When my friends Jody Freeman and Richard Lazarus defend the legality of the EPA’s power plant rule by saying that no one would take the constitutional arguments against the rule seriously were my “name not attached to them,” they no doubt mean to be complimentary. But I take my arguments very seriously indeed and hope, by bringing them into the public forum, that I will be able to help others understand why – despite my lifelong devotion to environmental causes, my deep concern about climate change, my agreement with the need to address it urgently, and my admiration for the president whose plan to address that vital problem is at stake and for those (including Jody and Richard) who are defending that signature initiative – I regretfully feel obliged to oppose their views.

In their response to my analysis, Professors Freeman and Lazarus describe as “ridiculous” and “wholly without merit” the arguments they seem to think I am making. But it would help to be clear about what those arguments are – and what they are not.

The Remarkable Claim That There Exist Two Versions of the Relevant Statutory Provision

My central argument is that the text, context, and history of the 1990 statutory provision that EPA invokes to support the rule it seeks to impose on all fifty States does nothing of the sort but in fact destroys EPA’s claim of congressional authority. Jody and Richard do not quote the relevant text; instead they rely on the claim that there actually exist “two versions of” the relevant law, “both passed by the Congress and signed by the President.” One “version” – the “version” everyone has assumed for 25 years to be the real law at issue (based on a substantive bill originating in the House of Representatives and adopted there on May 23, 1990) – would not authorize what the EPA proposes to do. That “version,” as I’ll elaborate shortly, in fact would forbid EPA’s proposal. At the same time, my friends argue that the other “version” – which is nothing more than a clerical or “conforming” amendment adopted more than a month earlier by the Senate on April 3, 1990 – “clearly authorizes EPA’s proposal.”

Actually, even that is quite an overstatement. The Senate’s clerical amendment provided, in its entirety: “Section 111(d)(1) of the Clean Air Act is amended by striking ‘112(b)(1)(A)’ and inserting in lieu thereof ‘112(b).’” This amendment, which merely updated a statutory cross-reference in the previous version of the statute in question, hardly provides EPA the basis to remake the U.S. economy. All it did was to delete the text “(1)(A).” But I’ll proceed on the assumption (just for the sake of argument) that my colleagues would at least have a leg to stand on (even if a wobbly leg) if the Senate “version” had become the law.

More interesting by far is the argument Freeman and Lazarus make when they find me guilty of “shrug[ging] off” what they call the “truly hard legal question” of how to decide which of the two conflicting laws supposedly passed by Congress and signed by the President actually governs. To resolve what they depict as a novel whodunit, they invoke “the Chevron principle, established by the Supreme Court three decades ago, which asks simply whether the agency’s view of an ambiguous statute is ‘reasonable.’”

Wrong, wrong, and wrong again. Where to begin?

Let’s begin with the scene in which the President signed the 1990 amendment to the Clean Air Act. Although the president at the time was George H.W. Bush, one struggles to imagine even our current multitasking (not to say ambidextrous) President signing two different laws at once, leaving to EPA the extraordinary task of deciding which was the real 1990 amendment to the Clean Air Act. Perhaps the one he signed with his left hand while signing the other with his right?

Imagining that scenario is what got me started doubting the EPA’s novel argument (and Jody’s and Richard’s remarkable defense of that argument): I was fascinated by the utter implausibility of the agency’s story of why the clerical amendment deleting the text “(1)(A)” – which the agency and my colleagues implausibly say would have permitted EPA to regulate CO2 emissions from existing power plants (mostly by requiring many of them to shut down despite the billions they had invested in meeting EPA’s regulations for toxic pollutants emitted by those plants ever since 1990)—was in truth the version that ought to appear in the United States Code instead of the version that in fact has appeared there from 1990 to this day.

You’ve got to admire my colleagues’ chutzpah. They, like the EPA, are undaunted by the fact that the version you’ll find if you read the U.S. Code is precisely the substantive amendment adopted by the House on May 23, 1990, more than a month after the Senate voted on its clerical amendment. It’s the substantive amendment that my colleagues concede makes any legal defense of the EPA’s proposal vastly more difficult if not downright impossible.

