One of the great public policy controversies of recent years centers on the effectiveness of funding for education. The dimensions of this controversy tend to be confused, however. The most common ways of discussing school spending are not very helpful for thinking about school policy; they simply divert attention from the real issues. Those issues currently have little to do with resources and much more to do with educational organization, incentives, and decision making.
Some six years ago, a panel of economists convened to consider a set of questions that is critical to the health and well-being of the American people: How important are schools in bettering our society? How have they done? What—if anything—should be done to improve how they function? The answers to these questions are clear: very; poor; and a lot. Following is an overview of the evidence and conclusions from that panel's deliberations, Making School's Work (Hanushek et al. 1994).
Some Baseline Facts
Economists view schooling as an investment. The time and money spent on an individual's education are rewarded by better jobs, higher earnings, improved health, enhanced parenting skills, and so forth. Society as a whole benefits from the greater productivity of its workers, more democratic participation, a lower crime rate, and a higher rate of economic growth.
But all investments are not equally good. During the past quarter century, we have shifted our attention away from increasing graduation and college attendance rates and toward improving the quality of education for all students and ensuring that schooling is provided equitably. Although we have dramatically increased our investments, they have not yielded the high returns that previous efforts did. That is the heart of today's problem, and the issue we must address if America is to reach its potential.
During the past 25 years, per-pupil expenditure has more than doubled in real terms (that is, after allowing for inflation). This increase is the result of dramatically smaller pupil-teacher ratios, the growing numbers of more experienced teachers and teachers with master's degrees, and the higher real salaries teachers receive. For example, the teacher-pupil ratio fell from 1 teacher to 26 students in 1960 to 1 teacher to 17 students in 1990. During the same period, the percentage of teachers with a master's degree more than doubled, from 23 percent to 56 percent.
The results have been disappointing. The performance of 17-year-olds, as measured by the National Assessment of Educational Progress, shows that average reading achievement is essentially the same in the 1990s as it was in the early 1970s; science achievement has fallen; and math achievement is only slightly improved. SAT scores—even after adjusting for selective test-taking—are noticeably below those in the 1960s and early 1970s. Moreover, although the performance gaps between racial and ethnic groups and white students have narrowed somewhat, the differences remain unacceptably large. Finally, students in other nations consistently outperformed U.S. students. These performance differences are large; they cannot be dismissed as the result of more selective test-taking by foreign students.
These aggregate statistics tell a simple story: Enormous increases in the resources devoted to U.S. schools have not yielded improvements in student performance. Student performance has at best been flat and at worst may have fallen substantially.
These aggregate data could be misleading, and schools should not bear all of the blame. After all, costs have been pushed up by such factors as special education programs and an increasingly diverse student population. In addition, living conditions for many children have worsened with increased divorce and poverty rates and the greater number of single-parent families. Unfortunately, these factors do not sufficiently explain the trends. Special education alone cannot account for more than one-fifth of the spending increase that has occurred since passage of the Education for All Handicapped Children Act in 1975 (Hanushek and Rivkin 1994). The changes in families also have not been all bad. Parents are more educated and families are smaller, so the net impact on children actually may be favorable (Grissmer et al. 1994).
More important evidence comes from the substantial national research base. The relevant research employs sophisticated statistical analysis to separate the impact of schools and teachers from other influences on student performance. Intensive investigation during a 30-year period has provided clear evidence of the ineffectiveness of common resource allocation policies. The conclusion is clear: There is no consistent, systematic relationship between school resources and student performance.
This is most easily seen in the case of class size. Perhaps no topic has been more extensively researched than the effect of class size on student performance. Only a few of some 300 econometric investigations suggest that student achievement would improve with smaller classes, and an almost equal number confidently point in exactly the opposite direction (Hanushek 1986, 1989). Most studies give no reason to expect a change in student performance as a result of smaller class sizes. Moreover, direct experimental evidence from Tennessee reinforces these results by demonstrating that small classes at best make a discernible difference in kindergarten but not at later primary grades (Word et al. 1990).
Similar studies relating to the effects of master's degrees on student performance have shown that graduate work does not result in superior performance. Teachers with master's degrees are, nonetheless, paid more.
None of these findings would be all that important were it not for the fact that smaller class size and more graduate training are viewed as powerful policy instruments. We have pursued these, along with similarly ineffective spending policies, and we have not gotten any return on our investments.
