Measuring Regulation

Cylinder dreams passing in stages
Lethargic grins left to bare
Broadway window, cubical cages
Where escape is fairly rare
--Ric Ocasek

Academic study provides empirical evidence of negative relationship between regulation and industry productivity. Such results should not be surprising, and we have discussed this relationship many times on these pages.

The paper might serve as a good reference for its conceptual development of the regulation:productivity argument as well as for its examples of behavior in regulated industries.

The paper is also interesting in how it measured extent of regulation. Simply counting the number of regulations that restrain industry behavior is problematic since some regulations have much greater inhibiting effect than others.

For example, Federal Register page counts, while providing some sense of general regulatory trends, ignores the extent to which some regulations might influence behavior. Applied to the regulatory context, the Pareto Principle suggests that relatively few regulations are likely to have outsized impact.

A better measure is to use counts of binding words, such as "shall," "must," and "prohibited" that appear in regulatory text. Binding word counts are appealing because they better reflect the constraints imposed by regulations rather than wordiness.

The authors used a binding word count index developed by RegData for 96 industries. They split the industries into three groups--high, medium, low regulated--and then compared them using various measures of productivity obtained from the BLS. Their data set was longitudinal, covering 1997-2010.

Comparison of means and regression analysis were straightforward. Results demonstrating the negative effects of regulation on productivity are clear.