Yet it’s the substantive amendment that in fact became and remains the law. How do we know that? No prizes for getting that one right: We know because the Senate itself recognized that the substantive amendment made by the House in May should prevail over the clerical amendment made by the Senate in April. To be specific, the Senate Conferees expressly stated in the Conference Report dated October 27, 1990 that they were “receding” to the House version, which is Washington-speak for saying “We’re caving. The other chamber’s language is the law.”

To be sure, both the clerical amendment and the later substantive amendment were included in the bill enacted by Congress after the Conference and signed by the President on November 15, 1990, and both appear in the Statutes at Large. But the presence of both the clerical and substantive amendments in the enacted legislation does not mean that there were “two versions” of the 1990 amendments. Rather, there is one law with provisions to be harmonized as part of Congress’s own codification process. Accordingly, the Office of Law Revision Counsel (the “Revisor”), which Congress has tasked by statute with handling the codification process, properly concluded that, once the substantive amendment was enacted, the clerical amendment was rendered moot and “could not be executed” because it referred to language that no longer existed in the amended statute. This decision was compelled both by the decision of the Senate conferees to “recede” to the House version, and by basic rules of legislative drafting, as reflected in the legislative drafting manuals established by the House and the Senate: An amendment fails to execute if a prior amendment in the same bill removes or alters the text that the subsequent amendment would amend. Congress legislates against the backdrop of that principle.

Hence, there were never “two versions” of the law, nor was there any confusion about what Congress was trying to do. To the extent there was ever a mistake, it was corrected twice – once by the congressional Conference Committee and a second time by the Revisor. Every Member of Congress receives from the Revisor a copy of the Code and its latest supplement. No evidence has been produced to suggest that any Member ever challenged the Revisor’s determination, distributed to every Member early in 1991, by raising a question of privilege or by any other means. An update completed by the Revisor and submitted to Congress in 2013 flatly rejects EPA’s current interpretation and adheres to what is the current version of the U.S. Code, with a text that specifically prohibits invoking Section 111(d) for “any air pollutant . . . emitted from a source category that is regulated under section [112] of this title.” That language indisputably includes the “source category” of existing coal-fired power plants, which since 2000 have been listed for regulation under Section 112. Case closed.

Such a situation – where a substantive amendment moots a prior clerical or conforming amendment – has occurred with great frequency in the U.S. Code. It has never in our Nation’s history resulted in the remarkable scenario in which an agency is allowed to claim that two different laws have been simultaneously enacted and that the agency has the authority to choose between them, selecting the law that might “arguably” sustain its proposed exercise of power, and tossing out the law that would not. What my colleagues claim supports EPA’s authority to remake the American energy landscape is indeed a first.

Make no mistake: The position that EPA is taking and that my colleagues support would have the effect of overriding the meticulous procedure Congress itself has put in place for dealing with clerical errors that come to the attention of congressional conferees when the time comes to reconcile the measures emerging from the House and Senate. In doing so, that position would call into question dozens and possibly hundreds of statutory changes throughout the U.S. Code, making the already challenging process of merging the work of the two legislative chambers essentially impossible.

Why The “Two Versions” Story Would Itself Render EPA’s Position Constitutionally Untenable Rather Than Rescuing It From Oblivion

When President George H.W. Bush signed the 1990 amendments into law on November 15, 1990, he noted “several provisions of the bill that” he thought “raise serious constitutional concerns” about such matters as his ability to remove members of the Chemical Safety Investigation Board and other provisions that he believed “invade the deliberative processes of the executive branch” or “impose on courts responsibilities inconsistent with their judicial function.”

Would not the President have raised even more momentous constitutional concerns had the bill placed before him by Congress genuinely left up to EPA the task of deciding which law it was that the President was in the process of signing – and thus which law he would henceforth be sworn by Article II to “faithfully execute?” Of course he would have.

