CRoNoS & MDA 2019: Start Registration
View Submission - CRONOSMDA2019
A0272
Title: Independent and conditionally independent counterfactuals Authors:  Marcin Wolski - European Investment Bank (Luxembourg) [presenting]
Abstract: A novel dependence filtering framework is proposed. A counterfactual random variable which is independent, or conditionally independent, from the effects of given covariates is characterized. Under error exogeneity the counterfactuals have causal interpretation. A fully nonparametric estimation technique and an inference roadmap for such counterfactuals is offered. A numerical exercise confirms that the approach performs well in nonlinear environments. Furthermore, bootstrap validity results are provided for the confidence intervals of the estimates. The approach is applied to filter out the sovereign risk spill-overs on corporate cost of borrowing in selected euro area countries.