Data and Code for: Labor Market Inequality And The Changing Life Cycle Profile Of Male And Female Wages

The code in this replication package assembles the data needed and replicates the analysis of the paper “Labor Market Inequality and the Changing Life Cycle Profile of Male and Female Wages,” by Richard Blundell, Hugo Lopez, and James P. Ziliak. The first file is a Stata file 0_InstallPackages.do which installs a number of plug-in ADO files needed for successful execution (only run once). The second is also a Stata file 1_FullDataPrep.do which calls a number of Stata DO files to compile all the necessary data files and prepares the data for the analysis. The resulting Stata dataset, RunningData_withtaxsim.dta, is found in the replication file /ProcessedData/, and because Matlab relies on csv files, the resulting Matlab input files are at /ProcessedData/MatlabDataInputFiles/. The replicator should expect the code to run for about 3 hours. Then, the Matlab file a2_QuantileEstimation.m should be executed. The parameter estimates reported in the figures and tables come from a Windows desktop version that takes about 5 hours for each of four subsamples for each model specification. Due to the computational complexity of the bootstrap quantile with selection estimator, we made use of a computing cluster with a SLURM job scheduler. There are 8 Matlab bootstrap programs—four to produce standard errors in Tables 1-4 of the manuscript and four to produce standard errors for Supplemental Appendix Tables D1-D4. These boostrap computations were submitted in parallel (i.e. all at once as separate programs) and each took on average 6 days when running on 4 cores. Then the Stata file 3_Figures&amp;Tables.do should be run to produce all 9 figures (30 in the appendix) and 4 tables (7 in the appendix).<br>