Qualifications and Elaborations
It is important to understand how to interpret the findings before we pursue any policy implications. People frequently distort the conclusions, either because of genuine misunderstanding or because of unwillingness to accept the implications.
First, the evidence does not indicate that all schools and teachers are the same. Teachers and schools differ dramatically. Unfortunately, the best teachers are not necessarily the ones who are paid the most, who teach the smallest classes, or who have master's degrees. The issue is how we ensure that the best teachers stay in the classroom while we encourage the worst to move on.
Second, the evidence does not indicate that schools never use financial resources effectively. The most widely circulated critique of this evidence (Hedges et al. 1994) inappropriately concentrates entirely on that issue. The findings indicate that good uses of resources in one place tend to be offset by poor or ineffective uses elsewhere. Just knowing that a few schools in specific circumstances use resources effectively has very limited policy value, particularly when there is no understanding of the circumstances leading to effective use of resources (Hanushek 1994).
Third, the evidence does not indicate that money could not matter. In fact, the major focus of Making Schools Work (Hanushek et al. 1994) is on ways to improve the chances that resources will be used effectively. The results here indicate that money is not used well within the current organization of schools or in the way schools use incentives. Change those and money could become a potent policy instrument.
Fourth, the evidence does not say that all programs fail, but it does say that what is “known” is often incorrect. A common appeal for added resources goes something like this: It is known that program XYZ works; program XYZ requires additional resources to implement; and student performance would increase if we provided funds that would permit installing program XYZ. If program XYZ is so uniformly good, why isn't it already found in more schools? Surely there are ineffective programs that could be eliminated to provide resources for XYZ. Either XYZ is not easily transferred to other places, XYZ does not work as well as advertised, or something other than resources prevents its adoption.
Fifth, parents and the environment as a whole play a very important role in determining the performance of students, but simply saying that does not obviate the need to organize schools to be as effective as possible. Nonschool factors cannot explain inefficient use of resources in schools, and they will not eliminate the gains from using resources more effectively.
Sixth, faced with the evidence that resources are not used effectively, some people argue that added resources must be necessary even if they are not sufficient to ensure improved performance. They usually maintain that more resources should be allocated to schools that are doing badly. This argument is true only if it is necessary to accept current inefficient uses of resources. The very meaning of inefficiency is that more resources are being used than are needed to achieve current performance. A corollary is that eliminating the inefficiency would release resources that could be used for improvements.
Many resist the notion that our schools are inefficient. The word inefficiency implies less urgency in meeting requests for added resources. It also suggests that responsibility for the performance problems of the schools falls in part on school personnel. Finally, it means that policy making becomes more difficult, as considerably more is required than simply adding resources.
Incentives and Organization
Noting that schools are not working well today does not answer the question of how they can be improved. Despite extensive effort, researchers have been unable to describe adequately the features of a good school or a good teacher. The historical effort to make regulatory policies that specify what is to be done and how it is to be done simply does not work.
For example, each state prescribes a series of hurdles that a prospective teacher must overcome to secure certification, but none involves actual performance in the classroom. Further, little evidence links the required hurdles to effective performance. While certification perhaps ensures that truly incompetent teachers are less likely to reach the classroom, it does little to ensure that really top-notch teachers populate the schools.
The alternative approach is built on stronger performance incentives. It comes as little surprise to an economist that expenditure and performance are not linked when few if any school incentives reward high student performance. It is a sad fact that really good teachers can expect essentially the same career path, pay, and rewards as really poor teachers. Similarly, really good principals and administrators find little incentive tied to high student achievement or to efficient use of resources. Schools might be very different places if rewards were linked to student performance.
A key element of Making Schools Work is aligning incentives with performance. Teachers and administrators should be given performance goals. They should be given the latitude to exercise their professional judgment about how best to achieve such goals. And, they should be held responsible for their performance and rewarded when they achieve their goals. The underlying theme is recognition of the central role of school professionals in implementing effective programs. Such alignment of incentives and performance is, of course, easier said than done.
Policymakers have proposed and debated many generic incentive plans. Merit pay for teachers is perhaps the most recognized plan. Good teachers receive extra rewards not available to poor teachers. This is the only incentive scheme that has been tried very widely, although without much past success (Cohen and Murnane 1986). A related scheme is former President Bush's plan for “merit schools,” a plan under which entire schools are rewarded for high performance. By rewarding schools instead of individuals, it is argued, unproductive competition among teachers would be discouraged.