Surely the President would have seen a profound constitutional problem with asking any agency to become, in every meaningful sense, a super-legislature unto itself, picking which of two statutes to make into law. Statutes leaving agencies with broad discretion to fill in the details of any particular statute have become commonplace since the mid-1930s, but this would be ridiculous: giving executive agencies or independent bodies the authority to decide for themselves which of two (why not two hundred?) alternative statutes has in fact become the law of the land would represent the paradigm case of unlawful delegation, not delegation to implement an ambiguous law but delegation to make law from scratch!

Why Leaving EPA With the Task of Harmonizing or Reconciling The Supposed “Versions” of the Law Would Still Leave EPA’s Position Indefensible

If Congress had included, either mistakenly or deliberately, both “versions” of the law in question and had expressly or impliedly directed the EPA to “reconcile” the two versions, then I assume the President’s counsel would have explained to him that the only possible reconciliation would have been for the agency to respect the prohibitions that each version of Section 111(d) purported to place on EPA. That would have meant that EPA would have to respect both (1) the prohibition against its invocation of that section to regulate any air pollutant being emitted by an existing source when that pollutant is already being regulated under Section 112 (a program for hazardous air pollutants), and (2) the prohibition against EPA’s invocation of Section 111(d) to regulate any air pollutant, including CO2, emitted by a source category (like coal-fired power plants) that is already being regulated under Section 112.

That’s deep in the weeds, I know, but the main point is extremely simple: When an agency does indeed confront two different congressional rules limiting its authority, its duty is to reconcile or harmonize them if possible, and here reconciling the two rules that the EPA and my colleagues (mistakenly) insist Congress inserted into the 1990 amendments would clearly mean enforcing both rules at once, not choosing which to enforce and which to ignore as though they were incompatible. They are clearly additive, not conflicting.

EPA’s Phony Invocation of the Chevron Principle

Beyond that, I cannot leave unaddressed the claim that EPA makes, and that my colleagues defend, that choosing the clerical amendment over the substantive amendment to Section 111(d) is nothing special but is akin to what agencies do all the time when they resolve a textual ambiguity in a single statute – a process that poses no constitutional problem entailing an insufficiently constrained delegation of authority but, on the contrary, triggers a standard judicial duty under the Chevron principle to defer to the agency’s choice in resolving ambiguities left in a statute that Congress meant to have the administering agency resolve.

I’m afraid that position entails a basic category error, confusing the process of interpreting what an ambiguous statutory command means with the process of deciding which of two distinct statutory commands one is engaged in interpreting. Stripped of irrelevant detail, the narrative that EPA sets forth– and that my colleagues echo – is that Congress tossed two versions of S.1630 into the air, packaged them in one bill for the President to sign, and then watched with detachment while the two floated to earth so that EPA could select which of the two to catch and run with. Really? No-one as familiar with the lawmaking process under our Constitution as I know Professors Freeman and Lazarus to be could expect anyone to see this account as anything but a fantasy.

The “Thirteenth Chime” Explanation That Decisively Undercuts EPA’s Legal Position

But why would such knowledgeable students of the legal process indulge in such a fantasy? There is only one possible reason: because they understand that the one law that the President signed does not sustain EPA’s claim of legislative authority to undertake the plan they are determined to defend. That’s why this whole surreal excursion into a historically revisionist account of the process of enacting the 1990 Amendments to the Clean Air Act reminds me of the proverbial 13th chime of a clock, the chime that makes one doubt all those that went before. Replacing the law that Congress actually enacted with one that it decisively did not enact in order to supply an otherwise missing source of authority for what EPA seeks to do is the only conceivable reason the defenders of EPA’s plan would go to such desperate lengths to pull a supposedly second version of the key statutory section out of the hat.

My colleagues claim to find in the legislative history, notwithstanding this strikingly clear text, evidence showing that “Congress was trying to prevent duplicative regulation of particular pollutants,” not to “exempt entire categories of industry, like power plants, from regulation under separate Clean Air Act programs,” an approach they say “makes no sense.” But there is no such evidence. I know: I’ve looked for it with care.

