Badi Baltagi and Pallab Kumar Ghosh, "Replication of 'Unconditional
Quantile Regressions' by Firpo, Fortin and Lemieux (2009)", Journal of
Applied Econometrics, Vol. 32, No. 1, 2017, pp. 218-223.

Two separate data sources are used in this article. First are the data
used in

  Sergio Firpo, Nicole M. Fortin, and Thomas Lemieux, "Unconditional
  Quantile Regressions", Econometrica, Vol. 77, No. 3 (May, 2009), pp.
  953-973

to replicate the authors' results.

We have applied the proposed unconditional quantile regression
estimation method using a separate data set used in the paper

  David H. Autor, Lawrence F. Katz, and Melissa S. Kearney, "Trends in
  U.S. Wage Inequality: Revising the Revisionists", Review of
  Economics and Statistics, Vol. 90, No. 2 (2008), pp: 300-323
  
to study the recent trend of wage inequality across different income
groups in the U.S. labor market.


1. Firpo, Fortin, Lemieux

The data file and the original programs in STATA used to obtain the
results in the paper may be found at

https://www.econometricsociety.org/publications/econometrica/2009/05/01/unconditional-quantile-regressions

Additional programs to compute more general RIF-regressions are
available on the web site: http://www.econ.ubc.ca/nfortin/

The data file is men8385.dta, a STATA10 data file containing an
extract of variables from the Merged Outgoing Rotation Group of the
Current Population Survey of 1983, 1984 and 1985. The file contains
266956 observations on males with 17 variables whose definition is
given by the variable labels. More detail about the data selection and
recoding (e.g. top coding, wage deflator, etc.) is found in Lemieux
(2006).

The file men8385.dta is zipped in ffl-dta.zip. The same data may be
found in the ASCII text file men8385.txt, which is zipped in
ffl-txt.zip. Since men8385.txt is in DOS format, Unix/Linux users
should use "unzip -a".

The files men8385.dta and men8385.txt contain 266,956 observations on
males and have 17 variables. The dependent variable is the logarithm
of hourly wage from the current population survey outgoing rotation
samples for 1983, 1984 and 1985. The independent variables are
workers' demographic characteristics, which consist of five education
categories (dropout, high-school graduates, some college, college
graduates, post college graduates), nine potential experience
categories (each of five years gap), and dummy variables for union,
married, and nonwhite.


2. Autor, Katz, Kearney

The data sets we used are available on Professor David Autor's MIT web
page. We used the following three years of data from Autor et al.
(2008) Table 4 data1 to generate the results in our Table 2. There are
three Stata files:

1.   quant_morg1973_w.dta

2.   quant_morg1989_w.dta

3.   quant_morg2005_w.dta

All the three years data are from the March current population survey
for full time workers ages in between 16 to 64. Both 1973 and 1989
years data have 98 variables and 2005 data has 97 variables. The
number of observations in 1973, 1989 and 2005 years data are 32,944;
128,147; and 104,250 respectively.

We have divided each year's data into two parts based on gender for
our specific empirical application. Thus we have the following six
data sets, with the specified numbers of observations:

  (i)   Year1973 Female  n=13,442
  (ii)  Year1973 Male    n=19,502
  (iii) Year1989 Female  n=61,618
  (iv)  Year1989 Male    n=66,529
  (v)   Year2005 Female  n=51,465
  (vi)  Year2005 Male    n=52,785
  
The six .dta files are zipped in akk-dta.zip, and the six
corresponding ASCII files, in DOS format, are zipped in akk-txt.zip.

The dependent variable is the logarithm of weekly wage for full time
workers ages 16 to 64 from the March current population survey. The
data consists of five education categories (Dropout, High-school
graduates, some college, college graduates, post college graduates),
four potential experience categories (0-9, 10-19, 20-29, and 30-39
years), square, cubic and fourth degree of experience, black and other
race dummies, and interactions of experience with the education
category variables.

For more detail on data preparation and variable construction,
see the Autor et al. (2008) Data Appendix.

Please address any questions to:

Pallab Ghosh
Dept. of Economics
University of Oklahoma
Norman, OK 73071
USA