The Clean Power Plan “Shadow Papers”: the Process is Smarter than John Roberts
1. John Roberts Should Have Done Some Homework about the Michigan Ruling 2. When “Executive Acton” Means “Regulatory Rulemaking” it’s a Model of Good Governance
This article is an exercise in shameless self-indulgence. As someone deeply involved in the Clean Power Plan, reading the recently published 2016 memos circulated by John Roberts and five of his colleagues on their sprint to staying the Clean Power Plan was painful, with Roberts’ sloppy citing of an EPA blog post crystallizing, for me at least, everything that was wrong with the Court’s way of doing business in that instance and subsequently via the Shadow Docket. Meanwhile, I’ve been searching for a long time for an occasion to write about how much value the regulatory rulemaking process brings to our government when it is keeping faith with the values of a rule-of-law democracy. The “Shadow Papers” gave me just enough of a pretext for doing that. Successors to the current administration, should they more resemble earlier administrations than the current incumbents, will face myriad choices about what currently eclipsed or discarded attributes of governance they want to recapture. Understanding fully the track record and promise of regulatory rulemaking will be essential to how they make their choices.
Introduction
Last month the New York Times’ published a block buster report on how in 2016 the US Supreme Court went about deciding to stay the Obama EPA’s Clean Power Plan (CPP). https://www.nytimes.com/2026/04/18/us/politics/supreme-court-shadow-docket.html?utm_source=substack&utm_medium=email At the center of the story were memoranda exchanged among six or the justices apparently leaked to reporters Adam Liptak and Jodi Kantor.
At the time the stay itself was shocking; the Court’s issuing it without a word of explanation was appalling to anyone who believes that the Court has an existential obligation to engage in fully informed and careful deliberation before it acts and then to explain clearly to the public, the lower courts, and the justices’ successors, their principled reasoning. The memos, particularly the two authored by Chief Justice John Roberts, show that he and the four colleagues who joined him in issuing the unexplained stay acted without understanding or caring about that obligation.
Experts like Professor Stephen I. Vladeck of Georgetown Law School have provided trenchant analysis of the leaked memos and of the practice, which the CPP stay seems to have inaugurated, of the Court’s issuing high-impact decisions, including decisions that block carefully developed national policies, via the “Shadow Docket”, a practice that more often than not entails short-circuiting the normal sequence of judicial review, full briefing by the parties, and an explanation of the Court’s actions.
As Professor Vladeck put it: “Behind the scenes, Roberts led the charge for the Court to blaze a new trail—relying on statements outside the record; invoking the wrong standard for the kind of relief the applicants sought; failing to even acknowledge the irreparable harm the government (and the environment) would suffer from the Court intervening; and pushing back aggressively when Justices Breyer and Kagan both urged a compromise that should have accounted for his ostensible concerns.”
Many of the points – assertions really – the Chief Justice makes are surprisingly superficial and that superficiality illustrates just how much is lost when the Court eschews the kind of careful, well-informed deliberation so drastic a move as staying a national policy should have received.
The senior leadership team of the Obama EPA, who like me were deeply involved in the CPP, and I shared our thoughts in @TheContrarian:
The upshot of our piece was that the Obama EPA Clean Power Plan was a monument to public engagement, exhaustive and fully informed deliberation, and careful fully explained reasoning. The New York Times’ publication of the memos exchanged among six SCOTUS justices in their mad 5-day dash to stay (and effectively kill) the Clean Power Plan reveals just how little regard Chief Justice Roberts and the four other justices voting for the unprecedented and unexplained stay had for careful, well-informed deliberation and public transparency. Americans are still reaping the bitter harvest of the Court’s stymieing of climate policy and the Court is still reaping the bitter harvest of the self-inflicted harm it has done to its own credibility in the eyes of the public.
The CPP stay proved not to be the last time the Roberts majority summarily dispatched a national policy issued after meaningful deliberation or actions taken by lower courts that had provided carefully built records and presented extensive explanations. There’s no sign that the Court will desist from this Shadow Docket conduct. What’s new, though, is that in his memos, John Roberts displayed distinct habits of mind that are likely to persist going forward.
One of the more curious moves Roberts makes in his first memo is to cite an EPA blog post responding to another, unrelated Supreme Court case. That citation is a microcosm of the hazards Roberts’ extreme haste created. It also may reveal a blind spot in Roberts’ understanding of the chronically and widey underappreciated and misunderstood value and functions of the rulemaking process.
Considering the facts underlying the EPA blog post Roberts cited, citing it was a misfire. Separately, those same facts reflect the way in which the EPA’s rulemaking process embodies the attributes of fully informed deliberation, public transparency, and accountability that are critical to government decision-making in a rule-of-law democracy – attributes missing from the CPP stay decision and increasingly from the Court’s actions under the Shadow Docket.
