectively.

Measure progress on a regular basis.

Persevere and learn from mistakes; there are no magic bullets.

Putting the principles into practice is a different story. Building parental understanding of the need for higher student achievement, designing teacher training programs that change behavior in the classroom -- this is the hard and dirty work of education reform. It is human nature to try to avoid this wrenching work and instead to search for "magic bullets" that promise to improve performance without painstaking effort.

For GM, the magic bullet was robotics. And, of course, robots had their place: they could weld and paint and do other physically demanding jobs very well. The bad idea was seeing the introduction of robotics as a substitute for the hard work of retraining workers, improving incentives, and redesigning jobs.

In a similar way, choice plans, school-based management proposals, and charter schools have their place. When they are well designed -- an issue we return to below -- they contribute to school improvement by providing strong signals of the need to change. These market-based initiatives can also make change easier by shifting decisions about resource allocation from centralized bureaucracies to school-based teams.

At the same time, these initiatives raise two dangers. The first is that they will be seen as a complete answer to the nation's educational problems -- a foolproof path to educating children for tomorrow's jobs. This is simple hype. Just as the market could not tell GM Hamtramck how to change, market-based initiatives cannot tell schools how to accomplish the work embodied in the Five Principles. Believing otherwise wastes crucial time and sets the stage for failed expectations.

The second, more subtle danger involves the design issues noted above. The nation's schools and GM both produce a product. But, continuing the analogy, they operate in different markets using different technologies. In terms of technology, GM faces a clean separation between production and demand. The quality of GM cars depends on activities in the plant -- not on who the customers are -- and so GM can focus on selling cars to anyone who wants to buy them. A school's technology is quite different. Classroom learning depends on the quality of the teachers but also on the nature of the students. Student "consumers" are part of the production process, not only for themselves but for their classmates as well. Some students are assets to their classmates; others are liabilities. From a school's perspective, finding ways to exclude the liabilities is a powerful means of improving performance.

Turn next to the market, where GM is judged by the level of its profits. No one assumes that GM should provide cars to families that cannot afford its prices. Schools operate under different rules. Under the constitution of virtually every state, every child is entitled to a good public education -- a very different regime.

Taken together, the differences in technologies and markets suggest that an appropriate standard for judging a school's performance is the extent to which it provides a good education to its students without making it harder for other schools to do the same. A market-based initiative that enables some schools to improve their performance by selecting only students who are assets to their classmates and leaving to schools of last resort students who are perceived as liabilities does not contribute to the goal of providing all students with a good education.

Well-designed market-based education initiatives have much to recommend them. But as we show below, these two issues -- design details and providing the "other half" of school reform -- will determine whether the initiatives will actually improve U.S. education, will be just another political fad, or will reduce the quality of education provided to some children.

School Choice

To begin to get a feeling for the Five Principles in action, consider the case of school choice. The most enthusiastic advocates of school choice are convinced that "Choice is a panacea."3 Even those with less enthusiasm argue that, if schools had to compete for students, they would be forced to provide what parents want. And because parents want their children to achieve, the market will force schools to place greater emphasis on improving student skills. If these assertions were true, the first and fifth principles would be wrong. If parents really understood the problem, well-designed school choice would be a magic bullet.

But what does the evidence say? The assertions rest on an important assumption: that parents find their children's current achievement wanting. Attitudes are beginning to move in this direction, but they have a long way to go. Consider these facts:

In the 1996 Phi Delta Kappa/Gallup poll, 66% of parents gave their oldest child's school a grade of A or B.4

In 1992 almost half of the nation's 17-year-olds could not compute with fractions, decimals, and percentages; recognize geometric figures; and solve simple equations.

