According to a theoretical model, school autonomy and parental participation in schools, can increase student learning through separate channels. Greater school autonomy increases the rent that can be distributed among stakeholders in the school, while institutions for parental participation (such as school board) empower parents to command a larger share of this surplus - for example, through student learning. Using a rich cross-sectional data set from Argentine schools (sixth and seventh grades), the authors find that autonomy, and participation raise student test scores for a given level of inputs, in a multiplicative way, consistent with the model. Autonomy has a direct effect on learning (but not for very low levels of participation), while participation affects learning only through the mediation of the effect of autonomy. The results are robust to a variety of robustness checks, and for sub-samples of children from poor households, children of uneducated mothers, schools with low mean family wealth, and public schools. It is possible that autonomy, and participation are endogenously determined, and that this biases the results - the data available do not allow this to be ruled out with certainty. Plausible predicators of autonomy, and participation are also plausible predicators of test scores, and they fail tests for the over-identifying restrictions. Heuristically argued, however, the potential for correlation with unobserved variables may be limited: the data set is rich in observed variables, and autonomy and participation show very low correlation with observed variables. Subject to these caveats, the results may be relevant to decentralization in two ways. First, as decentralization moves responsibility from the central, toward the provincial or local government, the results should be directly relevant if the decentralization increases autonomy, and participation in schools. Second, if the results are interpreted as representing a more general effect of moving decision-making toward users, and the local community, the results are relevant even if little happens to autonomy, and participation in schools. More important, perhaps, the authors illustrate empirically the importance of knowing who is empowered when higher levels of government loosen control.
