CO2 tax vs Cap-and-Trade from RFF's Weathervane

Ian Parry says:

The CO2 tax has a number of advantages over pure emissions trading systems. In particular, if the revenues from the tax are used to reduce other taxes, the policy might actually benefit the economy overall, and the tax also avoids problems that might be caused by permit price volatility under an emissions trading regime. 

... an influential study by my colleague William Pizer suggests that the net benefits over time (climate change benefits less emissions compliance costs) under a cap-and-trade system might be only a small fraction of the net benefits under an appropriately scaled CO2 tax.

Am I convinced to get behind a carbon tax over cap-and-trade? Not quite yet, I prefer a carbon tax over cap-and-trade but cap-and-trade over nothing. I think the latter choice is the one we are most likely facing these days. Here is Parry's list of reasons:

Three practical arguments are made against the CO2 tax shift. First is that influential producer groups - refineries, steel companies, airlines, and electric utilities, for example - must be compensated if climate legislation is to go forward and that this compensation is easily provided by giving away free permit allowances to firms under a cap-and-trade system. Second is that voting for any new tax - even if offset by tax reductions elsewhere - can be electoral suicide for members of Congress; the first Clinton administration failed to implement a broad energy tax, despite a major effort. Third, even if a tax regime does go forward, based on how Congress has used new revenue sources in the past, CO2 tax revenues may end up being wasted in special interest spending, rather than being used to substitute for other taxes.

Comments

[dcbob]

I can't deny that Ian's arguments have a ring of truth, but at the same time, I think each point is open to serious question. Most importantly, I don't think anybody ever got a smart piece of legislation through Congress by obsessing about all the reasons it couldn't be done. Our job as economists is to argue for the first best and grudgingly accept the second best after a good fight.

On the first point: if producers really need to be bribed into accepting the program, targeted tax breaks are just as easy as handing out permits. What's so simple about giving the Administration a mandate to decide how to allocate perimts?

On the second: my reading of the polling data is that people know taxes will have to rise at some point, and they know that something has to be done on climate. I don't see most people being terribly keen to learn that Congress gave energy companies the green light to restrict output and gouge them on prices, and I don't think they're generally going to favor giving a windfall to Exxon over a tax swap. And I think they're smart enough to figure that out. And even if I'm wrong on all of that, a revenue-recycling approach that swapped emissions taxes for income taxes wouldn't actually raise taxes, in which case the argument is moot.

On the third: potential waste isn't an argument against using a tax; it's an argument against not recycling. But that's the whole point: you can sell it as a tax swap rather than a tax hike. I just don't see how that's any harder than selling a massive giveaway to the energy companies at the expense of the general population. Bush Administration policies notwithstanding ....

[John]

How difficult would it be to create the marginal cost estimates that would be needed to set the taxation levels for a CO2 tax shift? Are there any groups or academics who have created estimates? Does anyone know any papers that talk about this?

[web tasarım]

............. ........... ................ ....... .......... ................ ......... ......... .......................... .................. ................. ................... .............. ................... .................. .........

[dcbob]

I don't think there's much recent modeling work, but among the people who published on this in the 90s were Larry Goulder, Dale Jorgenson, Warwick McKibbin and Pete Wilcoxen, the Energy Information Administration, and several of the then-prominent macro modeling groups (now all pretty much absorbed into Global Insight). Probably others I've forgotten. Independent of such modeling, though, there's plenty of information about how changes in marginal rates have shifted revenues from different income groups, so the tax estimate gnomes at Treasury, JCT, and CBO, would be able to come up with a fair approximation of a set of sufficient rate changes without too much trouble. Those estimates might not include much in the way of dynamic feedbacks but they would certainly be a good first approximation.

[John]

Oh, I guess I should have been more specific. I meant "are there estimates for the marginal costs of CO2 emissions?", not the tax shifts.

[John Whitehead]

John, check out the "says" link in the post above. there are some estimates there.

[powderhouse]

There are estimates of the marginal cost of CO2 abatement in the recent Stern report -- $66/ton in 2025 and $109/ton in 2015. If the tax were set at these levels, we'd supposedly be on a path to stabilize at 500-550 ppm by mid-century.

A tax is no better than cap-and-trade in terms of revenue recycling if allowances are auctioned. I know this sounds crazy, but I do believe this is the trend after the disaster that was the EU ETS.

NY came out for 100 percent auctioning of its RGGI allowances in its December draft
implementation rule; this was followed by 100 percent announcements by MA,
VT, and CT. The other RGGI states haven't announced a decision yet.

Then a few weeks ago, the National Commission on Energy Policy came out in
favor of 50 percent auctioning (up from 10 percent in their 2003 report);
Bingaman's proposal from last year was based on their work and he's said he
will incorporate many of their more recent recommendations in his upcoming
proposal (which is considered moderate). Also, the Feinstein-Carper bill proposes 15 percent auctioning initially, increasing to 100 percent
in 2036.

