On June 2, the Obama administration's Environmental Protection Agency (EPA) released its long-awaited proposed regulation to reduce carbon dioxide (CO) emissions from existing sources in the electricity-generating sector. The regulatory (rule) proposal calls for cutting CO emissions from the power sector by 30 percent below 2005 levels by 2030. This is potentially significant, because electricity generation is responsible for about 38 percent of U.S. CO emissions (about 32 percent of U.S. greenhouse gas (GHG) emissions).
On June 18, EPA published the proposed rule in the Federal Register, initiating a 120-day public comment period. In my previous essay at this blog, I wrote about the fundamentals and the politics of this proposed rule (EPA's Proposed Greenhouse Gas Regulation: Why are Conservatives Attacking its Market-Based Options?). Today I take a look at the economics.
Cost-Effective, Perhaps — but Efficient?
The proposed rule grants freedom to implementing states to achieve their specified emissions-reduction targets in virtually any way they choose, including the use of market-based instruments (the White House has referenced cap-and-trade in this context, although somewhat obliquely as "market-based programs," and state-level carbon taxes might also be acceptable — if any states were to include them in their plans to implement the regulation). Also, the proposal allows for multi-state proposals and for states and regions to establish linkages among their state and multi-state market-based instruments. Some questions remain regarding the temporal flexibility (banking and borrowing) that the proposed rule will allow, but it's reasonable to conclude at this point that although EPA may not be guaranteeing cost-effectiveness, it is allowing for it, indeed facilitating it. As Dallas Burtraw of Resources for the Future has said, the proposed rule ought to be judged to be potentially cost-effective.
Cost-effectiveness (achieving a given target at the lowest possible aggregate cost) is one thing, but economists — and possibly some other policy wonks — may wonder if the proposal is likely to be efficient (maximizing the difference between benefits and costs). This is a much higher mountain to climb, and a particularly challenging one for a regional, national, or sub-national climate-change policy, given the global commons nature of the problem.
The Challenge of this Global Commons Problem
GHGs mix globally in the atmosphere, and so damages are spread around the world and are unaffected by the location of emissions. This means that any jurisdiction taking action — a region, a country, a state, or a city — will incur the direct costs of its actions, but the direct benefits (averted climate change) will be distributed globally. Hence, the direct climate benefits a jurisdiction reaps from its actions will inevitably be less than the costs it incurs, despite the fact that global climate benefits may be greater — possibly much greater — than global costs.
(An Aside: This presents the classic free-rider problem of this ultimate global commons problem: It is in the interest of no country to take action, but each can reap the benefits of any countries that do take action. This is why international, if not global, cooperation is essential. See the extensive work of the Harvard Project on Climate Agreements.)
On June 2, EPA released its 376-page Regulatory Impact Analysis (RIA) of the proposed "Clean Power Plan" rule, the same day it released the 645-page proposed rule itself. An RIA is essentially a benefit-cost analysis, required for significant new Federal rules by a series of Executive Orders going back to the presidency of Jimmy Carter, and reaffirmed by every president since, including most recently President Obama.
Given the fundamental economic arithmetic of a global commons problem, it would be surprising — to say the least — if EPA were to find that the expected benefits of the proposed rule would exceed its expected costs, but this is precisely what EPA has found. Indeed, its central estimate is of positive net benefits (benefits minus costs) of $67 billion annually in the year 2030 (employing a mid-range 3 percent discount rate). How can this be?
Two Answers to the Conundrum
First, EPA does not limit its estimate of climate benefits to those received by the United States (or its citizens), but uses an estimate of global climate benefits.
Second, in addition to quantifying the benefits of climate change impacts associated with CO emissions reductions, EPA quantifies and includes (the much larger) benefits of human-health impacts associated with reductions in other (correlated) air pollutants.
