Different versions of Grunfeld dataset

1958. Original version of data in Grunfeld's PhD dissertation.
1960. Some of the firms were used by Grunfeld and Griliches (1960), including American Steel.
1960. The first widely published version of the data (10 firms) in Boot and DeWit. These 10 did not include American Steel.
1970. Swamy used some of the original firms not included in Boot and DeWit for a random coefficients computation.
1971. Theil reproduced data from Boot and DeWit for 2 firms, for a SUR computation.
1977. Maddala reproduced the Boot and DeWit version (10 firms), although there was a typo for I(1953) for Atlantic Richfield/Refining. Maddala used 81.90, but Boot and DeWit had 91.90 (a slight printing glitch in the IER made it a little bit unclear). This affected Maddala's results on p.216 for Atlantic Richfield. There is apparently also a typo for U.S. Steel K(1946) - 232.6, but apparently should be 132.6 to reproduce results on p.216 for U.S. Steel.
1993. Greene used a subset of 5 firms (firms 1, 4, 3, 8, 2 in Maddala's order). However, the data for U.S. Steel (Maddala's firm 2) were apparently changed: I(1940) from 361.6 to 261.6, and K(1946) from 132.6 to 232.6 .
I have just (5/25/01) discovered these discrepancies, and I haven't checked to see what Boot and DeWit have for these 2 values. The Greene subset has been widely distributed on web pages.
1995. Baltagi reproduced the Boot and DeWit version (10 firms), without the Atlantic Richfield typo of Maddala, and with the apparently correct values for U.S. Steel. His data file is on a web page, and is the basis of the benchmarks here. It was checked by running the regressions on p.216 of Maddala, reproducing all of them except Atlantic Richfield, and checking the original data listing in Boot and DeWit to resolve the typo in Maddala.

References

Baltagi, Badi H., Econometrics, Springer, 1995.
web site with data file: http://www.springer.de/economics/samsup/baltagi
Baltagi, Badi H., Econometric Analysis of Panel Data, second edition, Wiley, 2001.
web site with data file: http://www.wiley.co.uk/baltagi/
Boot, J.C.G. and DeWit, G.M., "Investment Demand: An Empirical Contribution to the Aggregation Problem," International Economic Review, 1960, p. 27-28.
Greene, William, Econometric Analysis, Prentice-Hall, 1993, ..., 2000
Grunfeld, Yehuda, "The Determinants of Corporate Investment," Unpublished Ph.D Thesis, Department of Economics, University of Chicago, 1958.
Grunfeld, Yehuda and Griliches, Zvi, "Is Aggregation Necessarily Bad?" Review of Economics and Statistics, February 1960, p. 1-13.
Maddala, G.S., Econometrics, McGraw-Hill, 1977, p. 214-216.
Swamy, P.A.V.B., "Efficient Inference in a Random Coefficient Regression Model," Econometrica, 1970, p. 311-323.
Theil, Henri, Principles of Econometrics, Wiley, 1971, p. 296.

The above material was last modified in January 2002, while working with Baltagi's (accurate) downloadable data file.
In December 2008, Christian Kleiber and Achim Zeileis conducted a very thorough examination of the different versions of the Grunfeld data and reproduced published analyses of the data in their paper The Grunfeld Data at 50:
http://statmath.wu-wien.ac.at/%7Ezeileis/grunfeld/
They examined Grunfeld's original dissertation and their analysis is quite definitive and interesting.
Kleiber and Zeileis are careful to note that the company referred to as AR is Atlantic Refining, not Atlantic Richfield (ARCO), which they point out was not formed until Atlantic Refining merged with Richfield Petroleum in 1966.
They also provide the original 11-firm version of the data.