Economics professors at two major universities just issued their reviews of industry-funded assessments of the costs of EPA’s proposed boiler rule (via NRDC). The professors’ conclusions: “the methodology is fundamentally flawed;” “the resulting estimates of job losses are completely invalid;” “the results reported are useless;” “if I were grading this, I would give it an F.” These strongly-worded indictments should make us sit up and take note.
Professors Charles Kolstad and Jason Shogren were asked to review industry-funded estimates of the costs of EPA’s proposed boiler MACT rule. These estimates have been cited in support of recent industry claims that it would be too costly and result in a large loss of jobs. The professors’ reviews usefully reveal the serious flaws in the “evidence” around which industry has been spinning its anti-regulatory story. In an earlier post, I examined another aspect of the industry story, showing how it neglected to mention that the estimated benefits of regulating particulates, heavy metals, and dioxins would dwarf the estimated costs (and that’s without even counting benefits for mercury reductions). With the assistance of Kolstad and Shogren, we can see how, in addition to belittling the benefits, the industry story has attempted to exaggerate the costs.
Here is some background reading for this assignment:
How does environmental regulation affect jobs?
As Kolstad explains, environmental regulation can be expected to have three impacts within the regulated industrial sector:
“Changes in employment from regulatory action are from three sources: losses due to contraction of the domestic industry due to price increase induced by regulations, abatement activities that require additional labor, and production processes post-regulation that may have different labor requirements. The first of these is negative, the second of these is positive, and the third of these is ambiguous.”
Professor Kolstad points out that the consultants funded by the American Forest & Paper Association counted only the first of these — that is, the job losses that would be experienced in the pulp and paper industry. Historically, however, when all three of these impacts are considered, air pollution regulations have had a “statistically insignificant number of jobs lost” in the pulp and paper industry. The three impacts effectively cancel each other out for this sector.
But there is more to this lesson:
“In addition, there will likely be job gains in the sector providing the pollution control equipment purchased by newly regulated firms.”
This facet of the inquiry is often neglected – an error that has undermined a thoughtful accounting of the interplay between environmental regulation and jobs. Professors Frank Ackerman and Lisa Heinzerling (respectively, current and former CPR Member Scholars) have observed that the “jobs versus environment” rhetoric perpetuates the myth that environmental regulation creates no new jobs. In fact, as recent experience in the context of mercury emissions control demonstrates, environmental regulation is often a crucial driver, creating control technology markets and the jobs these markets require. Consider testimony delivered in March before the Senate Committee on Environment and Public Works by Dr. Michael Durham, CEO of a company that develops and commercializes air pollution control technologies. Durham said a robust commercial market in mercury control “is well under way,” driven by new regulations of mercury emissions from power plants in nineteen states, as well as existing federal regulations on new plants. As of the spring of 2010, over 150 contracts had been awarded for mercury-specific control technologies, such as activated carbon injection (ACI). Moreover, “a great deal of expansion of activated carbon production is currently being planned, but is contingent on a federal mercury regulation” for power plants. A rule that requires a 90% reduction in mercury emissions would create the need for new activated carbon production plants, resulting in an estimated $2 billion in construction jobs. In addition, these regulations would result in the creation of operating jobs at the activated carbon production facilities, plus mining jobs to supply the feedstock material needed to make activated carbon.
Professor Shogren points to a similar flaw in the estimates of the consultant funded by the Council on Industrial Boiler Operators. While their estimates counted projected contraction of jobs in the regulated industrial sectors, they ignored entirely the expansion of jobs in other sectors; among other things, “workers will be needed to produce, install, maintain, and monitor pollution control equipment.”
Although the industry-funded studies were quick to tally the negative impacts, they ignored entirely the positive impacts of the proposed boiler rule. The result is a sloppy accounting – and one that undermines informed and thoughtful regulatory responses to environmental harms.
How accurate are projections of the costs of environmental regulation?
The costs of environmental regulation are generally more amenable to quantification than the benefits, but these cost estimates still tend to prove inaccurate. Importantly, they tend to overestimate – sometimes by orders of magnitude – the actual costs of complying with environmental regulation. This is so because once there is certainty that regulation is forthcoming, markets respond to stimulate innovation by the private sector. This point has been demonstrated empirically; one study, for example, found that advance estimates of the cost of compliance were more than twice the actual compliance costs for eleven out of twelve regulations. The more recent experience of mercury regulation similarly bears out this claim, as Dr. Durham’s Senate testimony illuminates. According to this testimony, back in 2004, the difficulties related to capturing mercury from western coals were thought to make 90% reductions “not achievable for many plants” and led to costs estimates of more than $100,000 per pound of mercury removed. “However, technology developers working in concert with their power producing customers discovered the root cause of this limitation, then they developed new chemically-treated sorbents to overcome the problem.” Now, a 90% reduction in mercury emissions from power plants burning western coals is readily achievable at costs under $10,000 per pound of mercury removed. The costs of control are expected to drop even further in the future, moreover, as improved sorbents are expected to overcome other limitations.
Yet the industry-funded estimates of costs for the proposed boiler MACT appear to assume worse than the worst. As Professor Shogren explains, the consultants for the Council of Industrial Boiler Operators used “high-end estimates of capital costs to upgrade” and assumed no innovations in technology, arriving at costs “nearly 100 times greater” than EPA’s estimates. As Professor Shogren observes, while economists might assume different rates of cost-cutting technological innovation, the consultants’ failure to envision any new technological development in the face of such a large projected cost “is an assumption challenging to defend.”
In the end, the criticisms leveled by Professors Kolstad and Shogren are not mere quibbles with this or that input to the industry consultants’ estimates. Their reviews identify fundamental methodological flaws – errors that, they make clear, they wouldn’t tolerate from their students.
As I have argued elsewhere (“Mathematics of Mercury” at page 108 in Reforming Regulatory Impact Analysis), there are serious limitations in relying on cost-benefit analysis to determine the appropriateness of environmental, health, and safety regulation. When government casts the debate in terms of costs and benefits, it arguably invites the same from other participants. This most recent chapter in discussions of the proposed boiler MACT rule highlights two of these infirmities. First, cost-benefit analyses are highly malleable. Whoever wields the calculator has enormous power over the inputs to and, so, the output of the cost-benefit equation. Second, cost-benefit analyses lack transparency. How are members of the public supposed to judge consultants’ claims about job loss when these claims are buried within a jargon-laden economic analysis? As Professor Rena Steinzor once put it: it can take “an experienced team of economists with weeks of free time on their hands” to review all of the many judgments that go into a cost-benefit assessment for environmental regulation.
Thankfully, in this instance, there was an experienced team of economists.