Of course, even if benefits exceed costs at the given level of stringency of the proposed rule, it does not mean that the rule is economically efficient, because it could be the case that benefits would exceed costs by an even greater amount with a more stringent or with a less stringent rule. However, if benefits are not greater than costs (negative net benefits), then the rule cannot possibly be efficient, so I will stick with the all-too-common Washington practice and simply ask whether the analysis indicates a winner or a loser at the proposed rule's given level of stringency. In other words, the question becomes, "Is the proposed rule welfare-enhancing (even if it is not welfare-maximizing)?"
Now, let's take a look at the numbers from these two key aspects of EPA's economic analysis and the issues surrounding the calculations.
U.S. versus Global Damages
There are surely ethical arguments (and possibly legal arguments) for employing a global damage estimate, as opposed to a U.S. damage estimate, in a benefit-cost analysis of a U.S. climate policy, but employing a global estimate is a dramatic departure from the precedent of decades of Regulatory Impact Analyses.
In a recent working paper, "Determining the Proper Scope of Climate Change Benefits," Ted Gayer, Vice President and Director of Economic Studies at the Brookings Institution, and Kip Viscusi, University Distinguished Professor of Law, Economics, and Management at Vanderbilt University, review the history of RIAs, including their virtually exclusive focus on national impacts (defined by geography or U.S. citizenship) in benefit and cost estimates of regulations.
In the context of a conventional RIA, it does seem strange — at least at first blush — to use a global measure of benefits of a U.S. regulation. If this practice were applied in a consistent manner — that is, uniformly in all RIAs — it would result in some quite bizarre findings. For example, a Federal labor policy that increases U.S. employment while cutting employment in competitor economies might be judged to have zero benefits!
Another example, this one courtesy of Tim Taylor via Ted Gayer: Under global accounting, if a domestic climate policy had the unintended consequence of causing emissions and economic leakage (through relocation of some manufacturing to other countries), that would not be considered a cost of the regulation (and with diminishing marginal utility of income, it might be counted as a benefit)!
On the other hand, a counter-argument to this line of thinking is that the usual narrow U.S.-only geographic scope of an RIA is simply not appropriate for a global commons problem. Otherwise, we would simply restate in economic terms the free-rider consequences of a global commons challenge. In other words, a domestic-only RIA of a climate policy could have the effect of "institutionalizing free riding," to quote my Harvard Kennedy School colleague, Professor Joseph Aldy. Of course, if global benefits are to be included in a regulatory assessment, it can be argued that global costs (such as leakage) should also be considered.
I leave it to legal scholars and lawyers to debate the law, and I defer to the philosophers among us to debate the ethics, but let's at least ask what the consequences would be for EPA's analysis if a U.S climate benefits number were used, rather than a global number. For this purpose, we can start with EPA's estimates (from Table ES-7 on page ES-19 and Table ES-10 on page ES-23 of its Regulatory Impact Analysis of the proposed rule) for 2030 benefits and costs, using a mid-range 3 percent real discount rate. The estimated (global) climate benefits of the rule are $31 billion.
In order to think about what the domestic climate benefits might be, we can turn to the Obama administration's original calculation of the Social Cost of Carbon in 2010, where the Interagency Working Group estimated a central global value for 2010 of $19 per ton of CO, and noted (and explained in more detail in a subsequent scholarly paper by several members of the Working Group) that U.S. benefits from reducing GHG emissions would be, on average, about 7 to 10 percent of global benefits across the scenarios analyzed with the one model that permitted such geographic disaggregation.
(The Interagency Working Group also suggested that if climate damages are simply proportional to GDP, then the U.S. share would be about 23 percent. However, given the IPCC's prediction of highly unequal geographic distribution of climate change effects worldwide, combined with the exceptionally heterogeneous nature of climate sensitivity among the world's economies, which vary from those with trivial reliance on agriculture to those dominated by their agricultural sectors, I find the argument behind this second approach unconvincing.)
Taking the midpoint of the Obama Working Group's 7–10 percent range, U.S. damages (benefits) may be estimated to be 8.5 percent of global damages, which would reduce the $31 billion reported in the new RIA to about $2.6 billion, which is considerably less than the RIA's estimated total annual compliance costs of $8.8 billion (assuming that the states facilitate cost-effective actions). This validates the intuition, explained above, that for virtually any jurisdiction, the direct climate benefits it reaps from its actions will be less than the costs it incurs (again, despite the fact that global climate benefits may be much greater than global costs).