In 1992 more than half of the nation's 17-year-olds did not read well enough to find, understand, summarize, and explain relatively complicated information, and almost two-thirds could not respond to queries by writing complete statements that contained sufficient information.5

These math and reading skills that almost half of our 17-year-olds lack are some of the critical skills that labor market entrants -- males and females, blacks, Hispanics, and Anglos -- need in order to obtain and keep a middle-class job today.6

These facts should not surprise us, since evaluating a child's education is much more difficult than evaluating a car. When parents visit a school, certain things stand out -- the safety, the discipline, the teachers' enthusiasm, the size of the classes. The school's ability to prepare students for the job market is much harder to judge. What parents do see is the fact alluded to above -- that today's schools are teaching to slightly higher standards than the schools of 20 years ago. What parents cannot see is that the labor market has changed far faster than schools have improved. In this kind of market, it is easy to picture how choice can lead to safer and more pleasant schools -- no small accomplishment. It is harder to picture how choice automatically raises student skills.

The picture is confirmed by an experiment now under way in Milwaukee. In the spring of 1990 the Wisconsin legislature enacted the Milwaukee Parental Choice Program, a voucher program aimed at low-income children in the Milwaukee public schools. The program is restricted to children in families with incomes less than 1.75 times the poverty line (about $20,000 for a family of four). The students can use the state payment, $2,987 in 1993-94, to attend private, nonreligious schools. The private schools accept the state payment in lieu of tuition.7

Has the program worked? The Wisconsin State Department of Public Instruction commissioned an evaluation team headed by John Witte, a political scientist from the University of Wisconsin. Supporters of school choice have criticized the evaluation, calling Witte a "hired gun" whose mission is to discredit the program.8 Despite the vitriol, Witte and his critics agree on four main results:

1. The participating children tend to be children who were not doing well in the Milwaukee public schools.

2. Parents of participating children are more satisfied with their children's private schools than they were with the Milwaukee public schools that their children had previously attended.

3. The participating private schools have lower average per-pupil expenditures than the Milwaukee public schools do.

4. Children in the choice program do not have higher reading and mathematics achievement than low-income students attending the Milwaukee public schools.9

For those with an open mind, these are important findings. The evaluation shows that choice in Milwaukee has given options to students not doing well in public schools, not just to high-achieving students who want to flee the system. Equally important, the plan has increased parents' satisfaction with their children's schools. But to this point, the plan shows no sign of raising student skills.

As always, there are caveats. Catholic schools, traditionally strong in educating low-income students, cannot participate.10 The private schools that do participate work with less money per student than the Milwaukee public schools. The Milwaukee plan has been operating for only five years.11 Nevertheless, the results are consistent with our expectation: choice (or any other broad reform) can complement a detailed program to raise student skills, but it is no magic bullet. There is no avoiding the dirty work embodied in the Five Principles.

The successes of the Milwaukee plan also highlight the importance of design details. Choice in Milwaukee satisfied the parents of low-income students because the Wisconsin legislature designed the plan with these families in mind. But other choice plans distribute benefits in other ways. California's Proposition 174, a 1993 ballot initiative, is a case in point.

Jack Coons is a law professor at the University of California, Berkeley, with a 30-year passion for improving the education of low-income children. Coons led the challenge in the 1960s against tying public school funding to local property taxes. Fifteen years later, he concluded that school finance reform had not done much for poor children's achievement. The way to make schools responsive to low-income parents, Coons decided, was to move to a system of vouchers and school choice.

