2023/07/27 by Jad Beyhum, Beyhum, Jad, Elia Lapenta +3
Computer Science · Mathematics · #Neural Networks and Applications #Advanced Statistical Methods and Models #Gaussian Processes and Bayesian Inference
paper · pdf · doi:10.48550/arxiv.2307.14867
We extend nonparametric regression smoothing splines to a context where there is endogeneity and instrumental variables are available. Unlike popular existing estimators, the resulting estimator is one-step and relies on a unique regularization parameter. We derive rates of the convergence for the estimator and its first derivative, which are uniform in the support of the endogenous variable. We also address the issue of imposing monotonicity in estimation and extend the approach to a partly linear model. Simulations confirm the good performances of our estimator compared to two-step procedures. Our method yields economically sensible results when used to estimate Engel curves.