Thread by @bprest

New @rff working paper w/ @AlanKrupnick! Have you heard of the ~$1 billion federal subsidy for burning coal? It’s a $7/ton tax credit for the use of “refined coal”, which is normal coal that has been sprayed with chemicals meant to make it burn cleaner. But does it work? 1/22

First, some background: refined coal barely existed in the U.S. 10 years ago, but now it makes up more than 20% of power sector coal & growing. At least 140 million tons of this stuff was burned in 2018 (see EIA Form 923). $7/ton * 140 mln tons = ~$1 billion in tax credits! 2/22

Now about the tax law: the statute requires that, to be eligible for the $7/ton credit, companies must demonstrate that the refining process reduces NOx emission rates by 20% and SO2 or mercury rates by 40%. Less pollution sounds great! But… 3/22

…the IRS lets companies demonstrate compliance through laboratory tests. And operating conditions in the lab can differ significantly from actual operations at the power plant! So we might not actually get the lab-shown reductions in practice. 4/22

There are many reasons for this: (i) lab tests use a small boiler run under a fixed set of operational parameters, whereas conditions at plants can vary a lot from hour to hour, 5/22

(ii) refiners are often outside tax equity investors—not the plant operators—and may not be involved or have a say in how the plant is operated. This can negate any emission reductions from refined coal. 6/22

(iii) the chemical process by which refined coal is supposed to work requires certain types of pollution control technologies to already be installed, and not all plants have them. But language describing eligibility in the tax law doesn’t differentiate along these lines. 7/22

So how does it look in the field? These graphs show two plants’ emissions rates just before/after they switched to refined coal. For one plant, perhaps we see the needed 20% reductions in NOx but not the 40% on SO2 or mercury. For the other, we see no obvious reductions. 8/22

These are anecdotes, but we do something more statistically rigorous and comprehensive using a data set on almost all coal in the U.S. power sector. We estimate reductions in emissions attributable to refined coal, accounting for installed pollution control technology. 9/22

The results: negligible reductions in SO2 emission rates, and the reductions for NOx and mercury are roughly half as large as the 20%/40% targets require. In addition, the mercury reductions only appear at plants with certain kinds of pollution controls installed. 10/22

These are averages, but we also do a boiler-level analysis where possible. We find that some might be achieving the 20% NOx reductions, but we can only find one boiler in the country achieving the 40% reduction in SO2 and one other achieving the 40% reduction in mercury. 11/22

Further, neither of those two boilers that got the 40% reduction also got the 20% NOx reductions! In other words, we can’t find any evidence that any individual boiler is achieving the joint 20%/40% reductions required by the tax law. 12/22

So the reductions in the field fall short of the reductions required by the tax law. But at least they’re getting some reductions! Were the 20%/40% targets were just overly ambitious, but the more modest reductions actually achieved still justify the costs? In short: no. 13/22

We estimate benefits from reduced premature mortality around $500 million, compared to ~$1 billion in costs. So it doesn’t pass a cost-benefit test based on observed emission reductions. Further, the $500 million in benefits is probably overstated for a number of reasons: 14/22

(i) We haven’t accounted for the negative impacts on water quality. The chemicals used in refined coal can end up in drinking water, which combines with chlorine during water treatment, resulting in the formation of carcinogens called disinfection byproducts. 15/22

(ii) The “waterbed effect”. Since many refined coal plants are covered by NOx or SO2 cap-and-trade programs, reductions at such a plant may simply free up emission permits that they can sell to another plant, who can use it to increase emissions. No net change in emissions. 16/22

(iii) We haven’t accounted for the effect on coal use. But if you subsidize something you’ll get more of it. The subsidy has likely increased coal use, both by increasing its dispatch and delaying coal retirement. This effect would result in more emissions, not less. 17/22

Even though we are likely overstating the benefits, the benefits still fall far short of the costs. That said, if refined coal did actually reduce SO2 emissions by 40%, that would be an amazing achievement and could merit the subsidy (see graph in tweet 14 above). 18/22

As it turns out, there is an immediate impetus to reconsider the policy. It was last changed in the 2010 extension of the Bush Tax cuts (and the 2008 TARP bill before that). As a result, it will effectively expire in 2021, unless Congress affirmatively votes to extend it. 19/22

Speaking of which, just last month, two new bills were introduced in the Senate to extend the credit. This creates a great opportunity to re-evaluate it.
congress.gov/bill/116th-con…
congress.gov/bill/116th-con…
20/22

One option would be to change the enforcement rules to require field tests, meaning companies would qualify for the credit if and only if they can demonstrate the 20% and 40% reductions at the plant (not lab). This could be an improvement (although see caveats above). 21/22

But since plants aren't getting the legally stipulated reductions in practice, the policy fails a cost benefit test, so letting it expire would be better, from a social welfare perspective, than continuing it.

Now here's a link to the full paper: rff.org/publications/r…

22/22

Hey, @electronecon @GavinBade, you guys might be interested in this. ☝️

#EnergyTwitter #RefinedCoal

And if you want to hear more, check out this podcast where I talk about it with @rff's @DanielRaimi.

resourcesmag.org/resources-radi…