To say that my view makes power plants “exempt” from regulation – even from CO2 regulation – is not correct. EPA (assuming it complies with the statutory prerequisites) can still regulate CO2 emissions from power plants under Section 111(b) (which governs new and modified sources) and under the agency’s permitting (or “PSD”) program. Plus, EPA itself has touted its Section 112 rules as ways of reducing CO2 emissions, and the agency even claims a $320 million annual “co-benefit” on the basis of those reductions. All told, power plants spend billions of dollars annually on regulatory compliance. There is no “exemption” from regulation.

Moreover, my colleagues’ policy argument is properly addressed to Congress, and only to Congress. A mere agency, even one as important as EPA, isn’t constitutionally authorized to say “Never mind” when its legislative mandate doesn’t include a regulatory authority that its administrator and the agency’s supporters believe it “should” have been given. As the Supreme Court wrote just last year in UARG v. EPA, the case in which the Court first encountered greenhouse gas regulation outside the context of cars and trucks, EPA “may not rewrite clear statutory terms to suit its own sense of how the statute should operate.” That’s no mere technicality: it goes to the heart of our constitutional framework.

Besides, it is simply untenable to claim that Section 111(d) is a vital part of EPA’s arsenal, or that EPA must be allowed to invoke it even when it is already regulating the source category in question under Section 112. The bottom line is that never before has EPA attempted to adopt a Section 111(d) standard for a source category it is regulating under Section 1112. In fact, EPA has imposed a Section 111(d) standard precisely one time since the 1990 amendments, for gases arising from municipal landfills. When the Clinton administration’s EPA adopted that rule in 1995, it endorsed my understanding of the relevant statutory provisions and rejected my colleagues’. In particular, EPA noted that Section 111(d) does not permit mandates for emissions that are “emitted from a source category that is actually being regulated under section 112.” It also confirmed that the substantive House amendment, and not the clerical Senate amendment, was the governing provision.

Ten years later, EPA repeated that it “cannot” issue a mandate “under CAA section 111(d) for ‘any pollutant’ . . . that is emitted from a particular source category regulated under section 112,” so “if a source category X is ‘a source category’ regulated under section 112, EPA could not regulate” any emissions “from that source category under section 111(d).” In 2008, EPA told the D.C. Circuit that “a literal reading of this provision could bar section 111 standards for any pollutant . . . emitted from a source category that is regulated under Section 112.” That reading was expressly embraced by the D.C. Circuit in New Jersey v. EPA in 2008 and by the U.S. Supreme Court, in an important passage contained in footnote 7 of the majority opinion in  A.E.P. v. Connecticut in 2011.

Even before the 1990 amendments, a leading Senate architect of the legislation described Section 111(d) as an “obscure, never-used section of the law.” To be precise, it had been used a handful of times before 1990, to regulate three unique, localized pollutants from four source categories, such as acid mist from sulfuric acid plants. None concerned a ubiquitous, inherently benign substance like CO2 emitted from sources throughout the nation and indeed the world and causing harmful environmental effects only in its aggregation throughout the earth’s atmosphere with all the other sources of the same substance. None involved an attempt to revolutionize the U.S. energy sector, as EPA’s latest rule seeks to do, and none required States to coordinate their energy and emissions policies with one another, linked as they are through the electric distribution grid. In 1990, an EPA official testified to Congress that imposing double regulation on existing sources, even for different pollutants, would be “ridiculous.” Both the language of Section 111(d) and the hearings leading to its revision in 1990 leave no doubt that the provision had a vastly less ambitious purpose than the one to which EPA would put it today.

