Microeconometric Analyses of Individual Behavior in Public Welfare Systems : Applications in Health and Education Economics

Sammanfattning: This thesis consists of four self-contained essays.Essay 1: Women have more absenteeism while simultaneously live longer than men. This pattern suggests that men and women's preferences for sickness absence might differ due to e.g. distinct health behaviors. These behaviors could, in turn, arise from the traditional gender division of labor within households, in which it might be more important to invest in the woman's health. We empirically analyze these hypotheses using administrative health data and find robust evidence for gender differences in preferences for health-related absenteeism.Essay 2: The paper analyzes whether residential proximity from an emergency room affects health outcomes from suffering an acute myocardial infarction (AMI). Previous research has suffered from empirical problems relating to health-based spatial sorting of involved agents and data limitations on out-of-hospital mortality. Using policy-induced variation in hospital distance, arising from emergency room closures, and data on all AMI deaths in Sweden over a twenty-year period, results show a clear and gradually declining probability of surviving an AMI as hospital distance increases.Essay 3: Although learning-by-doing is believed to be an important source of productivity, there is limited evidence that increased production volume enhances productivity. We document evidence of learning-by-doing in a high-skill activity where stakes are high; advanced cancer surgery. For this purpose, we introduce a novel instrument that exploits changes in the number of public hospitals across time and space, affecting the number of cancer surgeries performed in Swedish hospitals. Using detailed register data, our results suggest substantial positive effects of operation volume on post-surgery survival rates.Essay 4: The paper analyzes whether student choice of college financing affects study durations by exploiting an intervention in the Swedish student aid system. The reform provided incentives for college students to reallocate time from studies to market work. We evaluate this time reallocation hypothesis by estimating relative changes in earnings and completed academic credits attributed to the intervention for students from different socioeconomic backgrounds. Applying detailed Swedish administrative data, we find that the intervention both increased relative earnings and decreased the relative study pace for students from a lower socioeconomic background.