Association for Public Policy Analysis & Management

Countries differ on the arrangements between different levels of government to manage their education systems, and particularly to regulate and manage schools at the elementary and secondary levels. Some countries place most of the responsibilities on school boards at the district level, others grant high levels of autonomy to schools, some others centralize most of the responsibilities at the national level, and finally, in other countries, a great deal of responsibilities are placed at the municipal or regional level. We explore if giving autonomy to municipalities or regional bodies has an effect on students’ performance in a comparative perspective. Do students in countries where municipalities or regional bodies have more responsibilities regarding the school system have better academic results when compared to students from countries where policies are either centralized at the national level, or where schools have higher levels of autonomy?

To answer the previous question in a comparative perspective, we will use information from the Organization for Economic Cooperation and Development (OECD)’s Program for International Student Assessment (PISA) for years 2003, 2006, 2009 and 2013. In addition to the test on mathematics and language, each PISA’s iteration includes context questionnaires answered by students, parents and school principals.  The context questionnaire includes two types of questions of interest for this research. First, questions regarding the source of funding for the schools and the percentages coming from different sources, including regional or local authorities. Second, questions about the authority or body responsible for a variety of tasks regarding the education system, including teachers’ hiring and firing, establishing salaries and their increases, formulating schools’ budgets and defining budget allocations, among others.  

Our model will estimate the effect of the municipal or regional autonomy on students’ performance, measured by PISA results. The identification strategy relies on a panel of about 40 countries that have participated in at least three of the tests’ iterations. Using panel data with country fixed effects allows us to exploit within country variations on levels of municipal or regional autonomy and estimate their effect on students’ achievement, while controlling for unobservable factors that are specific to that country and could influence both the levels of administrative autonomy and students’ achievement (e.g. cultural preferences towards a more centralized regime, or changes about the importance placed on education by public opinion). Additionally, the possibility of using time fixed effects allows controlling for any factors that could vary over time but are constant across states (e.g. increased awareness about performance in PISA). Furthermore, the richness of PISA’s context questionnaires and the availability of international data at the country level permit including students, schools and countries’ covariates in order to increase the precision of our estimates. Finally, the comparison of levels of municipal or regional autonomy aggregated at the country level avoids any issues of “selection” of students into municipalities or regions with different levels of autonomy (i.e. the issue that students with high performance self-select themselves into a particular type of municipality or region).