Le-Yu Chen and Sokbae Lee, "Exact Computation of GMM Estimators for 
Instrumental Variable Quantile Regression Models", Journal of Applied 
Econometrics, Vol. 33, No. 4, 2018, pp. 553-567.

All programs are zipped in cl-programs.zip. These are Matlab programs   
used to reproduce the numerical results of the paper. The zipped file 
also encloses a data file "fish.mat", which consists of the fish demand 
dataset used for the empirical application. This dataset is stored in the
Matlab data matrix format.

The original file for the fish demand dataset is available at
http://people.brandeis.edu/~kgraddy/datasets/fish.out

To replicate the simulation and empirical results of the paper, put
all program and data files in the same work directory. The Matlab
version of the Gurobi solver has to be installed for running the codes
for solving the MIO based IVQR GMM estimation problem. The Gurobi
solver is freely available for academic purposes.

Here is the description of the program files.

IVQR_GMM.m implements the Matlab function for computing the exact GMM
estimator and its standard error for the IVQR model via the MIQP
formulation (3.3).

IVQR_MIO.m implements the Matlab function for solving the MIO
formulations (3.3), (C.1) and (C.10).

miobnd_fn.m implements the Matlab function for computing the bounding
quantities M(i) defined by (3.6) of the paper.

Two_stage_LS.m implements the Matlab function for computing the
coefficient estimates and the estimated asymptotic variance for the
two-stage least square (2SLS) regression of y on x using z as
instruments.

inv_qr.m implements the Matlab function for computing the inverse QR
estimator of Chernozhukov and Hansen (2006).

rq.m implements the Matlab function for solving the standard quantile
regression problem. This implementation is based on the codes provided
by Christian Hansen via the link:
http://faculty.chicagobooth.edu/christian.hansen/research/iqrmat.zip

simulation_3endog_MIO.m contains the codes for generating the
simulation data used in all the simulation studies of the paper on the
performance of the MIO based IVQR GMM estimation approach.

asym_std_dev_computation.m contains the codes for computing the
asymptotic standard error reported in Table 4 of the paper.

simulation_coverage.m contains the codes for computing the coverage
probabilities reported in Table 5 of the paper. 

simulation_3endog_inv_qr.m contains the codes for reproducing the
simulation results reported in Tables 7 and 8 of the paper on the
performance of the inverse QR estimation approach.  

demand_estimation.m contains the codes for reproducing the empirical
results reported in Table 6 of the paper.