Debating the EPA’s Clean Power Plan Proposal — The Clean Power Plan’s Equity Problems

Editor’s Note: The post is part of a discussion on the EPA’s Clean Power Plan proposal between Philip Wallach, a fellow in Governance Studies at Brookings and the author of the upcoming book, Legality, Legitimacy, and the Responses to the Financial Crisis of 2008 (Brookings Press, 2015), and Bob Sussman, former senior policy counsel to the EPA administrator.

In his response to our July 16th Wonkblog article, former EPA Deputy Administrator and Senior Policy Counsel Bob Sussman argues that we fail to give the EPA the credit it is due for its Clean Power Plan.  Sussman argues that the EPA’s approach to limiting greenhouse gas emissions from existing power plants is far from arbitrary, as it relies on a four-part formula based on four of the most cost-effective ways of reducing carbon intensity from the power sector.  We disagree with Sussman’s characterization of our argument and stand by our initial criticisms.  EPA’s hugely disparate treatment of states, including those with similar energy mixes, does not result from arbitrary deviation from its formula, but precisely because its formula makes a number of assumptions that are difficult to justify as fair.

Where do EPA’s goals come from?

At the outset, Sussman wants to emphasize that EPA’s action is “rooted in section 111(d) of the Clean Air Act,” which requires state-driven development of performance standards within a framework provided by the EPA.  He rightly notes that, working from this legislative premise, EPA decided that its framework should be defined “broadly,” going beyond plant-specific improvements to “system-wide changes in how electricity is produced and used.”

That’s all fair enough, but most readers are probably unaware of just how little guidance 111(d) actually provides to the agency.  The main pollution-control features of the act run through different, better-fleshed-out sections, with 111(d) providing a backstop for otherwise unregulated pollutants—but EPA has never been in a position to rely on it for such an important rule before.  The section has some legal issues that may undermine EPA’s efforts moving forward, but even putting those aside, the law tells EPA very little about how it is to develop standards.  EPA has interpreted “best system of emissions reductions” to apply to power production writ large, rather than thinking about each plant’s capabilities, and by doing so it aims to support development of cost-effective plans for overall emissions reductions—a laudable goal.  But, without any real guidance from the statute, it has cast its net so wide that it is essentially prescribing energy policy goals for each state.  Doing so is unprecedented under the Clean Air Act, and means that the burden is on EPA to show that its interpretation is appropriate.  When states seem to be treated inequitably under the EPA’s proposal, EPA should be prepared to offer substantive justifications for those discrepancies better than those it has given so far.

What drives the differences between states, and what are those differences?

The crux of Sussman’s (and EPA’s) argument is that the differences in expectations between states are perfectly reasonable, the result of a consistent application of a formula that accounts for differences in energy mix, geography, electricity market structure, and past progress in developing renewables.  Differences in treatment are based on meaningful differences in circumstances.

But saying that EPA’s rule is fair because it adheres to the agency’s formula merely shifts the question: what exactly is in the formula, and what kinds of outputs does it have?  Answering these questions is far from easy, but we have done our best to dig into the technical documents supporting the proposed rule, explaining the different steps of the formula for interested readers in a Brookings explainer.

Let us start with the outputs—49 state goals for improvements in carbon intensity (measured in lbs of CO2 per megawatt hour) by 2030 ranging from 10.6 percent (North Dakota) to 71.9 percent (Washington).  While most states are in the 20-40 percent range, the extent of variation is large enough to cause political resentments.  And, as we pointed out in our original article, the differences are in many ways counter-intuitive. 

Our headline takeaway—that states emitting greenhouse gases at the highest rates are generally asked to reduce their carbon intensity by the lowest proportions—is based on a complete look at EPA’s prescribed goals for all 49 states covered by the rule—not just by looking at three examples (as Sussman suggests we do).[1]  Sussman points to three low-emissions states asked to make relatively modest reductions, in part because they cannot shift away from coal that they already do not use.  But these are outliers from the general trend, and many other states that use little coal are expected to make large emissions reductions (e.g., Washington, Oregon, Idaho, New Hampshire, and New Jersey).

Renewables: Realism or Politics?

Turning to the process that produces these standards, embodied in the four-part formula, does not dispel concerns of unfairness, especially for the sources affected by Block 3.

EPA’s expectations for renewable growth vary enormously from region to region, from as little as 6 to as much as 17 percent annual growth.  Sussman and EPA say that these differences are justified by the geographic variations across states, so that any sense that treatment is inequitable is misguided.  But the calculation method, described in our explainer, embedded in EPA’s formula makes this claim problematic.   

