This research is based on a database described as “as close to an annual census of American business as exists,” enabling researchers “to focus on the components of growth.” It is known as the National Establishment Time Series database, and it includes employment and sales data for individual companies.[*]
The NETS database tracks establishments across the U.S. over time, each with their own unique identifier. Scholars can follow firms over the years, where they move, how many opened or closed, how many workers they employed and their annual sales. The specific dataset used for this analysis contains more than 60 million American establishments, spanning 1990 through 2015.
By matching companies in the NETS database to firms known to have been provided state incentives, we can measure the performance of these firms and compare them to similar firms that did not receive incentives. This is arguably the closest we can get to a controlled experiment when studying the efficacy of state economic development incentives. One group in the study is the treatment group — having been offered incentives — and the other serves as a control group, allowing for comparisons.
Did the treated firms (offered incentives) create more jobs or have higher sales than the control group? If so, at what cost? Those are research questions we intend to answer for nine types of Michigan incentive programs, including the previously studied MEGA and MBDP. This report will focus exclusively on firms that were offered incentives by the state of Michigan, analyzing more than 7,300 incentive deals from nine programs or program types going back to 1983.
[*] The National Establishment Time Series is produced by Walls & Associates using data from Dun & Bradstreet.