[dcbob]

The IPCC's Working Group II has a central estimate for present discounted damages of $43/MTC (bottom of p.20). Richard Tol has a two-year old lit review in Energy Policy ("The Marginal Damage Costs of Carbon Dioxide Emissions") that found a mode of $2/MTC, a median of $14/MTC, and a mean of $93/MTC. Nordhaus puts it at about $23/MTC. Note that these are all for carbon, not carbon dioxide, so divide by 3.67 to get $/MTCO2. Note also that the Stern estimates (which are for CO2) are nearly an order of magnitude larger, primarily because he uses a very low discount rate (justified by a nearly zero social rate of time preference and a low elasticity of marginal utility). Nordhaus has a good discussion of the Stern estimates on his website, as does Tol (with Gary Yohe) on his.

[powderhouse]

Stern's low discount rate is on the benefit side (i.e., the benefit of avoiding climate change), not on the mitigation cost side.

By using a low discount rate, he is able to say that it makes sense to set a tax at $66/tCO2 to stabilize at 550ppm, but the low discount rate doesn't get you the mitigation cost estimate -- he's just able to say that the marginal benefit is >= the marginal cost of a 550 target.

His mitigation cost estimates are distinct from his modelling of benefits (see Ch. 9 and technical paper by Anderson, on Stern website). Stern estimates that a 11 GtCO2 reduction from BAU in 2025 to stabilize at 550 ppm will have an average cost of $33/tCO2 and a marginal cost of $66/tCO2. The discount rate on the cost side seems to be 10 percent (this is hard to find, though; buried in the Anderson technical paper).

[powderhouse]

As for the IPCC estimate of the social cost of carbon, it's an average of $12/tCO2 with a significant amount of uncertainty; there's a lower bound of -$10 and an upper bound of $130.

I don't know what the correct number is, and neither does anyone else -- that's the point. There's enormous uncertainty as to both the marginal benefit and cost of abatement, so it's difficult impossible to know where the curves intersect and where the tax should be set. Granted, the same argument could lead to the conclusion that it's impossible to know where the cap should be set.

But if we can agree that some ppm level is too high (1000? 750? 550?, take your pick), then we can map out the remaining carbon budget and a cap-and-trade system will get us to that target. A tax will bring emissions down, but there's less certainty as to whether we'll wind up at the desired target.

[powderhouse]

Found one more estimate -- the IEA assumes that a tax of $25/tCO2 in 2050 will get us to a 650+ ppm target. This is from their Energy Technology Perspectives report (2006). Getting to a lower target (like Stern with 550) would cost more.

All of these mitigation cost estimates are contingent on ppm-target and year (as well as technology cost and deployment assumptions, etc, but those are even harder to unpack). I think it's important to separate the mitigation costs of reaching a particular target from the benefit estimates of that target.

If the marginal benefit is less than the marginal cost -- as suggested by a quick comparison of the Nordhaus estimates with the Stern estimates -- then maybe we don't want a 550 ppm target? Or if we do want to prevent reaching more than double the pre-industrial level, maybe we doubt Nordhaus' estimate of the marginal benefits?

[dcbob]

powderhouse, I think you're you're conflating estimates of prices necessary to achieve particular emission, concentration, or temperature targets with estimates of prices necessary to balance net present expected costs and net present expected benefits. One of the Stern report's many oddities is that it present one (very high) value for marginal benefits but then argues that marginal mitigation costs are very low. It's a jarring statement from an economist, to say the least. The estimates I cited are all estimates of the expected net present value of benefits (or damages avoided) from a marginal ton of carbon (or carbon equivalent) emitted today. The IPCC value of $12/MTCO2 that you cite is the CO2 equivalent of the $43/MTC I cited. Start at $43/MTC and let it rise at a real rate of 3% to 5%, and by standard estimates you'll price carbon out of the market over a century or so, hitting an ultimate conentration of around 550ppm and a temperature of somewhat under 3 degrees. Start at Nordhaus' $23/MTC and you end up a bit over 3 degrees. All of that with a good deal of uncertainty, of course, so you'd certainly have to adjust course over time - maybe up, maybe down.

[powderhouse]

Stern doesn't say that the marginal cost of mitigation is low at all -- $66/tCO2 in 2025 is fairly high, no? He obscures the marginal cost, preferring to assume for convenience that it's equal to the average cost (see ft 23 on p. 304).

I'm just saying that it makes sense to break down these estimates, so that it's clear what the underlying assumptions are in terms of the year we're talking about and the implicit ppm target. The estimates of the present-day social cost of carbon (a la Nordhaus) hid a lot of this type of detail. Better to be explicit about the marginal abatement cost and the marginal benefit for a specific target at a particular date.

For example, if $43/tC increased at 3-5% per year and led to a phase out of carbon in 100 years, I think we'd be at a lot higher than 550ppm CO2e (which is Stern's target). Of course, that depends on your BAU assumptions about economic growth, etc. It's all very complicated and uncertain, which is why a cap-and-trade system is more likely to get you to a fixed target than a tax.