Prescribed renewable growth rates are derived from what states in a region have previously aimed to achieve through their own Renewable Portfolio Standards.  These goals may, in part, reflect geographic potentialities, as EPA argues, but just as often they are based on politics, aspirations for a green economy, or elected officials’ desire to look like they are doing something about climate change.  Some states have set goals that are probably quite unrealistic—which is one thing when the state legislature can recalibrate them (or pause them, like Ohio legislators recently did) as they take stock of progress, and another thing when EPA has locked them in with a federally-prescribed goal.  The affront to the states’ democratic sovereignty is all the worse because it effectively imposes the goals set by neighboring states as enforceable goals.  Just as much as regional geography, then, EPA’s approach makes states’ renewable expectations dependent on the politics of their region.  Sussman’s only problem with EPA’s choice is that it is “too lenient toward states with a poor record of promoting renewables and should spur them to be more aggressive in its final rule.”  But this complaint amply proves our point: it is quite arbitrary to derive differential treatment from political climates, whether that results in stringent or lenient standards.

EPA would better serve its long-term goals if it focused the renewable portion of its rule exclusively on state-by-state geographical conditions, which directly impact the costs and benefits of renewable development, rather than including state-set (and, up until now, state-breakable) goals that may not be realistic.

Nuclear Fission: The Divergence between Alabama and Georgia

When we turn to the EPA’s treatment of nuclear power, the basic arbitrariness embodied in EPA’s formula is even starker.  To demonstrate, consider the requirements for Alabama and Georgia, two geographically similar neighboring states with nearly identical energy mixes (around one-third coal, one-third gas, a bit more than a quarter nuclear, and a couple percent renewables).  In spite of their similar profiles, to satisfy the EPA’s proposed Clean Power Plan, the two states will need to achieve very different improvements in carbon intensity by 2030 (26.7 percent for Alabama, 44.4 percent for Georgia).  How can it be that states that are almost identical in their energy profile, geographically similar, and in the same region are expected to make such different reductions in their overall emissions rates?

The difference comes largely because of the rule’s confusing treatment of nuclear power.  Existing plants are treated identically in all states—itself problematic since the age and expected operating life of nuclear plants differs significantly.[2]  What drives the divergence is EPA’s treatment of two nuclear reactors currently under construction in Georgia’s Vogtle plant, which will have their expected output added in full to Georgia’s denominator—significantly increasing the state’s expected carbon-intensity reduction.  If, for some reason, the construction were to end unsuccessfully, Georgia would be in dire straits trying to fulfill the standard EPA has set for it under the assumption of completion.  Had physical work on these reactors started after the rule comes into effect, they could potentially have been a large asset in Georgia’s efforts to comply with the rule; instead, they become significant potential liabilities.  Alabama, meanwhile, has a reactor in the planning stage; if it manages to encourage construction of that plant as a part of its compliance plan, doing so would take it a long way toward meeting its goal.  That this proposed treatment of under-construction nuclear reactors (which include just the two in Georgia, two more in Tennessee, and one in South Carolina) is technically embedded within a formula does not make it fair, and EPA should consider revising its treatment of them in a way that would reward recent investments rather than effectively penalizing them.

Closing Thoughts

Climate change is among the greatest challenges that our nation and world will contend with in the decades to come, and it is understandable that environmentalists and the EPA feel the imperative to get something meaningful into U.S. law right away.  The Clean Power Plan is most definitely something—and, indeed, in its attempts to promote cost-effective reduction measures and accommodate important state differences, there is much that one can commend in it.  But the interstate equity issues we raised in our original post and in this response strike us as serious shortcomings.  At least in part, EPA can work to remedy them on its way to the final rule, and we hope it will do so.  Even better would be if Congress provided a clearer way forward—with a carbon tax being the most appealing alternative—so that EPA is not forced to tortuously interpret the subtleties of a forty year old law and to slog through the years of bitter legal and political battles that await the Clean Power Plan

[1] Our finding is based on a regression analysis that shows a relationship that is statistically significant (p=0.02) and important in terms of effect size (the least carbon-intensive state should expect to be charged with a reduction in carbon-intensity about 15% greater than the most carbon-intensive state).  Significant regressions are also behind our assertions that more coal generally leads to smaller percentage reduction demands, as does having less nuclear.

[2] Existing nuclear plants are referred to as “at risk,” because of the possibility that they could close down as they reach the end of their licensed operation periods—are multiplied by 0.058 and added into the denominator of the EPA’s formula, meaning that both states are effectively held responsible for maintaining their nuclear output less 5.8 percent from 2012 to 2030.  That figure is calculated based on EPA’s understanding of the proportion of nuclear power at risk nationwide, which it then applies uniformly.  We see no reason why, if EPA is already using such a complicated formula, it should not explicitly adjust for plant ages on an individualized basis using data already available to the agency.