In 1991 Coons and his longtime collaborator, fellow University of California, Berkeley, law professor Stephen Sugarman, initiated a petition drive to get school vouchers on the California ballot. They designed a plan for all California students that included protections for low-income children. But to get the needed signatures, they had to bargain with groups that had other priorities. As the petition deadline neared, Coons and Sugarman found that they no longer had control of their coalition. Groups that wanted school choice without protections for the poor rewrote the plan to the point that Proposition 174, the voucher initiative that appeared on the California ballot in November 1993, bore little resemblance to the professors' original version. In the Coons-Sugarman plan, for example, a participating school had to select a certain fraction of its students by lottery, and the $5,000 state voucher had to be accepted in lieu of tuition. Together, these requirements meant that low-income students would have expanded options. The revised plan dropped both conditions: a school could select whichever students it chose, and it could impose additional tuition on top of a $2,600 state voucher. What began as a universal choice plan with protections for the poor became a program directed at the well-to-do.12 Coons and Sugarman, angered by the changes, withdrew their support. In their view, no choice was better than a choice plan that didn't help the poor. They were relieved when Proposition 174 was defeated. They have since regrouped and continue to advocate a voucher plan aimed at improving education for poor children. Now as before, the level of support for such plans is unclear. This evidence makes a simple point: school choice is no panacea. To begin with, the impact of any choice plan -- including who wins and who loses -- is determined by its details. Talking about "choice" in broad terms is not useful.13 But even a fair choice plan, by itself, has limited power to boost student achievement. A fair choice plan will stimulate school change. But as in all market systems, the nature of change will be driven in large part by schools' attempts to satisfy consumers. Families that value skills will seek out schools that have organized around the teaching of skills -- schools that embrace the Five Principles. Families that are satisfied with their children's skills will seek out other kinds of schools, and achievement levels will not rise. Choice by itself is only part of a solution to the nation's school problem.

Charter Schools and School-Based Management

The logic that applies to school choice -- the importance of design details and the "other half" of school reform -- applies with equal force to school-based management and charter schools.

Advocates of school-based management argue that constructive school-based initiatives are hampered by school district bureaucracies. In their view, schools would be more effective if program decisions and resource allocations were made by the adults involved in a given school -- principals, teachers, and parents. School boards would hold schools accountable not by checking compliance with rules but by monitoring student achievement levels.

On paper, there are many experiments that test these ideas. In reality, the ideas have rarely been tried. One-fourth of the nation's public schools instituted some nominal form of school-based management during the 1980s. But in the vast majority of cases the central office retained control of crucial decisions.14 In many cases, decision-making authority was also constrained by collective bargaining agreements. The result was many "non-events," and it is not surprising that most attempts at "school-based management" have not led to improved student achievement.15

But even in a proper experiment, a move to school-based management would only begin the process. A school that wanted to raise student skills would still have to begin by getting agreement that this was the paramount goal -- important enough for parents to back up increased demands for homework. It would have to seriously investigate changes in the curriculum, find strategies for providing teachers with the skills to teach that curriculum, and find ways to measure student progress -- in short, the work embodied in the Five Principles.

Under school-based management the school would have more freedom to undertake this work. But the school could also use its freedom to go in quite different directions.

In the years just ahead, we will learn more about decentralized management from the new wave of charter schools. As reported in the September Kappan, 25 states have enacted charter school legislation that permits groups of teachers, community agencies, or profit-seeking firms to contract with a school board to operate one or more schools.16 Charter schools are a promising innovation because they may actually provide the autonomy missing from many school-based management programs. This makes it easier for the school's faculty and parents to agree on a small set of goals, including, if they choose, that all students should master critical skills. Because they stand as distinct entities, charter schools may also have more staying power than a typical school-based management program in which authority is constantly renegotiated. Will charter schools increase mastery of critical skills among the nation's students? The concept is too new to have a track record, but observers should keep their eyes on how individual charter schools answer four questions.

Does the charter school commit itself to a goal, such as mastery of critical skills for all its students, or will it emphasize other goals?

Does the charter school commit itself to serve a fair share of the most difficult-to-educate children, and does it have a strategy for attracting such children -- or will it discourage applications from such children?

Does the charter school's contract with the school district provide enough time and enough financial support for the school to persevere and learn from the mistakes that are inevitable in any ambitious new venture?

Does the charter school commit itself to providing information about student achievement that will allow parents to make sound judgments about the quality of the education their children are receiving?

Individual charter schools will answer these questions differently. Those that do commit themselves to providing all students with critical skills can be a positive force for improving American education, but they will still need the Five Principles to live up to their commitments.

