2026 Theses Doctoral

Essays on Disability, Inequality, and Causal Inference in Education and Labor Economics

Caron, Laura

In this dissertation, I examine research questions related to inequalities in labor and education economics, particularly for disabled individuals. This work highlights the importance of education policy and access to education for disabled students, along with the challenges of implementing and assessing the impacts of such policies. Throughout, this work depends on methods for causal inference in dynamic, staggered difference-in-difference designs, where heterogeneous treatment effects are likely. This setup is of particular interest in the last chapter, which extends prior econometric work on triple difference-in-difference designs to this setting. This thesis contributes new insights on previously understudied topics relating to how economic policy around disability impacts the education and labor market outcomes of disabled people and their peers.

In chapter 1, I study the impacts of a major expansion in public education services for disabled students in the US. Between 1949 and 1980, every U.S. state mandated public schools to provide educational services for disabled students. This is one of the largest education reforms in U.S. history, but little is known about its impacts. Given scarce data in this period, I compile survey and administrative datasets and set up a difference-in-difference design using variation in the mandates' timing. I show that the mandates increased both services for disabled students and preschool enrollments. In adulthood, disabled individuals below school age at a mandate's implementation became about 20% less likely to have no education, attained up to 0.23 more years of education, and were more likely to have worked. Although this policy could have taken away resources from non-disabled students, in fact, education and employment also increased for non-disabled individuals. These effects align with evidence that the mandates increased spending per student by up to 15%. Families were also impacted: the mandates increased employment among mothers of disabled children and the probability that disabled individuals became household heads. Over the long term, the mandates paid for themselves by generating government revenues in excess of their cost. These results provide new evidence on the large, broad impacts of expanding access to education for disabled students.

In chapter 2, coauthored with Jeppe Johansen and Jesper Eriksen, we highlight the importance of understanding student mobility when trying to assess the peer impacts of placing disabled students in classrooms with their peers. Prior work has shown that placing disabled students in classrooms with non-disabled peers ("mainstreaming") can have negative impacts on these peers, but has not fully considered the role of student mobility in understanding these impacts. Using Danish administrative data and a difference-in-difference design, we show that the arrival of a mainstreamed student in a classroom makes peers 11% more likely to switch schools in a given year. Because school switching is selective by socio-economic background and choice of public versus private schools, this may increase socio-economic segregation. We isolate the direct impacts of the arrival of a mainstreamed student from the indirect impacts of school switching by studying a sizeable --- although potentially selected --- subgroup of students who never change schools. For this subgroup, we find little evidence that the arrival of a mainstreamed student causes negative effects on attendance, test scores, or social tensions in the class. In contrast, students who do change schools experience negative impacts, which may be due to direct effects from the arrival of a mainstreamed student or the indirect disruption caused by switching schools. Research and policy on the inclusion of disabled students in classrooms must account for the role of student mobility in shaping outcomes.

In chapter 3, I address causal inference in a popular research design, building on previous econometric literature cited and applied in chapters 1 and 2. Triple difference designs have become increasingly popular in empirical economics. The advantage of a triple difference design is that, within a treatment group, it allows for another subgroup of the population -- potentially less impacted by the treatment -- to serve as a control for the subgroup of interest. While literature on difference-in-differences has discussed heterogeneity in treatment effects between treated and control groups or over time, little attention has been given to the implications of heterogeneity in treatment effects between subgroups.

In this paper, I show that the parameter identified under the usual triple difference assumptions does not allow for causal interpretation of differences between subgroups when subgroups may differ in their underlying (unobserved) treatment effects. I propose a new parameter of interest, the causal difference in average treatment effects on the treated, which makes causal comparisons between subgroups. I discuss assumptions for identification and derive the semiparametric efficiency bounds for this parameter. I then propose doubly-robust, efficient estimators for this parameter. I use a simulation study to highlight the desirable finite-sample properties of these estimators, as well as to show the difference between this parameter and the usual triple difference parameter of interest. An empirical application shows the importance of considering treatment effect heterogeneity in practical applications.