The Illusory Claim That EPA’s Proposal Leaves States Free to Choose Among Numerous Means to Achieve EPA’s Mandated Goal

Importantly, my two colleagues downplay the ambitious character of the way EPA seeks to deploy that obscure section when they say that it leaves States “in full command of their energy supply, just as before.” According to my colleagues, the new rule merely sets “carbon intensity targets for each state, which they can achieve using whatever measures they prefer, including by substituting natural gas for coal, using more renewable energy, and investing in energy efficiency.” They tout EPA’s sensitivity and flexibility in taking into account the different situations in which the several States find themselves, and they applaud EPA’s setting lower targets for “states that depend heavily on coal . . . than [for] states with a cleaner energy supply.” But, with all respect to my colleagues, that flexibility is an illusion. There is no flexibility in the rigid numerical emissions limits EPA has already set for each State. And those limits in turn dictate the energy mix for each State, requiring the shut-down of many coal-fired power plants with a shift to natural gas, just as the Obama administration has at times candidly described the limits as designed to do. (Secretary of State John Kerry on U.S. policy regarding coal-fired power plants: “We’re going to take a bunch of them out of commission.”) There is no other way for most States to meet EPA’s preset targets, which is why EPA can confidently predict that by 2020 its rule will cause the retirement of 46 to 49 gigawatts of coal-fired capacity and a 25 to 27% cut in coal production.

Montana’s PSC Commissioner, to name one of many who made essentially the same observation, testified before the House last September that the “much heralded flexibility” is “meaningless” in light of the “underlying . . . inflexible” emissions target set by EPA. Seventeen State Attorneys General filed comments with EPA last November describing how EPA’s plan would undercut “state authority” by forcing many States to enact “demand-side control measures” to “reduce electricity consumption or increase energy efficiency.” And they stressed how EPA’s theory would enable the agency to “require states to mandate that consumers dim their lights on alternate days, limit home builders to constructing only two-story buildings, or shutter public schools during periods of peak energy usage.” They objected that “EPA’s approach converts the obscure, little-used Section 111(d) into a general enabling act, giving EPA power over the entire grid from generation to light switch.”

Even though Section 111 as written provides only for standards that regulate the emissions performance of individual stationary sources, EPA is invoking 111(d) to force States to regulate the entire network of electricity generation, distribution, and consumption – everything from the power to run iPhones to the power to use electric toothbrushes – much as would be the case if EPA were to invoke its separate statutory authority to regulate the tailpipe emissions (including greenhouse gases like CO2) from cars and trucks in order to regulate how often people could drive to work rather than work at home via the internet, or whether they could use cars at all in lieu of public transportation.

My objection to the lack of any limiting principle underlying EPA’s assertion of power isn’t answered by saying that EPA has yet to order any particular State to regulate the electricity consumption of any specific group of consumers and has not expressly directed any State to enact a law putting particular coal-fired power plants out of business or specifying which alternative sources of electrical energy the State must opt to rely upon. The threat is real. It’s true, as Jody Freeman and Richard Lazarus say, that States are left to their own devices in deciding precisely which drastic measures to take internally in order to meet EPA’s mandatory targets with plans they are required to put in place by 2016. But what my colleagues fail to mention is the fully settled constitutional principle, reaffirmed by the Supreme Court’s 7-2 ruling about Medicaid in the 2012 decision in NFIB v. Sebelius (the decision upholding the rest of the Affordable Care Act by a single vote), that the Tenth Amendment protects States from being commanded to enact (or coerced into enacting) particular laws at the behest of the Federal Government.

As my colleagues point out, the EPA has yet to announce exactly which measures it would use to penalize noncomplying States. The agency’s regulations point to numerous possibilities, from cutting off a State’s highway funds to yanking the federal funds a State would otherwise have received to help it cope with more conventional forms of air pollution. But the uncertainty about which precise penalties might be invoked misses the point. In the leading Supreme Court cases holding that the Federal Government may not command States to do its bidding – either by enacting laws of their own meeting federal specifications, or by implementing laws enacted by Congress in ways that meet with federal approval – it was likewise unclear just what fate would befall those States that opted to stare down the Feds. The point stressed by the Supreme Court was that commanding state action in such ways in itself exceeds federal authority under our form of government.

