As we have previously described in Part I and Part II of our analysis, the CPP is essentially a national mass-based cap & trade system, with EPA assigning emissions limits (X million tons of CO2) to each state based (somewhat) on the historic (2010- 2012) emissions of all the fossil fuel plants within its borders. In turn, each state is responsible for ensuring that its sources meet the state limit, which take effect in 2022, and decline until reaching the final limits in 2030.
States have three way they can do this. The first two distribute to these sources emissions allowances equal to the number of tons of CO2 in the state cap; at the end of the year, each source turns in one allowance for each ton emitted (with significant penalties for any source with fewer allowances than tons emitted). The difference between these two options is whether the state distributes the allowances for free or auctions them off. The third option is a carbon tax.
Distributing the allowances for free imposes the fewest immediate costs on the ratepayers, as they will only have to pay for reducing emissions down to the level of allowances (the “abatement costs”). But free distribution has two major problems that will inevitably increase ratepayer cost beyond any of the alternatives.
First is the allocation swamp: How does the state decide how many allowances to give to each source? Because EPA calculated how many tons a state could emit based on the emissions of its coal and gas plants in 2010-2012, the intuitive answer is to simply distribute the allowances in proportion to those emissions: if the 2022 state limit is 90% of the tons emitted in 2010-2012, then each source gets allowances equal to 90% of its historic emissions.
The problem is that the CPP does not begin to require compliance until 2022, and trying to distribute allowances based solely on plant emissions a decade earlier will be a nightmare. Some plants may have increased their emissions, some will have decreased theirs, some plants will only have come online after the 2010-2012 period (any plant that began construction before 2014 falls under this rule), and some will have retired.
Take just that last situation: what happens to what would have been “their” allowances? Easy, say some folks: just allocate those pro rata among the remaining sources. If plant X, one of 5 in the state, would have been entitled to 100 tons of allowances in 2022 but retired in 2019, then just give those 100 tons to the remaining four plants (A, B, C, and D) in proportion to their allocations.
But not so fast, says plant A: I owned plant X, and so I should get all of those allowances, or at least a larger share of them than B, C and D. Whoa, says Plant B: Because I’ve had to meet increased demand for my power, my emissions have gone up since 2012, and so I should get a larger share. Hey, says Plant C: I’ve been really cleaning up my act, and have cut my emissions since 2012, so shouldn’t I be rewarded with some of those allowances? I can then sell them – either here or out-of-state – to help pay for the costs I incurred in cutting those emissions. What about me, says Plant D – I came online in 2013 and thus have no slice of a purely historical allocation, and so I need X’s allocation more than any of the others.
The only thing we can be absolutely certain of is that the ensuing allocation process will be based on neither environmental nor market considerations. As the EU experience shows, legislatures and governors will create a process that benefits favored constituencies and will often make decisions based on purely political considerations: investments will be made not according to whether it is economically rational to do so but rather to satisfy key constituencies, plant closings will be determined not by whether it makes sense to do so in terms of meeting the state target but rather whether the politicians can afford the job losses in that locale, etc. Thus, this one simple consequence means that these “free” allowances will ultimately cost ratepayers quite a bit.
Rent-seeking issue is only the first problem, because free allocation does not eliminate the market in allowances. Unless the state gets it exactly correct, then by definition, every plant will wind up with either more or fewer allowances than it needs, and will become either buyers or sellers in the allowance marketplace. In turn, those allowance prices will have a significant impact on the marginal electricity price, which is a critical factor in PUC rate-setting decisions. Given this completely wild card scenario, we don’t envy PUCs trying to figure out in advance what the “right” price is that plants can charge, and avoiding either bankrupting the generators or giving them windfall profits. (And if we had to guess, we would predict the outcome will be better for generators than for ratepayers.)
Auctioning allowances is immediately more expensive for the ratepayers. They have to pay for the abatement costs reducing emissions down to the level of allowances, and then they have to pay the allowance cost for every remaining ton that gets emitted. The state, however, gets a new source of revenue to use in any way it sees fit; it can send it all back to the ratepayers (so that they wind up in the same situation as if the allowances were distributed for free), use it to cut taxes, or pocket it and see if no one notices.
