Machine Learning Instrument Variables for Causal Inference

Instrumental variables (IVs) are a commonly used technique for causal inference from observational data. In practice, the variation induced by IVs can be limited, which yields imprecise or biased estimates of causal effects and renders the approach ineffective for policy decisions. We confront this challenge by formulating the problem of constructing instrumental variables from candidate exogenous data as a machine learning problem. We propose a novel algorithm, called MLIV (machine-learned instrumental variables), which allows learning of instruments and causal inference to be simultaneously performed from sample data. We provide formal asymptotic theory and show root-n consistency and asymptotic efficiency of our estimators hold under very general conditions. Simulations and application to real-world data demonstrate that the algorithm is highly effective and significantly improves the performance of causal inference from observational data