My colleagues also emphasize that the EPA has refrained from specifying precisely what laws any given State must enact in order to comply with its command but has instead left them with some choices in this regard. True enough, but that provides little solace and no constitutional answer. The possibility of (a Hobson’s) “choice” for States is constitutionally irrelevant if none of the laws is a law that the Federal Government may command a State to enact. That was the precise holding of the Supreme Court’s 1992 radioactive waste decision, New York v. United States, a case in which Congress had given New York a choice between enacting controls over in-state radioactive waste that met federal specifications and taking title to all the radioactive waste generated within New York’s borders. Because Congress could command neither of those steps, the Court held, it could not command New York to choose between them. A robber who says, “Your money or your life,” can’t eliminate the coercion by saying, “And you can pay me in cash, or credit, or bitcoin.”

In fact, Justice Souter, who joined the Court’s decision in New York v. United States but dissented five years later in Printz v. United States (the decision invalidating the Brady Act’s requirement that each municipality’s Chief Law Enforcement Officer conduct a background check on the buyer as a precondition of any firearms sale within the State), made the point in his Printz dissent that it’s an even deeper usurpation of state sovereignty for the Federal Government to compel a State to engage its sovereign lawmaking machinery than it is for the Federal Government to require a State’s executive branch (or its judiciary) to implement a federally designated legal regime. After all, requiring a State’s executive branch to implement federal law is close to Congress’s well established power to pass laws that state courts are bound to enforce.

The Even More Illusory Claim That States Remain Free to Gamble by Sitting Back and Letting EPA Directly Regulate Their Energy Economies

Professors Lazarus and Freeman dismiss my “colorful suggestion” that EPA’s proposal confronts States with a “gun to the head” because they say “nothing in the proposed climate rule ‘commandeers’ state institutions for federal purposes,” which they concede “the Constitution forbids.” Their defense of being so dismissive of my anti-commandeering objection is that, if “the States choose not to act, then the responsibility falls on the federal government to regulate industry itself.” So no big deal, they seem to imply. Yet the fact remains that EPA’s proposed rule would require the States, despite occasional extensions of deadlines, to come up with their package of EPA-approved laws well before learning anything specific or concrete about what federal regulations might confront them; what those regulations might do to the States’ people and their businesses should the States decline to perform as EPA instructs them to perform; or how EPA could responsibly implement so ambitious a nationwide regulatory program under its own steam. And unless the States move quickly to draft and decide how to implement packages of laws sufficient to meet EPA’s rigid CO2 targets, it will be too late for them realistically to meet EPA’s timetable.

All of that is strongly reminiscent of the possible “death spiral” that Justice Kennedy recognized the health insurance exchanges operating in several dozen States might face if the Court were to endorse the challengers’ interpretation of Obamacare in the pending case of King v. Burwell, argued this March 4. There, too, it’s unclear just what the Federal Government’s backup plan might be, but that very uncertainty was part of what led Justice Kennedy to imply in oral argument that principles of state sovereignty push strongly against adopting the challengers’ interpretation and pulling the rug out from under those States that failed to follow the Federal Government’s preferred course and instead set up their own insurance exchanges. I would thus repeat my “gun to the head” metaphor and up the ante: It’s a form of Russian Roulette that the EPA – and, I might add, Majority Leader McConnell in his reckless suggestion that States defy EPA’s directive to enact plans that will meet the agency’s CO2 targets – are daring States to play with their consumers and economies when both EPA and Senator McConnell, each for very different reasons, dangle before them the prospect of thumbing their noses at the agency’s demands and saying, “Bring it on.”

I’ve been heartened to see how many of those who believe (as I do) that the Obama administration is acting within the law in implementing the Affordable Care Act by providing tax credits to purchasers of insurance on all fifty exchanges, including those set up by HHS on behalf of the States that accepted the federal invitation to operate exchanges for them, reacted favorably to Justice Kennedy’s suggestion at the oral argument that interpreting the Act to deny HHS that authority would undercut the sovereignty of States accepting that invitation. His reasoning was that this interpretation would confront States with uncertain but probably adverse consequences, with the result that they would either be pressured into going along with the Federal Government’s preference that States set up their own exchanges or be tricked by discovering, after they rejected the federal invitation, that their exchanges might confront a “death spiral” unless the Federal Government took some as yet undefined backup step. Yet many of those same people casually dismiss the pressure that EPA’s interpretation of the Clean Air Act enables it to exert on States that decline to promulgate their own CO2 plans in compliance with EPA’s requirements, saying that it’s no big deal because, should they refuse, the Federal Government will take some equally undefined backup step. To me, that smacks of fair weather federalism, and I want no part of it.

