In evaluating policy options, say, developing new textbooks or creating more vocational schools, policymakers must deal with a series of interrelated questions: Which options are likely to be the most effective in our country? Which relate most closely to the needs and wishes of the stakeholders? What are the relative costs? How have other countries addressed similar problems? The international research community has produced a wealth of studies that can provide guidance, yet policymakers are unable to incorporate development knowledge effectively into the policymaking process. As a result, policy decisions are often based on anecdotal information and the opinions of a few experts and consultants rather than solid evidence and the corpus of available knowledge and experience. Moreover, consultation with the stakeholders themselves is often limited, and in many cases it is not clear whether the right questions are being asked. These shortcomings in the policy process are not unique to developing countries but their effects are most keenly felt there. We aim to help policymakers in developing countries overcome these process shortcomings.
Typically, policy analysis focuses on, at most, a few input variables, rather than the full array of variables affecting a given public or social system. For example, in education, much of the policy research is based on the “production function” approach, in which inputs such as school funding, physical facilities, family attributes, teacher attributes, etc. are linked to student achievement using regression analysis. While these efforts have contributed to the understanding of the factors associated with student learning, research on education production functions simply has not shown a clear, systemic relationship between resource inputs and student outcomes.
In the last two decades, a growing body of research has focused on understanding and developing causal as well as diagnostic models and their applications to policy analysis. Our approach to development policy analysis is to build on these advances, develop theoretical models, and link them to a specific body of development knowledge. We use a Bayesian networks modeling approach (described in the following sections of this note) to organize the development knowledge into expert systems that enable policymakers in developing countries to analyze and formulate policy more effectively. Bayesian networks are particularly suited for specifying causal relationships among numerous variables in a complex system to analyze their interrelationship and the performance of the entire system. This modeling technique is now used in disparate fields, such as medical diagnosis, genetics, banking, and oil production. However, it has not yet been widely used in the development context.
Using a Bayesian network model as our inference engine, we have built a knowledge-based software engine tool, the Policymakers’ Workbench. Through this tool, policymakers can run unlimited iterations of various policy scenarios and connect existing knowledge to effective action. While much of the discussion in this paper is focused on education, the framework and the tool that we discuss is general and could be employed in policy analysis in other development areas, such as health, the environment, economic growth, migration, and more.