The Clean Power Plan
The Clean Power Plan was the Obama Administration’s flagship second term climate policy. EPA promulgated the Clean Power Plan under what EPA believed to be its well-established Clean Air Act authority. The CPP set carbon dioxide (CO2) emissions limits on the nation’s coal- and natural-gas fired power plants At the time, the EPA projected that with the CPP in place power sector CO2 emission would fall to 32 percent below 2005 levels by 2030. After more than two years of exhaustive nationwide engagement with the public, utilies, the states, advocates, and just about every stakeholder in the electricity sector, EPA modeled the CPP on the efficiencies intrinsic to the way the pervasively interconnected electricity grid operates, calculating power plant CO2 limits by projecting cost-effective substitution of lower- and zero-emitting generation from natural gas, solar, nuclear, and wind generation for high-emitting generation primarily at coal plants.
The CPP was challenged almost immediately after it was promulgated in October 2015 with its opponents asking the DC Circuit Court of Appeals for an order halting its implementation, but in January that court denied the request. The opponents then petitioned the Supreme Court for the same thing. Even prominent industry lawyers called that move a “long shot”, but in just 5 days of non-deliberation (call it “vibing’?) on the petition, the Court shocked virtually everyone following the story by issuing a stay. The vote was 5-4; the Court offered no explanation for its highly unusual action, which was seen as virtually unprecedented at the time.
A Single Detail Can Tell the Whole Story x 2
As the Vladeck One First post and YouTube discussions in the links above show, the Roberts’ memos include hasty, slap-dash, and ill-founded assertions as if surfaced by an author who first recognized the outcome he wanted to achieve and only then went in search of justifications. For people familiar with EPA’s rulemaking process and the federal courts’ treatment of those rules and, by extension, the process that produces them, one Roberts move stands out as – well, weird. His citation of an EPA blog post discussing an entirely unrelated Supreme Court case is especially revealing – at once of carelessness and of, perhaps, a failure to comprehend the functions that the regulatory rulemaking process performs before rules reach the stage of judicial review.
EPA BLOG POST: THE MATS RULE AND MICHIGAN
In his first CPP memo Chief Justice Roberts cites an EPA blog post pubished the day after the Court handed down its decision in Michigan v EPA in 2015. The EPA blog post does not make the point Roberts asserts that it does and the underlying case offers a contrast with the CPP that impugns the memo’s overall reasoning. The history of the rule and case addressed in the blog post also nicely illustrates the critical functions of the EPA rulemaking process that seem to elude the Chief Justice.
The Supreme Court’s Michigan opinion resolved a challenge to EPA’s Mercury and Air Toxics Standards (MATS), which required sweeping pollution reductions in mercury and other heavy metals from power plants. The 2012 rule included a compliance date of April 16, 2015, set by the EPA in adherence to the Clean Air Act’s express requirement that deadlines fall three years after promulgation of rules like the MATS rule. The Michigan majority ruled on June 29, 2015 that the EPA had acted illegally in promulgating the rule. In his CPP stay memo Roberts claimed that the EPA post addressing the Michigan ruling “makes the case for irreparable harm.” He quotes the blog post’s statements that notwithstanding the Court’s decision “EPA is confident that [the rule] was still on track to reduce” the targeted pollutants since “the majority of power plants are already in compliance or well on their way to compliance.”
At a first pass, this does seem like “irreparable harm” to the power plant operators since they had already spent the money needed to comply with the rule before the Supreme Court issued its Michigan decision. Roberts’ point seems to be that the Michigan decision, which came down after the compliance date had passed, was too late to prevent the “harm” to the operators.
A closer look at the facts at the very least muddles the story Roberts purports to be telling and it’s likely that if he had taken the time to review the basic facts he might have thought better of citing the blog post in the first place. The fact that he does not seem to have done so deepens the impression that the CPP stay was a shoot first and ask questions – if they were asked at all – later.
First, immediately after the MATS rule was promulgated in 2012, several parties sought a stay from the DC Circuit alleging they would suffer irreparable harm unless the court stopped the 3-year complinace clock until it could hear their challenge to the rule and issue a decision on the merits. That court denied the stay and the compliance clock continued to tick.
There was nothing remarkable in the DC Circuit’s stay denial. Staying a rule that has undergone the extensive public and notice rulemaking process is an extraorddinary step and it’s one that the circuit court rarely takes or had taken until that point. The litigation challenging MATS on the merits proceeded, with the parties contesting virtually every pollution reduction and compliance demonstration requirement in the rule. The DC Circuit upheld the MATS rule and the Supreme Court itself denied petitions for certorari on every single issue save one. The single issue the Court agreed to hear did not involve any of the rule’s pollution reduction or compliance requirements (or any aspect of the rule that entailed spending money by the operators).