Misunderstanding the Fifth Amendment Objection to This Unique Exercise of Regulatory Power

Finally, my colleagues misunderstand my Fifth Amendment argument when they attribute to me the view that the EPA’s rule unconstitutionally takes the industry’s private property “because government regulation of power plant pollution has not covered greenhouse gas emissions until now.” They describe my reading of the Constitution as a “novel” one that entails the “ridiculous” implication that “the coal industry, and the power plants that burn their coal, possess an absolute constitutional property right to continue to emit greenhouse gases in perpetuity.” But that’s a mischaracterization of my position and not a view I have ever endorsed.

I have argued, much more modestly, that there is a serious Fifth Amendment problem, one sounding as much in Due Process as in the Takings Clause, in the bait-and-switch the EPA would have the Federal Government undertake without just compensation. After requiring coal-fired power plants to install the very costly “Maximum Control Technology” that the Clean Air Act requires under Section 112, the agency turns around and now tells the States to take actions that would force those very same power sources to shut down or significantly curtail their coal-based operations, essentially stranding the billions of dollars that EPA has required them to invest. When EPA initially promised confidential treatment to pesticide makers who submitted proprietary data in their registration applications and then reversed course and publicly disclosed the data, the Supreme Court had no trouble concluding that the manufacturers could sue for a compensable taking. Ruckelshaus v. Monsanto (1984). So too, when federal regulators encouraged banks to take over failing savings and loan associations by promising that they could take advantage of special accounting treatment and then later disallowed that accounting treatment, the Court held that the banks could sue for breach of contract. United States v. Winstar Corp. (1996). This case isn’t exactly like those, but the resemblance is uncanny.

Unlike conventional pollutants that are harmful in themselves and that government has every right to regulate even to the point of driving some emitters of such pollutants out of business, atmospheric CO2 is not such a pollutant. Nobody suggests that it is toxic or dangerous. We are all CO2 emitters, and atmospheric CO2 is the intermingled result of all human activity. As Justice Ginsburg wrote for the Court in AEP v. Connecticut four years ago, “[W]e each emit carbon dioxide merely by breathing.” When an agency targets for extinction – let’s not beat around the bush, that’s what even Secretary Kerry has conceded the administration has in store for many old coal companies – a narrow set of CO2 emitters and thereby imposes costs that ought to be borne equitably by everyone on a particular group of companies after requiring those same companies to invest massive amounts of money in reducing their non-CO2 pollutants over the course of a quarter century, then the principles underlying the Fifth Amendment’s Due Process and Takings Clauses support a duty to compensate those thereby singled out. Such compensation would, of course, have to come from American taxpayers. And that in turn implies that only Congress, which alone possesses the power of the purse, may authorize such a course of action, just as Justice Douglas argued in his concurring opinion in the Steel Seizure case in 1952. Absent clear congressional authorization, such action violates the Constitution because it would indirectly trigger either “taxation without representation” or confiscation without compensation. That’s the parallel with President Truman’s seizure of the steel mills during the Korean War that Freeman and Lazarus find it so far-fetched for me to have mentioned in my testimony.

I know about the vagaries of Fifth Amendment analysis. It may be less of a slam-dunk than are my structural objections invoking the anti-commandeering principle and the separation of powers. But it remains the case that stretching the Clean Air Act to sustain the EPA’s proposed exercise of power would run right up against the Fifth Amendment and would thereby require any reviewing court to resolve yet another tough constitutional problem.

