Paper latest version: https://conference.druid.dk/ac_papers/d8gctpyh2tdp6tew282el2vo57mwmv.pdf
In this paper, we try to answer the question: "Has a scientific background helped leaders against COVID-19?". To do so, we use a regression discontinuity design in close mayoral races between STEM (science, technology, engineering, and math) and non-STEM candidates to estimate the causal impact of having a mayor with scientific background on COVID-19 epidemiological outcomes in Brazil. We find that STEM mayors reduced COVID-19 deaths and hospitalizations and one of their mechanisms was increasing the number of non-pharmaceutical interventions (NPIs), such as face mask usage. We also show that this estimated impact is not due to other observable mayoral characteristics, such as years of education, ideology, or gender. Our findings suggest that municipal leaders with STEM training better handled a scenario involving emergencies and small amounts of data and evidence. We argue that our results may indicate that investment in science and technology-related human capital produces non-expected externalities—such as improvements in public management issues.