Primary Submission Category: Heterogeneous Treatment Effects
The Heterogenous Treatment Effects of Compulsory Education Age Reforms – A Causal Forest Approach¥
Authors: Hannelore Nelissen, Krist De Witte,
Presenting Author: Hannelore Nelissen*
While existing literature often focuses on average treatment effects, there is limited understanding of how uniform policy measures, such as raising the compulsory education age, impact very specific subgroups of students. This paper addresses this gap by investigating the heterogeneous effects of a compulsory education age reform on school dropout rates. Using rich administrative microdata from Statistics Netherlands, we apply a causal forest model to estimate Conditional Average Treatment Effects (CATEs), revealing how policy impacts vary across individual and school characteristics. Our results show an average reduction of 1.06 percentage points in dropout rates attributable to the reform, with significant heterogeneity; approximately 29% of the estimated CATEs indicate statistically significant effects up to 4.6 percentage points. Vocational track students emerge as the most responsive group, with parental income, household composition, and school progress further influencing outcomes. Moreover, we find that only certain groups of at-risk students are most suitable targets for the policy reform, suggesting that a single policy may not address the needs of all students. The analysis advocates the need for complementary policies to better address diverse student needs. This study contributes to the literature by demonstrating the importance of nuanced, data-driven policy targeting to optimize educational outcomes across varied contexts.