actors may not direct the school's instructional program well. Decentralization provides a check on the distortions likely to emerge in an incentives-only system. The major hitch is that local monitors, usually parents, must be able to detect distortions and be able to make judgments about both the seriousness of the distortions and ways to correct them. This is largely an information problem. Ultimately, for an SBM system with incentives to work most effectively, central authorities should provide information and technical assistance to school-level actors. The content of the information and the way it is provided are consequential. The information contained in ''report cards" on schools, for example, has some of the same problems that incentives do, albeit less severe ones. It can divert attention to those things easily measured.
A second type of information, though more costly to provide, is likely to be more valuable. This is information to enhance local actors' understanding of the range of production possibilities in education. It might be acquired by visiting other schools to learn what they are doing; joining a network of schools where technical advice is shared, or bringing in consultants on particular issues. This type of information would help local school actors identify possible shortcomings in their school's instructional program as well as likely solutions.
In short, performance incentive schemes in education and decentralized management arrangements each have advantages and disadvantages. To some extent, however, they balance each other out. Decentralized management systems are important because local monitoring can correct some of the distortions created by imperfect incentive systems and incentives are important because of the weaknesses in the management capacity of many local schools. Even with both decentralization and performance incentives, quality performance and process information is needed to ensure success.
As Summers and Johnson observe in their chapter in this volume, information on the contribution of SBM and other education reforms to student performance is meager. But we should not be overly critical of education research for the lack of good evaluations. A number of research problems make it a difficult area in which to obtain clear results.
One problem noted earlier is the absence of agreed-upon goals. The extent to which all schools are pursuing the same objectives is unclear. Even if a common set of agreed-upon objectives could be defined, there would still be serious measurement issues. If it were easy to identify and measure value-added in education, it would be easy to design incentive schemes to elicit the desired behavior; management issues would be of little concern. Management in education is important, but difficult to evaluate and prescribe precisely, because identifying and measuring agreed-upon value-added is so difficult.
Second, any analysis of the value of governance reforms on student achieve-