Many statistical estimation problems that arise in economic applications and more generally in causal inference, require the estimation of nuisance quantities that are not the focus of the analyst but are simply aides to identify causal models from observational data. Examples include modeling the treatment policy when estimating treatment effects, modeling the pricing policy when estimating demand elasticity or modeling an opponents behavior when estimating the structural utility parameters in a game of incomplete information. I will give an overview of recent results that invoke and extend the principle of Neyman-orthogonality to develop estimation methods whose target parameter error is robust to errors in the estimation of the nuisance components of the model. This enables fast mean squared error rates and asymptotically valid confidence intervals, even when the nuisance components are fitted via arbitrary ML-based approaches. I will discuss applications in the estimation of heterogeneous treatment effects, estimation with missing data, estimation in games of incomplete information and offline policy learning.
Back to the workshop page