Brian Potts says EPA existing-source performance standards (ESPS) for power plant carbon emissions won't matter much since they can't or won't be very stringent. This is partly true, if a bit overstated. You're certainly kidding yourself if you're counting on ESPS to take care of US climate policy on their own - though I know of nobody making that claim. But if ESPS are modest it will be because of conservatism at EPA (and the states), not the legal or technological limitations Potts claims.
Potts argues that the Clean Air Act itself bars EPA from enacting stringent ESPS. EPA has already done a few case-by-case reviews of new power plants (NSR) and required those plants to reduce their emissions by only a few percent. Potts argues that the statute prevents EPA from issuing performance standards that are any tighter than this. I don't think that's true. When a new plant is built, the law requires it to use "best available control technology" or BACT, which must be at least as effective at reducing emissions as the prevailing performance standards for new or existing sources under §111. (This BACT review is what EPA has already done at a few new plants). But the reverse is not true - the law doesn't require performance standards to be less stringent than all current BACT reviews. If it did, EPA would have a hard time ever strengthening new-source performance standards (NSPS), or writing new ones, as it is doing now. Once the agency does finalize NSPS, any future BACT reviews will have to be at least as strict. But I don't see how one-off BACT analysis limits EPA's authority to set sector-wide NSPS.
More importantly, the fact that BACT reviews required only small emissions reductions at new plants says little if anything about whether greater reductions are possible at existing plants, many of which are much less efficient. ESPS could be a lot less stringent than BACT and still end up requiring significant emissions cuts or efficiency improvements from a lot of the fleet. Potts seems to believe that EPA will require all plants to improve their efficiency/reduce their emissions by the same percentage. EPA might be able to take that approach, but there's little reason for it to do so. ESPS are performance standards, not improvement standards. Instead, EPA will likely set a target heat rate for all new sources (as it has already proposed doing), and a somewhat less-stringent target heat rate for existing sources (though states will be the ones actually regulating). That approach might result in negligible reductions at new plants but significant reductions (or retirements) at old or inefficient plants.
Nothing about this approach conflicts with the law's requirement that BACT be more stringent than §111 performance standards. In fact, EPA's proposed NSPS would effectively ban construction of new coal plants without carbon capture. If the final standards indeed do so, EPA will never have to worry about whether future BACT is more stringent than the NSPS since it won't be doing any coal plant BACT reviews.
Despite all this, Potts' prediction that ESPS would only lead to less than 5% emissions reductions from coal is plausible, and is in the ballpark of what researchers at RFF have found under some relatively conservative assumptions. But there are other ways for ESPS to reduce emissions, most notably by giving incentives to switch from coal to gas. Potts argues that modest ESPS aren't likely to lead to much coal retirement. That may be true if ESPS are applied to coal alone - though coal is under pressure from other regulations and low gas prices, so even a little pressure might push plants into retirement. But it's not true if ESPS allow trading between coal and gas. Recent RFF research suggests that if such trading is allowed, power sector emissions could decline by as much as 14%, with substantial switching from coal to gas (and, presumably, associated coal retirements). Average coal plant emissions would still decline by less than 5%, with the rest of the emissions reduction driven by switching.
EPA might not be so bold - legal hurdles and agency conservatism may lead it to regulate coal and gas separately. If so, Potts' suggestion that the best ESPS could do is something like 1-5% emissions reductions from coal, driven entirely by efficiency improvements, is probably about right. That doesn't get you close to the President's target of 17% economywide emissions reductions by 2020. It's not worthless either, though. Even a modest ESPS is an important part of a package of policies that, combined with market trends, could get the country to that goal.
Moreover, Potts' arguments cut both ways. If ESPS have to be minimalist, they won't achieve massive emissions reductions. But they also won't be extremely costly. And there are more (and louder) voices claiming EPA carbon policy will destroy the economy than there are saying it will solve the climate problem.