“On August 3, President Obama and the EPA announced the Clean Power Plan – a historic and important step in reducing carbon pollution from power plants that takes real action on climate change.” So begins the Environmental Protection Agency’s homage to the President and itself. The harder question is whether it is true. On this point, there is a sharp division of opinion between the traditional supporters and traditional detractors of these sorts of measures, with few (if anyone) occupying a middle ground that finds some merit but expresses real concern over the structure and function of the plan. Nonetheless, it is better to back off for the moment from extravagant claims that the end is near if we don’t (or do) embrace this particular plan.

Let’s put aside the EPA’s shaky legal authority and concentrate on the plan itself. A sensible approach divides the regulatory inquiry into two halves. The first asks about the best institutional framework to regulate greenhouse gases (GHGs), most notably carbon dioxide. The second asks how to assess, on empirical grounds, the severity of the carbon dioxide problem that the EPA purports to tackle. The EPA falls short on both counts. I shall take them up in order.

The first point to note about the EPA’s clean coal initiative is that, given its inability to secure any congressional action on the subject, the agency is working solely within the existing statutory framework. That is a big mistake from the get-go. The pollution control scheme put into place under the Clean Air Act of 1970 (CAA), as modified by the Clean Air Act Amendments of 1990, is the wrong way to deal with any form of pollution.

The basic conceit of that statute calls for a division of control over pollution between the national and state governments, all under the banner of federalism. No one argues that pollution should be ignored, so the question is how best to combat it. The usual private law remedies allow individuals to collect damages for their private losses and obtain injunctions against future harm. But the diffuse nature of most pollution makes private actions unwieldy. The correct legislative response is to cut down on the enforcement costs by having government agents use fines or taxes on the one hand and explicit prohibitions on emissions levels on the other to control the loss.

The CAA founders in making the transition from private to public enforcement. Instead of imposing an output regime that taxes pollutants and sets emissions levels, it makes two key mistakes.

First, it leaves too much to the imagination of the states in the imposition of the statutory scheme. The key device is the creation of the National Ambient Air Quality Standard (NAAQS), which sets aggregate levels of pollution for a given region, allowing each state through its state implementation plan (SIP) to take the first crack at assigning burdens to individual firms. The SIP is in turn subject to override by a federal implementation plan (FIP) if SIP falls short. Second, the SIPs and FIPs regulate inputs by seeking to determine not pollution levels, but “the best system of emissions reduction” (BSER).

The combination of these two decisions has severely retarded environmental protection. The NAAQS system gives local governments too much discretion in allocating liabilities, so that each polluter takes strategic measures to foist the burdens off on someone else, thereby turning a technical decision into a political one. The BSER system offers no sensible answer to the question of whether some additional precautions are worth their salt. In principle, these decisions should be made at the margin, until the last unit spent on precaution yields a comparable health and safety benefit. But just ask this question: if we can stop 90 percent of pollution at $100, and 99 percent at $1,000, and 99.9 percent at $1 million, are we committed to the last as the best system of pollution control? Does it depend on the size of the firm? Its level of profit and loss? Or what?

One unfortunate consequence of this system is that the gaming has led to the creation of stringent standards for new source pollution standards (NSPS), and consequently far more lenient standards for the modification of existing facilities — the effect of which is to retard the introduction of new and efficient facilities as old ones are patched up year after year. The point is especially important with respect to coal plants, many of which should have turned over years ago. The problem today is that the EPA has to play catch-up for the old plants, which makes it look as if there is a concerted “war on coal” in light of these overdue corrective measures.

