Our dataset integrates several of the leading research projects on corporate networks in the postwar United States. The data from the 1960s were derived from the Mathematical Analysis of Corporate Networks (MACNET) project, directed by Michael Schwartz (see Mintz and Schwartz, 1985, for a description). Data from the 1970s were drawn from a project directed by Gerald Davis, Eric Neuman, and Mark Mizruchi. The data from the 1980s through 2003 were collected by Gerald Davis (see Davis, Yoo, and Baker, 2003). And the data from 2005 and 2010 were collected by Johan Chu (see Chu, 2012).
Although we drew on several data sets, the directors of these projects used similar data collection strategies. All of the efforts were designed to document the networks of leading American corporations using the best available information sources for the time period studied. In each case, the researchers began by identifying the set of leading corporations, usually by means of the lists in Fortune, although Chu used the largest firms ranked by Standard & Poor. For the networks prior to 1990, the researchers identified director names and affiliations primarily from Standard & Poor’s Register of Corporations, Directors, and Executives and Moody’s Manuals, with supplemental reference to Dun and Bradstreet’s Million Dollar Directory, Directory of Directors of New York, and records of the Securities and Exchange Commission (SEC). From 1990 to 2003, the primary data source is SEC data on board composition distributed through Compact Disclosure and Global Access. For 2005 and 2010, the information portals were RiskMetrics and BoardEx. For all of the data, the researchers matched names across boards, and then checked to ensure that the identical names across firms actually represented the same individuals. Further details are available in the references cited above.
After piecing together these datasets, we then selected subsets of the data files to conform to the requirements for this project. In particular, we selected the largest public firms by assets for each time point according to the quota scheme for large countries: 150 manufacturers, 50 financial institutions, and 50 other firms (utilities, mining, transportation, and service). All of the datasets on which we drew for our analysis collected data on more firms in each category than was necessary for our purposes. This ensures that we do indeed have the largest public firms in each subgroup for each year.