The Decisive Role of the Principle of Constitutional Avoidance

That’s where a principle that my colleagues simply ignore comes into play: it’s the well-established principle of constitutional avoidance – of interpreting even ambiguous statutes in a way that doesn’t require us to face and resolve difficult constitutional issues unless doing so is unavoidable. That principle has been repeatedly held by the Supreme Court to trump even the Chevron principle of deference to administrative agencies in interpreting ambiguous laws, and it’s a principle that leaves me convinced that the Fifth and Tenth Amendment problems that EPA’s proposal raises would clearly warrant reading the Clean Air Act more narrowly than EPA reads it even if the Act could plausibly be read to authorize what EPA seeks to do – which I don’t believe it can be– and even if the constitutional objections, once faced, might be overcome – which I don’t believe they could be. After all, if one had to resolve the constitutional doubts that a contested reading of a federal law would raise in order to invoke the avoidance principle, it would be a pretty pointless principle, because one would need to decide the very constitutional questions the principle was supposedly designed to help one avoid.

One big difference between this situation and that of the Truman Steel Seizure is worth noting. At least there the President believed – and on the basis of fairly solid though not necessarily conclusive evidence – that the government’s takeover of the steel industry was needed to keep our troops in Korea properly supplied with military materiel. Here, EPA won’t even officially claim that its proposal will have any effect on global temperatures or sea level. Using EPA’s approach in prior rulemakings, its figures indicate that, assuming its proposal were perfectly implemented and not simply offset by increased emissions abroad, the effect by 2100 would be to slow sea level rise by approximately 0.03 mm (the width of several sheets of paper) and to reduce the rise in global mean temperature by under 1/100 of a single degree centigrade. Those are EPA’s figures, not mine and not those of the businesses that are challenging the EPA’s authority.

The Reasons to Worry About the Constitution and the Rule of Law Even When Confronting Problems as Massive as Climate Change

To be sure, those who believe, as I do, that the global challenge of climate change is one we need to confront in every responsible way available to us, might be tempted to take just about any first step toward meeting that challenge. But this first step? It’s a first step that, by the account of those who urge it, barely gets us anywhere – and does so by putting much of our legal framework at risk, not to mention imposing serious risks of electricity blackouts and stranded workers and significant harm to whole sectors of our national economy. No, that’s not something I can just sit by and remain silent about. An old and wise maxim has it that when one needs to cross a chasm, leaping halfway (or, as in this case, a tiny fraction of the way) is not likely to represent any progress at all. Having incurred the harms that follow hitting the bottom of the chasm, one might be hard-pressed to climb back up and try again.

So too here. When I make such a fuss about the rule of law and the importance of obeying the Constitution in the means we choose to approach even this massive problem, it’s not because I underestimate the problem; it’s because I deeply believe that the solution we try shouldn’t be one that tramples on our constitutional system. Difficult as more rational alternatives might be to pursue, such alternatives – including investing in technologically advanced renewable sources of energy and carbon capture; exploring national legislation to address climate change; or paying people and businesses, both here and around the globe, not to engage in so much deforestation – make much more sense to me than tinkering with the carbon emissions of a selected group of power companies that emit just the barest fraction of the CO2 that is a source of worry. That’s why I have invoked such vivid metaphors as the one I used when I told Congress on March 17 that burning the Constitution must not become part of our national energy policy.

I knew when I undertook this challenge that many of my closest friends, and many with whom I have been in the trenches fighting for environmental sanity ever since teaching what I believe was the first environmental law course in the Nation, would take me to task for saying what I have said about this matter.  So be it.

I close by expressing my genuine gratitude that my good friends and esteemed colleagues Jody Freeman and Richard Lazarus have helped elevate this discussion above the name-calling level with which my inbox has been filled in recent days. I disagree with them strongly but respect their views and appreciate their willingness to engage mine in a substantive way.


Laurence H. Tribe, the Carl M. Loeb University Professor and a Professor of Constitutional Law at Harvard Law School, served as the first Senior Counselor for Access to Justice in 2009-10, taught at Harvard in 1969 the first course in environmental law offered in any American law school, has written over 115 books and legal articles, has represented clients in more than 50 Supreme Court cases, and has argued more than 35 cases in that Court, winning such cases as the one establishing a State’s authority to impose a moratorium on nuclear power plants until the problem of safely disposing of the radioactive waste has been solved.