Ironically, most of the needed changes for coal plants are important not because of carbon dioxide, but because of all the other acknowledged pollutants emitted by coal (and, in lesser amounts, by natural gas). The EPA therefore is right to attack this issue, but it has picked a ham-handed way of making the change. The first point is that it did not confess error, go to Congress, and ask for a fundamental change in system design to fix up the current problem. Quite the opposite: this plan doubles down on the current system. Here is one key passage (in a 1,560 page report) that shows the basic errors:

In this final rule, the EPA is establishing a CO2 emission performance rate for each of two subcategories of fossil fuel-fired [Electric Generating Units] EGUs — fossil fuel-fired electric steam generating units and stationary combustion turbines – that expresses the “best system of emissions reduction… adequately demonstrated” (BSER) for CO2 from the power sector. The EPA is also establishing state-specific rate-based and mass-based goals that reflect the subcategory-specific CO2 emission performance rates and each state’s mix of affected EGUs.

The twin vices are these. First, the BSER system works no better here than it does for any other form of pollution. There is no reason to make technology-specific rules when a uniform tax applied to both will allow parties to make the needed adjustments to less dangerous sources. The EPA takes pride in the element of “facilitating” choices at the state level. But there is absolutely no need for the EPA to subsidize wind and solar energy in the belief that these are superior to traditional forms of production. They should pay their own way and be subject to taxation and regulation for the pollution they create anywhere in their production cycle — no exemptions for dead birds, for example.

More concretely, let the needed choices take place without the subsidies and without the EPA’s targets, all of which are necessarily based on little concrete information about the proper mix of different technologies that will emerge in the 15-plus years in which this plan is supposed to stay in effect. The great advantage of the tax system (or an equivalent output system) is that it does not require or invite fundamental regime changes in regulation with each advance in technology. The key issue is what choices and why. Any robber gives you a choice between your money or your life. The real question is whether he can force you into that choice. The EPA is entitled to make polluters pay. It is not entitled to run subsidy systems that introduce yet another level of administrative confusion and political intrigue.

The situation gets no better when the EPA announces that it will set individual goals for each state. Why and how should this be done, especially for carbon dioxide which has the same effect wherever it is released globally? State quotas invite political retribution against those states that put up a fight against the EPA and gentle treatment for those that cooperate. That is no way to run a business when other methods for controlling pollution are available.

These structural concerns are amplified by the dubious empirics that underline this particular debate. The first question is: do we think that carbon dioxide is a pollutant? The answer to that question is complicated. No one wants to reduce levels of carbon dioxide to zero. Unlike sulfur dioxide, carbon dioxide is essential to life. So the key question is “how much?”

When the United States Supreme Court decided that the CAA meant to cover pollution in Massachusetts v. EPA, it indulged in some fancy footwork to reach that result. But it did so confident that Al Gore’s “inconvenient truth” lay less than a decade away. The evidence has not quite turned out that way, as global temperatures have remained essentially flat. It is risky business for anyone to extrapolate with confidence how the past relates to the future, but reading the EPA report one gets a sinking feeling that the weakness of the projections of a decade past have not altered its estimate of the risk one whit.

This matters because the correct way to set the tax is to take the best estimate of the severity of the risk at hand. If the short-term dangers seem weaker, then the plan should ease off on the aggregate level of restrictions that it imposes. The hard question in many cases is whether to invest in this technology today or to rely on general improvements to deal with the problem going forward. No lawyer should ever presume to have independent knowledge of the best estimates. But it is proper for them — and for ordinary citizens — to ask the EPA to deal responsibly with the critics, such as Judith Curry, who take the opposite position, stressing the uncertainty of existing models and the importance of factoring in natural variations as a source of climate change.

Surface temperatures have been constant for 16 years in the face of major increases in carbon dioxide concentrations. It will not do for the EPA to announce that everyone agrees that greenhouse gases (GHGs) contribute to global warming. The question has to be asked in a more fine-grained fashion. What level of GHGs will produce what level of temperature increase, which in turn will result in what level of dislocation? Carbon monoxide and carbon dioxide differ chemically by only one molecule of oxygen. But in every other way they are worlds apart. A little humility can go a long way. The EPA should take that simple reality into account before it goes into regulatory overdrive based on insufficient information about the role of carbon dioxide in climate change.