Econometrics, Quantitative Economics, Data Science



R and Matlab implementation of Vector Quantile Regression


This software implements Carlier, Chernozhukov and Galichon’s “Vector Quantile Regression” (Annals of Statistics, forthcoming). Vector Quantile Regression (VQR) is a multivariate version of the Quantile Regression procedure of Koenker and Bassett (1978). It relies on a multivariate extension of the notion of quantile via optimal transportation, and a representation of Conditional Vector Quantiles by a variational problem.

Practical information

(c) 2014 by Carlier, Chernozhukov and Galichon. Available in Matlab and R. See user manual in the zip archive. Requires Gurobi. This software may be used for non-commercial use only and comes with absolutely no warranty.


Available here.
Engel’s multivariate dataset is available here.


Carlier, G., Chernozhukov, V., and Galichon, A. (2014): “Vector Quantile Regression”. Arxiv:1406.4643.