There are plenty of caveats on both sides of this simple analysis. One of the most important is that if the proposed U.S. policy were to increase the probability of other countries taking climate policy actions (which I believe is probably the case), then the impacts on U.S. territory of such foreign policy actions would merit inclusion even in a traditional U.S.-only benefit-cost analysis. More broadly, although it has been traditional to use a U.S.-only benefits measure in RIAs, the current guidelines for carrying out these analyses from the Office of Information and Regulatory Affairs of the U.S. Office of Management and Budget (Circular A-4) requires that geographic U.S. benefit and cost estimates be provided, but also allows for the optional inclusion of global estimates.
Pending resolution (or more likely, discussion and debate) from lawyers and philosophers regarding the legal and ethical issue of employing domestic benefits versus global benefits in a climate regulation RIA, it is essential to recognize that there is an even more important factor that explains how EPA came up with estimates of significant positive net benefits (benefits exceeding costs) for the proposed rule (and would have even if a domestic climate benefits number had been employed), namely, the inclusion of (domestic) health impacts of other air pollutants, the emissions of which are correlated with those of CO.
Correlated Pollutants and Co-Benefits
The Obama administration's proposed regulation to reduce CO emissions from the electric power sector is intended to achieve its objectives through a combination of less electricity generated (compared with a business-as-usual trajectory), greater dispatch of electricity from less CO-intensive sources (natural gas, nuclear, and renewable sources, instead of coal), and more investment in low CO-intensive sources. Hence, it is anticipated that less coal will be burned than in the absence of the regulation (and more use of natural gas, nuclear, and renewable sources of electricity). This means not only less CO being emitted into the atmosphere, but also decreased emissions of correlated local air pollutants that have direct impacts on human health, including sulfur dioxide (SO), nitrogen oxides (NO), particulate matter (PM), and mercury (Hg).
It is well known that higher concentrations of these pollutants in the ambient air we breathe — particularly smaller particles of particulate matter (PM) — have very significant human health impacts in terms of increased risk of both morbidity and mortality. The numbers dwarf the climate impacts themselves. Whereas the U.S. climate change impacts of CO reductions due to the proposed rule in 2030 are probably less than $3 billion per year (see above), the health impacts (co-benefits) of reduced concentrations of correlated (non-CO) air pollutants are estimated by EPA to be some $45 billion/year (central estimate)! (By the way, I assume that the co-benefits estimated by EPA are based upon a comparison with a business-as-usual baseline that includes the effects of all existing EPA and state regulations for these same local air pollutants. If not, the RIA will need to be revised.)
The Bottom Line
The combined U.S.-only estimates of annual climate impacts of CO ($3 billion) and health impacts of correlated pollutants ($45 billion) greatly exceed the estimated regulatory compliance costs of $9 billion/year, for positive net benefits amounting to $39 billion/year in 2030. This is the key argument related to the possible economic efficiency of the proposed rule from the perspective of U.S. welfare. If EPA's global estimate of climate benefits ($31 billion/year) is employed instead, then, of course, the rule looks even better, with total annual benefits of $76 billion, leading to EPA's bottom-line estimate of positive net benefits of $67 billion per year. See the summary table below.
The Obama administration's proposed regulation of existing power-sector sources of CO has the potential to be cost-effective, and if you accept these numbers, it can also be welfare-enhancing, if not welfare-maximizing.
That said, I assume that proponents of the Obama administration's proposed rule will take this assessment of EPA's Regulatory Impact Analysis as evidence of the sensibility of the rule, and opponents of the administration's proposed actions will claim that my assessment of the RIA provides evidence of the foolishness of EPA's proposal. So it is in our pluralistic system (not to mention, in the context of the political polarization that has gripped Washington on this and so many other issues).
Benefits and Costs of EPA's Proposed Clean Power Plan Rule in 2030
(Mid-Point Estimates, Billions of Dollars)