Kai Sun, Daniel J. Henderson, and Subal C. Kumbhakar, "Biases in 
Approximating Log Production", Journal of Applied Econometrics, Vol. 26, 
No. 4, 2011, pp. 708-714.

There is one data file (shk_data.txt) and one (R language) program 
file (shk_code.txt) which can be used to replicate the results 
reported in the paper. Both of these are ASCII files in DOS format. They are
zipped in the file shk-files.zip. Unix/Linux users should use "unzip -a".

Data description:

All data are taken directly from Masanjala and Papageorgiou, "The Solow
Model with CES Technology: Nonlinearities and Parameter Heterogeneity",
Journal of Applied Econometrics, Vol. 19, 2004, pp. 171-201.

The data file is named shk_data.txt.  There are 98 observations for each of 
the following variables:

GDP85 = Per capita GDP in 1985.
lnGDP85 = natural logarithm of GDP85 
s_ik = Average ratio of investment to GDP (1960-1985). 
s_ih = Average fraction of working-age population enrolled in secondary 
school (1960-1985).
ngdelta = n(population growth rate)+g(technology growth rate)
+delta(depreciation rate)
lns_ikonngdelta = ln (s_ik / ngdelta)
lns_ihonngdelta = ln (s_ih / ngdelta)
taylor1 = (lns_ikonngdelta)^2
taylor2 = (lns_ihonngdelta)^2
taylor3 = the square of the natural log of s_ik/s_ih

taylor1, taylor 2, taylor 3 are from the Taylor expansion 
(see pp.175 of Masanjala and Papageorgiou, 2004)

Note that this dataset is a subset of that in Masanjala and Papageorgiou 
(2004). More detailed information, as well as the full data set, can be
found at:

http://www.econ.queensu.ca/jae/2004-v19.2/masanjala-papageorgiou/


Code description

The file shk_code.txt can be used to replicate the results in the 
paper. It first estimates the basic CD and CES models, then their 
extended counterparts. The software R needs to be installed:

http://www.r-project.org/

The code also requires four packages built under the software (car, sandwich, 
msm and lmtest).