In these pages four years ago (“Obfuscation at the EPA,” Summer 2011) we announced our discovery of the U.S. Environmental Protection Agency’s new methodology for ascribing job gains to costly new regulations: a sleight-of-hand whereby it attributes the workers hired to ensure compliance with new regulations as a benefit rather than an economic cost. It is execrable logic, compounded by the agency’s failure to acknowledge the much larger long-term job losses from rendering entire industries less competitive from higher costs resulting from the regulation. Other economists joined us in deriding the fatuousness of such claims and the EPA refrained from such inanity—for a while.

Now, the agency has returned to this duplicitous Keynesian logic, this time in defense of its greenhouse gas standards for power plants, the “Clean Power Plan” (CPP). According to the EPA, the plan would create nearly 80,000 energy efficiency jobs by 2020. That is a breathtaking claim even by the agency’s standards: the EPA offers no peer-reviewed data that come close to supporting that extravagant employment figure and it avoids acknowledging that compliance costs for the CPP, when fairly measured, would exceed $30 billion in 2020. What’s more, the rule would shutter approximately 100 power plants across the nation and increase energy costs for all consumers and businesses. To pretend that the CPP would merely transfer tens of thousands of coal jobs into renewable energy and efficiency jobs, with more people employed as a result of the rule, is a truly breathtaking distortion of reality.

Uncharted territory / The notion that imposing higher compliance costs on a company somehow results in more jobs is not just the blithe justification of some EPA mandarin. In her speech announcing the new greenhouse gas standards, EPA Administrator Gina McCarthy declared that “utilities like Exelon and Entergy are weaving climate considerations into business plans. All this means more jobs, not less.” It’s a claim that sits on a house of shaky inferences, improper extrapolations, and shoddy math.

The EPA acknowledges it is, in effect, inventing an entirely new cost-benefit methodology, conceding in the Regulatory Impact Analysis (RIA) for the CPP that “employment impacts of demand-side energy efficiency programs have not been extensively studied in the peer-reviewed, published economic literature.” But just because the entire economics profession has not figured out how to do something is no reason for EPA officials to not do it themselves.

One problem—but far from the only one—with the EPA’s job creation methodology is that the agency uses data from the Annual Survey of Manufacturers, which is a broad but far from in-depth data set, to derive a multiplier of energy efficiency spending to jobs. However, only half of the data it takes from the survey relate to the energy efficiency sector. With half a sample, the agency then assumes the multiplier in the construction industry is the same as the manufacturing sector.

After using data only tangentially related to energy efficiency to infer that construction and manufacturing support the same employment levels, the EPA posits a jobs multiplier whereby $1 million in demand-side energy efficiency creates 2.56 new jobs. That number is at odds with other studies used more often by the EPA; for instance, Richard Morgenstern, an economist at Resources for the Future, estimates that $1 million in environmental spending supports only about 1.5 new jobs, well less than the EPA’s new claim. However, Morgenstern’s own estimates have a standard error of 2.2 jobs, which suggests that the true effect of such measures on jobs is likely to be “insignificant.” Despite the EPA’s previous reliance on Morgenstern, the agency is now happy to jettison his research and instead embrace claims that energy efficiency spending will reap greater employment benefits.

The very notion of a spending multiplier for some government action is controversial and one that various government entities have spent significant time studying. For example, the Congressional Budget Office analyzed the effects of the 2009 economic stimulus and quantified various output multipliers ranging from 0.4 for deferral of income to 2.5 for the “purchases of goods and services by the federal government.” Each multiplier represents the direct and indirect effects on the nation’s output of a dollar’s worth of given policy. However, the EPA posits a higher multiplier than any of the CBO’s estimates, even though the EPA’s actions take place in a robust economy when any multiplier effect would be muted relative to a period like the Great Recession of 2008–2009.

