How Well Does the U.S. Government Do Benefit-Cost Analysis?
* Mr. Hahn is co-founder and executive director of the American Enterprise Institute–Brookings Joint Center for Regulatory Studies and a scholar at AEI. Mr. Dudley is a former researcher at the AEI–Brookings Joint Center. The authors would like to thank Gregory Besharov, Cary Coglianese, Scott Farrow, Art Fraas, Myrick Freeman, John Graham, Scott Jacobs, Randy Kroszner, Lester Lave, Randy Lutter, Al McGartland, Richard Morgenstern, Sheila Olmstead, Paul Portney, Eric Posner, James Prieger, Kerry Smith, Robert Stavins, Paul Tetlock, Scott Wallsten, Richard Zeckhauser, and seminar participants at Harvard University for helpful comments. The authors would also like to thank Caroline Cecot, Jordan Connors, Laura Goodman, Jesse Gurman, Elisabeth Irwin, Katrina Kosec, Troy Kravitz, Rohit Malik, Mary Beth Muething, and Shenyi Wu for help with this research. The views expressed in this article represent those of the authors and do not necessarily represent the views of the institutions with which they are affiliated
To make prudent recommendations for improving the use of benefit-cost analysis in policy settings, some measures of how well it is actually done are essential. This article develops new insights on the potential usefulness of government benefit-cost analysis by examining how it is actually performed in the United States.
We assess the quality of a particularly rich sample of benefit-cost analyses of federal regulations. The data set we use for assessing the quality of regulatory analysis is the largest assembled to date for this purpose. The seventy-four analyses we examine span the Reagan administration, the George H. W. Bush administration, and the Clinton administrations. The article is the first to assess systematically how government benefit-cost analysis has changed over time.
There are three key findings. First, a significant percentage of the analyses in all three administrations does not provide some very basic economic information, such as information on net benefits and policy alternatives. For example, over 70 percent of the analyses in the sample failed to provide any quantitative information on net benefits. Second, there is no clear trend in the quality of benefit-cost analysis across administrations. Third, there is a great deal of variation in the quality of individual benefit-cost analyses.