[dcbob]

powderhouse, you're right that $66/MTCO2 is fairly high - about $240/MTC or 10 times Nordhaus' esimate of the net present value of benefits and 5 times the IPCC mean estimate. (The box on page 304 says that Stern's central estimate of the marginal benefits is even higher - $85/MTCO2 or about $310/MTC.) But from my understanding of Stern - at least from his executive summary - his argument is that mitigation costs for the trajectory he advocates for reaching stabilization below 550 ppm is quite inexpensive for the first several decades. (I'm not sure how that can be, or how it squares with his high estimate of marginal benefits; I still need to digest the report. I think he estimates benefits using a very low discount rate to justify a low temperature target and then estimates a price path that gets there using a market discount rate to distribute the allowable emissions over time, but that's just a guess on my part. My own discussions with Stern's staff left me quite confused on this point.)

Nordhaus' model and presentation (which are a lot more transparent than Stern's) suggests that a price path starting at $40/MTC gets you to a 2 degree limit on temperature change, and a path starting at $33/MTC gets you to a 550ppm target. Nordhaus' preferred optimal path starts at around $30/MTC and ulitmately reaches 720 ppm and 3.3 degrees in 2200. But he also points out that reaching a 550 ppm target involves less than $1 trillion more in present value costs, compared with his optimum. Obviously there are huge uncertainties around those estimates, but they seem like reasonable ballparks. (Nordhaus also replicates the Stern results using his model, by the way; the price path starts at around $150/MTC.)

I agree with you that uncertainty reigns on this issue, but I fail to see why that argues for a cap-and-trade system rather than a tax. Why is certainty about quantities of emissions or about the ultimate target so important if it's not certain what the benefits are, especially if caps yield much higher (or lower) prices than are consistent with our best guess of marginal benefits? The standard economic view - best articulated by Billy Pizer at RFF - is that for a stock externality like greenhouse gases, a price instrument overwhelmingly dominates a quantity instrument in the face of uncertainty.

[powderhouse]

I work with scientists, and have tried to explain the idea that a tax is better than cap-and-trade given uncertainty about the marginal benefits (which I think is what you're talking about; I'm familiar with the Weitzman article from the 70s), due to the relative deadweight loss under the two instruments.

The response I get is -- "how could we possibly monetize the benefits? whatever the dollar figure is, more than 550 ppm is catastrophic." I've gradually been convinced that instead of trying to estimate the MAB and MAC curves, it makes sense to pick a ppm target and then figure out how to get there most cost-effectively. Cap-and-trade does both of those things.

[dcbob]

Yeah, that's right, I'm referring to the Weitzman-type argument, which Pizer and others have shown is even more compelling for a stock externality. I get the same argument from scientists that you do, and I've thought long and hard about it. I'd be willing to live with something approximating a 550ppm limit (in net present value terms it's just not that much more expensive than the Nordhaus optimum) but I still don't believe that a long-term concentration target require a cap-and-trade system - certainly not a system that involves hard annual targets. Starting with something like a $35-$40/MTC price would more than double the price of coal and over a decade or so it would give us a good gauge of how accurate the Nordhaus-type pathway is, giving us a chance to make course corrections along the way. The cap-and-traders bend themselves into pretzels to include banking and borrowing and/or safety valves and so on, but all of that just serves to move the proposed system to something more like a tax. Why not just have a tax and be done with it?

[Bates]

The one thing I don't like about most of the comparisons between the tax and C&T is that it compares a "best" case scenario to a "worst" case scenario. The "best" case scenario of the tax is a rebate, or tax reduction elsewhere, while the "worst" case scenario of C&T is that permits are given away for free.

Is there anything out there that gives a more 'apples to apples' comparison?

[Bates]

The following is a shot in the dark, but here it goes:

I like C&T for several reasons, but one of the major reasons I support C&T for GHG emissions is the potential to include it in an Ecosystem Multiple Market. I just came across this concept recently. It makes a lot of sense to me. See what you think:

http://www.envtn.org/docs/emm_es.PDF

If you don't want to read, the basic idea is that several activities (such as creating riparian buffer forests) produce a number of environmental benefits (i.e. water quality, reduced GHGs, wildlife habitat). A market that integrates all of these ecosystem commodities should create a higher demand for these services than fragmented markets for each problem.

[Vitch]

Is there not a concerning possibility that if a C&T system is used, there will be a shift in the thinking of environmental integrity from natural environments to man made environments?

When considering the TechnoGarden scenario of the Millenium Ecosystem Assessment (Page 254) it is suggested that people may start to develop limited, somewhat simple views of nature and consider, for example, a 'put back what you took out' answer to healthy ecosystem approaches. These approaches may include practices such as planting trees in an area different than that you are cutting them down in. This would result in the same desired affect as just leaving the trees standing to begin with (reducing CO2 emissions), but would not be ecologically sustainable. Essentially this could result in the replacement of wilderness with intelligent 'gardening' for human purposes.

Let me know what you think!