Bjung-Joo Lee, "Separability Test for the Electricity Supply Industry",
Journal of Applied Econometrics, Vol. 10, No. 1, 1995, pp. 49-60. 

There are two data sets for this paper.  All variables in both data sets
are arranged in a free format style, separated by blanks and/or tabs.
For some computer systems, it may be necessary to expand the tabs.  They
are ordered by observation, and there are 70 observations for each
variable.

The first data set is "lee1.data".  It contains the capital stocks and
the prices of seven different types of asset classes.  They are:  steam,
nuclear, hydroelectric, other power generating equipment, transmission,
distribution and general plant equipment.  There are 14 variables
arranged in the seven capital stocks first, and then seven prices.

The second data set, "lee2.data", has other than capital data.  There
are 11 variables in this file.  The first three variables are the
elasticities for different types of demand, commercial, industrial, and
residential.  Next three variables are labor inputs:  number of full
time, part time and the labor required for the generation stage.  The
next are the fuel price and the fuel input.  The last three variables
are the amount of output in three different demand categories:
commercial, industrial and residential.