When to Persevere

General Motors eventually learned at Hamtramck that large investments in robotics were no substitute for the hard work of redesigning jobs, improving incentives, and retraining the work force. The misguided attempt to sidestep those processes resulted in financial losses and alienated customers -- significant costs, but ones that caused no permanent damage.

In schools, however, innovations that are bad ideas can have larger consequences: children who pass through the schools without learning critical skills needed to prosper in a difficult, increasingly demanding labor market. The high cost of school failure poses a difficult dilemma for school officials charged with monitoring choice, school-based management, and charter school initiatives: when to persevere on a new initiative and when to give up or to change course. The question is difficult because, without perseverance, no initiative that has potential for real improvement will succeed. At the same time, some ideas are hopelessly flawed, and it is important to end these initiatives before children experience serious losses in achievement. Two tests provide guidance about when to persevere and when to end an initiative.

The first test. Within a school, any plan worthy of long-term support must ultimately embrace all of the first four principles. Thus the first test of a program is to determine whether it is trying to establish consensus among teachers, students, and parents with regard to goals; to provide all parties with appropriate training and incentives; and to develop measures of progress. While the test sounds simple, it is one that many programs fail. It is all too common to hear advocates argue for such reforms as school-based management or school choice as ends in themselves. Forcing advocates to explain how a reform will improve student achievement is a powerful aid to clear thinking.

The second test. Even with a sensible plan, it may be several years before it is known whether or not an innovation has improved student achievement. During this time, many people will ask whether the plan is on course or whether dramatic midcourse corrections are needed. Under these circumstances, the second test should be applied. Ask whether the plan is changing the daily activities of teachers and students. Unless the daily work of teachers and students changes, student achievement cannot improve.17

As a thought experiment, consider applying the tests to a new charter school. Does the school have clear goals? Are there professional development opportunities for teachers that are tied to the pursuit of clearly defined goals? Are there incentives for teachers to participate in professional development and to change the way they teach? Is there an agreed-upon strategy for measuring progress in student achievement? Does the classroom work here differ from the work in other schools? Only charter schools that provide affirmative answers to these questions will contribute to the goal of providing all children with the skills needed to thrive in today's demanding economy.

Schools for Tomorrow

There is every likelihood that the economy will continue to change rapidly. The good jobs -- the ones that pay enough to support a family -- continue to become more complex and to require employees with greater levels of skill. And so economic change hits hardest at the least educated. As long as our education remains unchanged, these trends will not reverse.

But as the newspaper tells you, economic change now affects more than high school graduates. White-collar jobs at General Electric and Citibank can end in a moment. People with more than a high school diploma have to resell themselves four and five times during their careers. As long-term commitments fade, strong education and skills become important for these workers too.

Together, these facts make up one-half of the future. They say that people must see themselves as economic free agents, prepared to prove their market worth at any time. It is a world where you go to war every day, and, short of being a millionaire, a very good education is your best armor.

The other half of the future will be created by our response to these facts. Many of today's schools continue to educate children for an economy that no longer exists, and many of today's parents are just beginning to recognize the problem. Improving our schools requires a long-term commitment to change based on the Five Principles: setting clear goals, improving incentives and redesigning jobs, continually retraining the teaching force, monitoring progress on a regular basis, and persevering in the face of adversity. The experiences of schools and private sector firms that have successfully improved show that this strategy works.

The promise of well-designed choice plans, charter school programs, and school-based management initiatives is that they can signal the need for change and can make it easier to bring about change. The danger of these initiatives is that they will be viewed as magic bullets -- as the educational equivalent of robotics at Hamtramck. This danger is particularly great this election year, when candidates for public office seek to attract voters with sound bites. Choice and charter schools lend themselves to sound bites. The details -- the differences between Proposition 174 and the Coons/Sugarman voucher plan, for example -- do not. Moreover, the central message, that market-based initiatives will contribute to better schooling only if they stimulate change based on the Five Principles, is too complicated for electoral speeches. The crucial question now is whether America is ready to hear the hard truth, taught by GM's experience, that there is no substitute for the difficult work of restructuring.