The private school personnel model concentrates more on decisions about who to hire and retain than on differential pay. Charter schools permit publicly sanctioned alternatives to existing public schools. These often attempt to mobilize the innovative talents of existing teachers.
Another approach concentrates on competition to supply services through performance contracting. Public school systems contract with outside suppliers to provide educational services. This approach was the subject of experimentation in the 1970s, and currently is being tried in Baltimore, Maryland, and Hartford, Connecticut.
Magnet schools, common to many desegregation plans but not restricted to such situations, also provide incentives by making demand for alternatives a mechanism for allocating resources. Finally, school choice—either restricted to public schools or including private schools—more directly links parental decisions to resource flows. Vouchers for private schools would open up much more competition.
While varying in details, each of these plans involves aligning rewards with performance. This alignment makes each plan appealing as a potential way to improve school performance. Each also involves considerable uncertainty. We've had little past experimentation. When experimentation has been attempted, little evaluation has taken place, so little has been learned.
The evaluation process in schools is another central concern of any movement toward incentives. Performance incentives work best when the goals are clearly stated, when performance is measured well, and when individuals know how to accomplish what is desired. This is hardly the state of education today. Goals are frequently ambiguous and can even be conflicting. They are seldom stated in a clear and measurable manner. The current controversies about the best way to measure student performance—too simply stated as a conflict between standardized exams and authentic testing—highlight some of the issues.
Here there is actually a twin evaluation issue. We must measure and relate student performance to incentives, but we also must compile information about performance in different kinds of incentive schemes. Each of these is a very large topic, going well beyond the scope of this article. Most discussion on this topic has focused on understanding individual student performance; there have been relatively few evaluations of alternative incentive approaches.
In brief, a key element of evaluation and incentives involves separating what the school and teachers control from what they do not. Specifically, incentive schemes must be based on value-added notions, where rewards are related to gains that can be attributed to individual teachers and schools, not just to the overall performance of students. Overall performance is important, but it involves the combined efforts of parents, peers, students, and schools. Teachers have, with good reason, resisted any evaluation schemes that fail to recognize the nature of the student population. Incentive schemes cannot ignore the importance of influences beyond the control of the teacher.
The second issue is our lack of sufficient experience with incentive approaches. We cannot gain experience from things we have never tried. This fact suggests a broad program of experimentation (with evaluation). We must try different approaches and evaluate whether or not they appear to be leading us in the direction of higher student performance.
Experimentation with random assignment of different treatments has accomplished wonders in the case of medicine, but it is virtually unheard of in education. The major experiments to date can be counted on one hand: The Office of Economic Opportunity's performance contracting experiment, Tennessee's class size experiment, and Milwaukee's voucher experiment. Now is the time to move toward experimentation—with measurement and evaluation—on a much wider basis in schools.
Implementation Issues
Getting from here to there involves major changes. Perhaps the most significant changes involve state legislatures and departments of education. These institutions must back away from being the regulators. Instead, they must work toward developing more precise performance goals, helping establish incentive schemes, and pursuing performance equity goals. States also must work closely with the federal government to learn more about what does and does not work in schools. Yes, we will face more uncertainty, but the potential gains are very large.
At the same time, local school personnel must assume greater responsibility for developing programs and designing ways to meet goals. This will almost certainly involve more decentralized decision making. Nevertheless, the character of decentralization will differ dramatically from the current approach to site-based management. Typically today, certain decisions are decentralized, but schools introduce little monitoring of student performance and few incentives for higher performance (Summers and Johnson 1994). Without assessment and incentives, decentralized decision making may actually worsen matters.
The federal government, never a major actor in school policy, should continue doing what it has been doing—supporting equity across states and districts and providing evaluation of programs and results. It should not take on more centralized policymaking.
Let the Professionals Be Professional
The public is becoming aware of the weak links between spending and student achievement. We are likely to see increasingly harsh and damaging decisions about spending for our schools unless there is clear demonstration of improved performance.
A popular slogan in education has been “Let the professionals be professional.” This slogan generally means removing the regulatory bonds on teachers and administrators and permitting them to make their own, educationally appropriate decisions. That is a portion of the central message here. This message, however, adds that professionals should also be responsible for results, and that rewards should accompany achievement.
The regulatory environment that shapes most school policy has not produced acceptable results. We have applied more and more resources without measurable student gains. This situation hurts a society that pays too much for schools and forgoes the advantages of a better educated population. It also hurts the individuals who are robbed of the schooling they deserve.
