The aggregate and 2-digit data used in Burnside, Eichenbaum and Rebelo (1995) 
"Capital Utilization and Returns to Scale" are contained in the file ma95ber.xls.  
The 3-digit data used in that paper are in the file 3digit.xls.  

The 2-digit wage and electricity data were deseasonalized using a frequency
domain technique described in the RATS 4.2 manual.  Suppose I had a quarterly 
series w, from 72:1 to 92:4 in the text file w.txt.  I could deseasonalize its 
growth rate as follows, and write it to the file wdeseas.txt.

cal 72 1 4
allocate 92:4
*
open data w.txt
data(org=obs) / w
close data
log w
difference w
linreg w 72:2 92:4 detrend
# constant
prj removed
*
frequency  2 204
rtoc 72:2 92:4
# detrend
#    1
fft 1
cmask 2 / (204/4) 9
cmult 1 2
ift 1
ctor 72:2 92:4
# 1
# deseason
*
set deseason  = deseason+removed
graph 2
# deseason 72:2 92:4
# w 72:2 92:4
open copy wdeseas.txt
copy(format=free,org=obs) 72:2 92:4 deseason
close copy
end

The 3-digit data were obtained in NSA form from the Federal Reserve Board, and were
deseasonalized using the X11 filter program that is a companion to RATS.