Perhaps the most glaring fault in the EPA’s calculus is the disconnect between its estimated 78,800 jobs created by the CPP and the economy-wide compliance costs of the regulations. While the agency claims demand-side spending would range from $8 billion to $12.3 billion by 2020, the EPA’s multiplier of $1 million in demand-side efficiency spending creating 2.56 new jobs implies compliance costs exceeding $30 billion.

If the EPA’s math is wrong about the short-term effect of the CPP rule when it is phased in beginning in 2020, what implications does that have for the agency’s economic analysis regarding CPP’s ultimate effect in 2030, when the agency predicts anywhere from $42 billion to $51 billion in energy efficiency spending? If the EPA is off by a factor of three in 2030, as it seems to be for 2020, it would imply that the total implementation costs imposed by the CPP are somewhere north of $150 billion.

Real employment implications / The EPA has been regulating power plants for some time. Its Maximum Achievable Control Technology standards for power plants was the costliest rule issued under President Obama when it was introduced, imposing $10 billion in compliance costs, while its Cross-State Air Pollution rule added another $1 billion in regulatory compliance costs. Last year the EPA finalized rules controlling cooling water intakes at power plants, creating another $300 million in compliance costs per annum.

The result of those regulations has been a gradual diminution of employment in power plants since 2008, which is the opposite of what the Obama administration would have us believe. Now, it is true that job creation should never be the sine qua non of government activity, and in this economy there has been undue weight placed on the employment gains—both real and imagined—from government activity. As Milton Friedman loved to point out, we could get to full employment in a jiffy if we mandated that people build roads with shovels instead of bulldozers. But to pretend that a rule that would close 100 power plants and boost energy costs throughout the economy would be a job-creator is a distortion of epic proportion.

Setting aside the half-baked Keynesian stimulus nonsense, here is what we know, using the EPA’s own numbers: Even with $8.8 billion in estimated compliance costs, the EPA admits that the CPP will result in the premature retirement of 50 gigawatts (GW) of coal-fired power, which represents one-fifth of current capacity. The agency’s previous forays into power plant regulation already managed to retire approximately 51 GW. The EPA projects that retiring 50 GW would destroy 15,600 jobs by 2020 along with an additional 14,100 employees laid off by 2025, with a total of 42,000 fewer employees in the power plant sector by 2030.

The EPA does not model exactly which power plants and which states would be most affected, but some sorting using the agency’s data allows us to infer how this might happen. Under the CPP, the EPA would calculate most states’ ratio of carbon dioxide output to megawatt hours of energy production (CO2/​MWh). Plants with high ratios would drag up state averages and make it more difficult to comply with the new standards. Shuttering inefficient plants is one compliance path to lower a state’s CO2/​MWh rate.

Using the EPA’s e‑Grid data, we sorted the power plants with the highest CO2/​MWh and nominal heat rates that might be candidates for retirement. We found 93 plants in 31 states with an average CO2/​MWh rate of 2,626, or roughly 20 percent less efficient than the average coal plant. Those facilities produce about 50 GW of power, or what the EPA forecasts for early retirements under the proposal. Based on those data, Pennsylvania is the most affected state, with 13 plants in danger of retirement. Michigan is next with seven possible retirements and Illinois has six that would presumably close because of the new rules.

The final flaw in the EPA’s RIA is that it does not appear to monetize the 42,000 lost power plant jobs. For a baseline, the University of Chicago’s Jonathan Masur and Eric Posner estimated $100,000 per displaced worker, which would add another $4.2 billion to the cost of the CPP, or almost half of the EPA’s cost estimate.

Worthwhile analysis? / The EPA analysis of the CPP illustrates the sad fact that the agency often uses RIAs more as a way to sell its preferred regulatory intervention than a dispassionate attempt to measure the costs and benefits and explore real alternatives to what it’s proposing.

Sorting through the rigmarole offered in the CPP RIA nicely illustrates the need for an independent agency to do the cost-benefit analysis of regulatory agencies’ proposals so that the White House can be sure that the agencies aren’t stacking the deck. It’s a solution we’ve offered in these pages before, and it remains as timely and as far from actually happening as when we first made that suggestion.