Census of Manufacturing
Much empirical work in economics uses census data to measure industry characteristics such as industry growth, concentration, number of firms, etc.. These data are contained in the Census of Manufacturing bulletin, Concentration Ratios in Manufacturing. These data for the years 1947 - 1992 are summarized in the linked document, concdat.hqx (EXCEL format or text format) which can be downloaded from this page. The latest concentration data (1997) are available in a pdf file online at the official Census Website.
Beginning in 1997, the census data use the new NAICS industry definitions rather than the previous SIC definitions. A "bridge" table that relates the definitions can be found here. You can use this table to translate the SIC code found in Compustat, for example, with the census industry data.
The data definitions for the 1992 EXCEL file are:
SIC Code = Standard Industrial Classification code (2000 - 4000).
YR = Census year (1947 - 92)
N. Firms = The number of firms classified in this industry code.
Value of Shipments classified in this industry code.
Concentration Ratios = Percentage of industry output produced by the largest 4 (8, 20, 50) firms in this industry.
Herfindahl Index = Another (better?) measure of industry concentration.
Specialization Ratio = Extent to which items produced by firms in this industry are classified in this industry.
Coverage Ratio = Extent to which items classified in this industry are produced by firms classified in this industry.
Common uses for these date include measuring competition (Concentration Ratios, Herfindahl Index, Number of Firms) and industry growth (percentage change in the number of firms or value of shipments).
Modified by Jws on 20 February 2002