But auctions, too, have their problems. First, the variable allowance price makes it much harder for utilities to plan investment decisions. There is one compliance strategy if allowances are $5/ton, and a completely different one if they are $10/ton, but you don’t know what the price will be ahead of time. Utilities will make guesses, probably based on the highest maximum allowance price they foresee (or can convince the PUC is plausible), and of course, will regularly get them wrong. And the ratepayers will pick up the bill – twice, since they will first pay for the investments and then, as taxpayers, get less back from the state’s lower allowance auction revenues.
Second, an allowance market worth billions of dollars will attract enormous amounts of speculation and – surprise – manipulation. EPA does not envisage any limits on who can buy allowances (in part because of the need for liquidity in the market), which is an invitation for people to get in whose only goal is to trade allowances for profit, not to generate power. Both speculation and manipulation increase price volatility, which exacerbates the underlying problem of variable allowance prices . . . to say nothing of creating the risk to ratepayers of paying far more for allowances than they would otherwise be worth.
But perhaps the biggest problem with auctioning allowances is that if the state has comparably low compliance costs, then the cost of those allowances will be bid up by buyers from states with higher compliance costs. If the plants in State X could meet EPA’s limit with $5 allowances, that is what the price would be if only those plants were allowed to bid on them. But EPA’s system does not work that way. The auctions are open to anyone, and the allowances are freely transferable to plants in other states. Thus, the relatively cheap allowances in State X will attract bidders from every state with higher compliance costs, and those allowance prices will rise accordingly. In its simplest example: If States X and Y have roughly the same level of emissions, and allowances in State X would cost $5/ton, and in State Y $15/ton, then once they share a market the price will stabilize at $10/ton. That is great news for state Y: it has just offloaded 33% of its compliance costs. But it is bad news for ratepayers in state X, whose costs have gone up 50% in order to help plants in State Y pay for compliance.
So if you’re a state with relatively high compliance costs, then you want to sign up for EPA’s cap and trade system in order to get out-of-state ratepayers to help pick up your tab. But if you’re a state with low compliance costs, cap and trade is a mug’s game.
Like a cap and trade auction, a carbon tax creates revenue that a state can use in any way. But unlike an auction, it sets a fixed price (and that price certainty results in utilities making far better capacity investment decisions), is not subject to manipulation, and the price cannot be driven up by other states in the system. The first two factors mean that it is superior to any cap and trade auction system; the second and third mean that it truly is the only rational choice for states with relatively low compliance costs.
The traditional objection to a tax from the environmental groups is that, unlike cap and trade, it does not provide certainty that emissions will be limited. But EPA has eliminated that risk: The CPP requires that states have a backup that automatically comes into effect within 18 months of the end of a compliance period if the tax does not achieve at least 90% of the required reductions. In other words, if the tax doesn’t work, then a mass-based cap and trade system will automatically replace it.
The only other objection to a carbon tax is the flip side to the issue of the certainty of the reductions: what if the state sets the tax at a level that results in over-compliance? Interestingly, this turns out to be a glitch is actually a feature: The state can sell those tons (in the form of allowances) to states in the cap and trade system, and can use the proceeds to refund the additional abatement costs. (The rule is silent on this issue but in order to maximize the allowances available, it seems logical that EPA would allow such sales.) Moreover, if the allowance price in the cap and trade states exceeds the marginal abatement cost of those over-compliant tons (which it almost certainly will, as a tax only makes sense in states with low compliance costs), then the carbon tax creates an incentive for plants to reduce emissions beyond what the state plan requires.
We believe the tax approach is clearly preferable for many states. It’s disappointing that EPA has chosen to load the scales so far in favor of cap and trade. It’s equally disappointing that so many governors seem determined to allow EPA to impose its preferred system rather than submit a plan that adopts the most rational compliance approach for their state.