1. The description of the experience at Hamtramck is based on Maryann Keller, Rude Awakening: The Rise, Fall, and Struggle for Recovery of General Motors (New York: William Morrow, 1989).
2. See Ina V. S. Mullis et al., Report in Brief: NAEP 1992 Trends in Academic Progress (Washington, D.C.: National Center for Education Statistics, July 1994), pp. 12-13.
3. John E. Chubb and Terry M. Moe, "Choice Is a Panacea," Brookings Review, Summer 1990, pp. 4-12.
4. Stanley M. Elam, Lowell C. Rose, and Alec M. Gallup, "The 28th Annual Phi Delta Kappa/Gallup Poll of the Public's Attitudes Toward the Public Schools," Phi Delta Kappan, September 1996, p. 46.
5. Mullis et al., pp. 12-14.
6. For a detailed discussion of critical skills, see Richard J. Murnane and Frank Levy, Teaching the New Basic Skills: Principles for Educating Children to Thrive in a Changing Economy (New York: Free Press, 1996).
7. John F. Witte, Andrea B. Bailey, and Christopher A. Thom, Third-Year Report: Milwaukee Parental Choice Program (Madison: Department of Political Science and Robert La Follette Institute of Public Affairs, University of Wisconsin, December 1993); and John F. Witte et al., Fourth-Year Report: Milwaukee Parental Choice Program (Madison: Department of Political Science and Robert La Follette Institute of Public Affairs, University of Wisconsin, December 1994). As currently operated, the program is limited to 1.5% of students in the Milwaukee public schools.
8. Daniel McGroarty, "School Choice Slandered," The Public Interest, Fall 1994, pp. 94-111.
9. Witte, Bailey, and Thom, pp. 22-23.
10. Anthony S. Bryk, Valerie E. Lee, and Peter B. Holland, Catholic Schools and the Common Good (Cambridge, Mass.: Harvard University Press, 1993).
11. The most recent evaluation of the Milwaukee choice program describes outcomes in the fourth year of the program.
12. The descriptions of the Coons and Sugarman voucher initiative and Proposition 174 come from the following sources: John E. Coons and Stephen D. Sugarman, Scholarships for Children (Berkeley, Calif.: Institute of Governmental Studies Press, 1992); and "Education. Vouchers. Initiative Constitutional Amendment," from California ballot pamphlet, analysis by the legislative analyst (Sacramento: Secretary of State, State of California, November 1993), pp. 32-35.
13. For a thoughtful discussion of the conditions under which choice may promote improvements in an organization's performance, see Albert O. Hirschman, Exit, Voice, and Loyalty (Cambridge, Mass.: Harvard University Press, 1970).
14. Joan First et al., The Good Common School (Boston: National Coalition of Advocates for Students, 1991), pp. 29-31.
15. See Anita A. Summers and Amy W. Johnson, "A Review of the Evidence on the Effects of School-Based Management Plans," paper presented at the Conference on Improving the Performance of America's Schools: Economic Choices, National Academy of Sciences, Washington, D.C., 12-13 October 1994; and Bruce Bimber, The Decentralization Mirage: Comparing Decision-Making Arrangements in Four High Schools (Santa Monica, Calif.: RAND Corporation, 1994).
16. Joe Nathan, "Possibilities, Problems, and Progress: Early Lessons from the Charter Movement," Phi Delta Kappan, September 1996, pp. 15-23.
17. For a rich description of this perspective, see Edward Pauly, The Classroom Crucible: What Really Works, What Doesn't, and Why (New York: Basic Books, 1991).

PDK Home | Contact Us | Search

Last modified January 14, 2008
http://www.pdkintl.org/kappan/k_v78/k9610mur.htm

This article appeared in the October 1996 issue of Phi Delta Kappan, Vol. 78, No. 2.
PDK International respects your privacy
© 2007 Phi Delta Kappa International