Monday, June 6, 2011

Does federal regulation impose costs of $1.75 trillion per year?

Last week, Cass Sunstein--a Chicago law professor and economist released a plan for regulatory reform.  Commentary on the proposals sometimes mentioned that the reforms were small in size, because federal regulations impose a cost of $1.75 trillion on the U.S. economy. As best as I can tell, that number comes from a study done under contract for the Small Business Administration by  Nicole V. Crain and W. Mark Crain, two economists from Lafayette College. Here, I ask two overall questions about that interpreting that estimate. 1) How much can this estimate of the costs of federal regulation be trusted? 2) Even if the estimate is true, what about potential benefits of regulation?

 1) How much can this estimate of the costs of federal regulation  be trusted? 

An estimated cost of $1.75 trillion is certainly large. Is its size plausible? As you would expect, the authors break down their estimates into some major categories, which looks like this:


I'm not familiar with all the underlying sources here, but the environmental and tax compliance numbers are roughly in line with other estimates I've seen. But the single largest factor here is the economic cost, which is determined by an "original regression analysis using World Bank Regulatory Quality Index." The Regulatory Quality index is one part of the World Bank's overall Worldwide Governance Indicators. As researchers who help to put together this data point out, it is based on survey evidence of from a mixture of commercial business information providers, public sector organizations, nongovernment organizations, and firms and households.

The key regression uses data on 25 OECD countries from 2002 to 2008. It uses per capita GDP as the dependent variable. The Regulatory Index for that country is one explanatory variable. Other control variables are foreign trade as a share of GDP, country population, primary school enrollment as a share of the eligible population, and fixed broadband subscribers per 100 people. Crain and Crain choose a "minimal" level of regulation, look at how far the U.S. is above that level, and then use the estimated coefficient to project how much U.S. per capita GDP is reduced by having that higher level of regulation.


As Crain and Crain point out, this broad approach is not new to them: other academic studies have used indexes of regulatory quality from the World Bank, OECD, and other sources. The Office of Management and Budget has experimented with a similar approach in the past. I'm not aware of any efforts to build up estimates of economic costs of regulation one regulatory agency and one main law at a time. But this regression approach is clearly open to a lot of questions. Cross-country regressions are notorious for having a hard time getting at cause and effect. How much does one trust a result based on seven years of data on perceptions of regulation? As the authors point out, they did a similar study back in 2005 using OECD survey data on regulation, and comparing that study with this one implies that the cost of U.S. regulation rose 70% from 2004 to 2008!

Overall, I wouldn't be shocked if the economic costs  of regulation are large. But I lack confidence in this specific estimate.

2) What about benefits of regulation?

Even if the $1.75 trillion in costs of federal regulation is accurate, it would not address whether those regulations have benefits that might exceed their costs. I don't know of overall evidence on this point (which is more than a little bothersome!), but in the Spring 2002 issue of my own Journal of Economic Perspectives, A. Myrick Freeman III of Bowdoin College wrote on "Environmental Policy Since Earth Day I: What Have We Gained?"

Freeman emphasizes that some environmental regulations are cost effective, and others not so much. He writes: 
"Among the winners in terms of net economic benefits are the following: the removal of lead from gasoline; controlling particulate matter air pollution; reducing the concentration of lead in drinking water under the Safe Drinking Water Act; the setting of maximum allowable concentrations on some volatile organic compounds under the Safe Drinking Water Act; the cleanup of those hazardous waste sites with the lowest cost per cancer case avoided under Superfund; and probably also the control of emissions of chlorofluorocarbons. ... The environmental rules that appear to be losers in terms of net economic benefits include the following: mobile source air pollution control; much of the control of discharges into the nation’s waterways, with the exception of some lakes and rivers that were especially polluted; and many of the regulations, standards and cleanup decisions taken under the Federal Insecticide, Fungicide and Rodenticide Act, the Toxic Substances Control Act, the Safe Drinking Water Act and Superfund."

It's bothersome but true that many government regulations have not received dispassionate cost-benefit analysis. My guess is that such analysis would encourage more regulation in some areas, where benefits are far in excess of costs, and less regulation in others. The goal should be to get more bang for the bucks spent on regulatory costs. For example, if it is costing corporations $100 billion a year to fill out tax forms, it's worth asking whether there are win-win ways to change the tax code so that corporations spend less and the federal government collects more.