The issue the Michigan Court took up was whether the EPA was obligated to consider cost in determining wihether regulating power plant mercury emissions was “appropriate and necessary” (a term used in the Clean Air Act). Beyond ruling that the EPA had erred in failling to read the statutory language as requiring the Agency to take cost into account in making the “appropriate and necessary” finding, the Court did nothing but send the rule back to the DC Circuit leaving it to that court to determine the remedy and taking pains to leave it to the EPA to decide how to take account of costs. The DC Circuit declined again to vacate the MATS rule and simply remanded it to the EPA to incorporate cost considerations in re-doing the requisite “appropriate and necessary” finding.
Again, the Court did not even take up, let alone find fatal defects in, the MATS emission reduction and compliance requirements, all of which entailed costs for power plants. Once the Court declined to grant certiorari to address those issues, the book was closed on those issues. The Court did not even hint in its ruling that the defect it did find in EPA’s “appropriate and necessary” analysis invalidated the rule, and nowhere did it suggest that power plant operators should be relieved of any ongoing cost-incurring compliance obligations. Contrary to Roberts’ citing of the case via reference to the EPA post, there was no “harm” to states, power plants, or stakeholders before the Michigan Court that its decision was too late to “repair”.
The trajectory of the MATS rule and its litigation history were, in fact, ordinary and typical. Why Roberts cited the blog post’s report on the progress of compliance as if it were a remarkable turn of events is baffling. That parties subject to national regulations must take on compliance before the entire judicial review process is complete has long been accepted. The EPA post portentously cited by Roberts was in fact an utterly anondyne recitation of a common place reality: once a regulation has worked its way from being proposed for public comment to being finalized sources either move quickly to seeking a stay from the DC Circuit or take the steps needed to comply with the rule. Here, that’s exactly what they did, which was all that the EPA blog reported. That the Court’s decision came down a couple of months or so past the compliance date is not significant, since the question of irreparable harm itself had long since been resolved by the DC Circuit and rendered moot not by the passage of time but by the power plant operators’ standard response of initiating compliance once the DC Circuit declined to stay the rule.
Roberts’ reference to the EPA blog post doesn’t even work as a “bloody shirt”-waving tactic to incite his colleagues by way of suggesting some kind of EPA defiance of the Court’s Michigan decision. Nothing in the post even suggested that the Agency would ignore or decline to respond to the Court’s holding.
Roberts’ invocation of the MATS rule and the Michigan decision actually set up a foil to the CPP that reinforces the reality that litigation would have had time to advance to completion or close to it before states or power plants woud have had to incur significant costs. The MATS rule established requirements that power plants had to meet within three years (with easily available one-year extensions); the CPP established requirements for states to submit plans in as many as three years while power plants would face a compliance schedule beginning more than 6 years after the final CPP rule was issued. In contrast to the extremely limited opportunity the MATS rule and the Clean Air Act afforded sources to seek extensions beyond one year, the CPP offered states an almost dizzying menu of options for constructing those plans with a heavy emphasis on features that would maximize flexibity, including some that woud allow power plants to extend their effective compliance deadline past 2022 – that is, beyond a six-year time horizon.
The Chief Justice’s memos show no evidence that he gave full weight to these details beyond just noting and then dismissing them. Instead, he insisted that the states were already spending resources at significant enough levels to be suffering irreparable harm to an extent as to justify the drastic, nearly unprecedented remedy he was pushing. Among the items he cited for this assertion was a statement made by EPA Administrator Gina McCarthy that states were already working on their plans. That statement made less than two months after the CPP was published in the Federal Register again does not tell the full story Roberts thinks it does. In reality, EPA and the states had been in vigorous, nearly continuous dialogue for two years for the duration of the rulemaking process. States were already exploring potential components of their plans throughout the process, in part to inform that comments they submitted on the CPP proposal. McCarthy’s comments were as much a reflection of what the EPA observed and participated in before the final rule was issued as they were a precise report of state activity in the months following the publication of the final rule.
TRUST THE PROCESS
The Chief Justice’s citing of the EPA blog post conveys, if faintly, a note of indignation – or at least surprise or unfamiliarity as if the post revealed a process that was news to him, as if compliance with regulations well ahead of the completion of the process of judicial review was not the default outcome, and as if the DC Circuit normally stayed regulations pending litigation as a matter of course making the trajectory of MATS compliance somehow exceptional. None of this, in fact, is novel. The reason pollution sources initiate compliance when requirements are finalized and that the DC Circuit is not in the habit of staying rules pending appeal lie in the widely underappreciated strengths of the process by which regulations are developed – a process that compares favorably with the Supreme Court’s unexplained CPP stay and its increasing reliance on unexplained or scantily explained orders via the Shadow Docket.
