A new paper by Mark Curtis of Wake Forest University, published by the National Bureau of Economic Research, provides estimates of the employment effects of a cap-and-trade program for electric utilities and large manufacturers. Curtis analyzes the impact of the cap-and-trade program for nitrogen oxide (NOx), which is a precursor for ground-level ozone. The EPA established the Nitrogen Oxide Budget Trading Program, which took effect in eight states and the District of Columbia in 2003 and – after a series of legal battles – took effect in an additional eleven states in 2004. The program would eventually regulate emissions of NOx from nearly 900 electric utilities and 270 large manufacturing industrial units.
It is challenging to determine the causal impacts of a regulation using observational data, since industries subject to the regulation are likely different in many ways from industries not subject to the regulation (even before the regulation is in place), and those differences could lead to different outcomes. Ideally, from a research standpoint, one would randomly assign the regulation to some industries but not others, and then could credibly conclude that the ensuing differences in outcomes were due to the regulation.
Given that EPA does not regulate randomly (thankfully!), Curtis instead exploits the variation in the NOx program that occurs across time (before versus after the regulation, with the regulation beginning in 2003 or 2004) and across geography (industries in some states are subjected to the regulation, industries in other states are never subjected to the regulation). In addition, Curtis exploits variation in industry energy-intensity levels, since high-energy industries would be expected to be more effected by the regulation than low-energy industries.
These variations allow for a difference-in-difference-in-difference estimation that compares the employment response of high-energy industries after the implementation of the NOx program, relative to low-energy industries and relative to industries in areas not subject to the program. He finds that employment in the manufacturing sector dropped by 1.3 percent as a result of the NOx program, amounting to about 110,000 jobs in the affected regions. The reduction in employment occurred primarily through a reduction in hiring, rather than an increase in firings. Young workers experienced the largest drops in employment and earnings of newly hired workers fell by as much as 4.2 percent for industries in the top quartile of energy intensity compared to the bottom quartile.
The NOx cap-and-trade program did lead to a substantial reduction in emissions. Cap-and-trade, like an emissions tax, provides the lowest-cost means of reducing pollution, but lowest cost does not imply no cost. Curtis’s paper provides important evidence of the labor market consequences of regulating pollution from the energy and manufacturing sectors.