Smoothed instrumental variables quantile regression

In this article, I introduce the sivqr command, which estimates the coefficients of the instrumental variables quantile regression model introduced by Chernozhukov and Hansen (2005, Econometrica 73: 245–261). The sivqr com- mand offers several advantages over the existing ivqreg and ivqreg2 commands for estimating this instrumental variables quantile regression model, which com- plements the alternative “triangular model” behind cqiv and the “local quan- tile treatment effect” model of ivqte. Computationally, sivqr implements the smoothed estimator of Kaplan and Sun (2017, Econometric Theory 33: 105–157), who show that smoothing improves both computation time and statistical accu- racy. Standard errors are computed analytically or by Bayesian bootstrap; for nonindependent and identically distributed sampling, sivqr is compatible with bootstrap. I discuss syntax and the underlying methodology, and I compare sivqr with other commands in an example.