The process EPA follows in its rulemaking is prescribed by the Administrative Procedures Act and a key section of the Clean Air Act. Frequently, the process begins with EPA staff formally or informally collecting data and other information about the pollutants and sources they intend to draft a rule for. This stage of the process often involves discussions with affected industries and outside experts. The formal process begins with a proposal, followed by a period, ranging from 45 to 90 days (at least), during which regulated sources, states, environmental and public health advocates, other stakeholders, interested experts and scientists, and members of the public may and always do submit detailed comments about the proposal. Often Agency staff meet with regulated sources and advocates to be briefed on and discuss the information they are submitting and their concerns and objections. The comments include data, engineering information, information about the operation of the regulated sources along with their business concerns, and cost analysis. They identify factors, arguments, and objections for the agency to consider, including information about the health impacts of the covered pollutants.
EPA is required as a matter of law to consider everything in the record created by this material and consider every submission with sufficient care to be able to respond to each submission with a substantive explanation of how the final rule accommodated the comment if it did, changes to the proposal incorporated in the final rule to the extent they were based on comments, and reasons for rejecting or declining to accept a comment.
More often than not EPA makes changes to its proposals that reflect comments and information submitted and writes its reponses knowing they will be scrutinized by both the public and the courts when, invariably, the final rules are challenged, usually in the DC Circuit Court of Appeals. In most cases, EPA makes significant changes to its rules after considering the comments, information, and suggestions presented to the Agency. For example, the MATS requirements include significant limits on PM 2.5. Through comments from stakeholders and the utility industry and meetings EPA staff held with utility representatives and engineering consultaants, EPA found a way to modify the ways the limits were formulated in the proposal so that the final limits achieved the same level of PM 2.5 reductions but at a much lower cost and with much greater workability.
Extensive discussions with utilities and states while the CPP proposal was pending led to an even more significant change. Outside analysts and stakeholders showed EPA that the level and timing of initial power plant reductions created an excessively abrupt and expensive compliance “cliff”. EPA captured the key ideas that emerged from those discussions and issued a supplemental proposal seeking additional public comment; ultimately, the CPP incorporated a cluster of changes to ensure a more gradual compliance ramp. Similarly, the extensive menu of options states could consider in developing their state plans offered in the final CPP grew out of written comments, meetings, and discussions in which a host of suggestions were presented. Almost all were addressed in detail in the final rule.
These examples and the history of countless EPA rulemakings illustrate the defining attributes of the rulamking process: EPA presents to the public its proposed action along with the information and analysis it used to develop it; the public, affected businesses, and communities most affected by the proposed rule engage with EPA via meetings and formal commentse submitted on the proposal providing informationa, analysis, suggestions, and objections; EPA painstakingly reviews the material and responds to it while also developing a detailed explanation of every decision it makes in finalizing the rule. Its explanations, analysis, and reasoning must be rigorous enough to stand up both to public scrutiny and review by the DC Circuit in litigation challenging the rule.
The DC Circuit’s general reluctance to stay rules reflects, among other things, simple common sense. The rulemaking process itself is a kind of litigation since comments submitted on proposals often challenge or object to the proposal or its central features including the legal arguments the EPA itself makes as part of the justification. Thus, the Agency itself will have proceeded from a record in which any of the issues that may be brought before the court will already have been contested and substantively addressed by the Agency in full public view and in anticipation that the issues and EPA’s treatment of those issues are going to be contested again before the court and adjudged by the court.
The process of judicial review includes a telling feature: the court considers only arguments that were made to the Agency via comments submitted on the proposal and scrutinizes not the merits of the rule per se but whether the Agency gave due consideration to the issues raised by the litigants during the rulemaking process and offered a reasonable response to those issues when the rule was finalized. The court is specifically not a forum for re-litigating the substance of regulations, not only for the sake of finality but also in tacit acknowledgement of the distinctive strengths of the rulemaking process as the venue for raising, addressing, and resolving the substantive issues involved – strengths that are different from and extend well beyond those of courts and judges.
This implied acknowledgement reflects the fact as well that since public participation is intrinsic to the rulemaking process and EPA must explain itself comprehensively and in complete detail, the rulemaking process comprises the cardinal democratic properties of providing the public with an efficacious voice and of accountability to that public.
Until the CPP stay, the public could take for granted most of the time that the corresponding strengths of the Supreme Court and the judiciary branch, tailored to meet the specific role of courts, but including the functions of deliberation and transparency and of accouontability to precedent and principled reasoning fully explained. Since then, with the increasing use of Shadow Docket practices that eschew those attributes the public can no longer do that.
Instead of standing side-by-side with the Court’s processes as a mutually reinforcing model of accountability, transparency, responsiveness to the public, and disciplined constraint, the regulatory rulemaking process stands in reproachful contrast to the Court’s new way of doing business, first augured by the CPP stay.