Fortune 500 data for size distribution analysis

In 2006, for the investigation of the goodness-of-fit of the Pareto Type III distribution, I manually downloaded the annual revenues of the top 500 firms in the Fortune database. You can download the compressed test file Fortune 500.

This file was further processed to obtain the distribution of revenues, available here. These data can be directly used with the gbhill utility in the gbutils package. For instance the command

gbhill pareto1 1 10 -v 1 -O 0 -u .05 < fortune_2006_size.txt.gz

will print the estimated upper tail behavior of the dataset using observations in the top 5%.