v11i1 Schmidt

This study applies advanced methods of causal inference, specifically the Double Machine Learning (DML) framework, to estimate the causal effects of public health insurance on individual health outcomes using data from the 2008 Oregon Health Insurance Experiment (OHIE), a randomized controlled trial. DML integrates modern machine learning with the econometric principles of causal identification to obtain
unbiased treatment effect estimates in high-dimensional data. Interactive Regression Models (IRM) and Interactive Instrumental Variable Models (IIVM) are employed to estimate effects of Medicaid coverage on perceived health, number of doctor visits,
satisfaction, access to medical services, and quality of care. The results indicate small but positive causal effects of Medicaid coverage on perceived health and healthcare utilization, statistically insignificant effects on satisfaction and access to medication, and a slightly negative effect on perceived quality of care. The findings highlight the potential of Double Machine Learning as a robust framework for causal analysis in empirical research.