How the Senate can fix cost containment in the climate bill with ‘price collar plus’

The climate and clean energy bill that narrowly passed the House has three problems related to cost containment (CC) that the Senate should — and I expect will — address:

  1. Fence-sitting Senators (and industries) worry that its CC provisions aren’t hard-nosed and specific enough to protect the public and businesses from carbon prices that get too high too fast, possibly driven by speculators.
  2. Progressives worry that its CC provisions are already too strong and that some proposals floating around to strengthen CC would environmentally weaken an already weak bill.
  3. The major CC provision in the bill — the strategic reserve — is so opaque that it is understood by a handful of people at most and none of them are U.S. Senators.

I’m going to try to take the best of all the current CC proposals and propose an alternative that I think might actually be appealing to all sides, what I’m calling “price collar plus.”

Two weeks ago, the Brookings Institution — which I’d view as center-right on the energy and climate issue now that David Sandalow has left — proposed a traditional price collar in Politico, “Time for a price collar on carbon.”  To their credit, they did suggest this was a way to “rein in offsets” but offered no specifics on how to achieve that important end.

The benefit of a price collar to Brookings:

By preventing the policy from being either unexpectedly lax or unexpectedly stringent, a price collar protects both investors in green technologies and households and preserves strong incentives to abate.

The House climate bill already has a price floor for the auction, which starts at $10 a ton in 2012 and rises 5% plus inflation every year thereafter.  I believe most everyone understands the need for a rising price floor — giving some certainty to businesses about investment decisions they make, say, in biomass cofiring or natural gas fuel switching.  The floor in Waxman-Markey is, by almost every independent analysis, on the low side in the sense that the CBO and EPA and especially the EIA project the price for a CO2 allowance in 2020 will be above the floor — in EIA’s estimation, double the floor price.

The fossil fuel industry, of course, funds economic analyses that project incredibly high allowance prices to scare people into opposing the bill entirely.  If their analyses were anywhere near accurate, the floor in the House bill would be utterly irrelevant.  I’d love a higher floor, but since it has already passed the House, we’re probably stuck with it.

A price collar, of course, requires a ceiling to go with the floor.  Brookings explains:

The price ceiling could work like the “safety valve” included in a 2007 bill introduced by Sens. Jeff Bingaman (D-N.M.) and Arlen Specter (D-Pa.), which would have allowed the government to sell additional emissions allowances if permit prices rose above a preset ceiling.

That kind of cap-busting safety valve is not good from an environmental perspective (see “Safety Valves Won’t Make Us Safer“).  That’s why I have long opposed such safety valves (see “The history of the ‘safety valve’ debate“), especially when set at ridiculously low levels, such as $7 per metric ton of CO2-equivalent (and rising a tad above inflation annually), as the National Commission on Energy Policy proposed in 2004.

NCEP’s new report, “Managing Economic Risk in a Greenhouse Gas Cap-and-Trade Program,” also endorses a safety-valve-type ceiling, but then wisely offers up this proposal:

An allowance reserve coupled with a price floor offers, in our view, many of the benefits of a simple price cap and has the not insignificant advantage of providing greater certainty about cumulative emissions reductions over the time horizon of the program.


You don’t want the government to sell an unlimited number of allowances that represent no emissions reduction whatsoever at the ceiling price.  You want to borrow the best feature of the strategic reserve, which is that the allowances the government sells are, to start, skimmed off of the emissions caps from 2012 to 2050.

In Waxman-Markey, a pool of allowances is made available for strategic reserve auctions consisting of

  • 1% of the allowances established for each year from 2012 to 1019
  • 2% of the allowances established for each year from 2020 to 2029
  • 3% of the allowances established for each year from 2030 to 2050

That was I think a good compromise by environmentalists.  It acts a lot like a safety valve, but maintains environmental integrity (at least to start).  The enviros (and whoever else signed off on this deal), however, made two mistakes.

First, in the final House bill, they set an initial trigger price for the strategic reserve of $28 — which is the equivalent of the “ceiling” or “safety valve” price — but that price quickly shifts to 160% of the average auction price of allowances over the previous 36 months.

Zzzzzzzzzz.  Crickets chirp.  Glaciers melt.

That approach was dissatisfying to everybody — or rather it was confusing to everybody and dissatisfying to the few people who wasted time figuring out what it meant.  For progressives who think there are an overabundance of domestic clean energy solutions available, and hence that the permit price will stay close to the floor for at least a decade (see here), it meant the reserve auction trigger price — aka the effective ceiling price for allowances — might be maybe only $22 a ton in CO2, a ridiculously low ceiling.  And that meant if we turned out to be wrong about, say, the supply of moderate-cost natural gas, then even a tiny allowance price spike would trigger the reserve auction.

But for moderates and conservatives, who tend to believe that the allowance price in 2020 will be much higher, even higher than EIA’s $36 a ton, then the ceiling in 2020 might be $60 a ton or higher, which for them is no protection at all from speculators or from the technology optimists being wrong or from offset prices being much higher than they thought.

The point is, the strategic reserve “ceiling” price in the House bill was designed in a manner to make everybody unhappy.  For instance, NCEP — which I’d characterize as center right today (see here) — was worried the ceiling/safety-valve price in 2015 might be as high as $49 a ton [though I think they did their math a little wrong].

Now NCEP does say:

We do not take issue with the initial allowance trigger price proposed in Waxman-Markey (at $28 per ton)“”rather our concerns focus on the method used to calculate the trigger price in subsequent years.

Me, too.

A majority of House members voted for the reserve trigger price to rise 5 percent plus the rate of inflation for 2013 and 2014 until the complicated formula kicked in for 2015 on.

Sp I’m going to propose what I think is the simplest and most obvious fix:   The floor price for the regular allowance auction should start at $10 a ton in 2012, and the reserve trigger price (aka the effective allowance ceiling price) should start at $28 a ton in 2012 — and those collar prices should rise 5% plus inflation every year thereafter.

NCEP elaborates on the benefits of a price collar:

A “price collar” retains the economic efficiency benefits of a price ceiling alone, which has been shown to be nearly as efficient as a carbon tax.  Moreover, recent research — [RFF’s”Alternative Approaches to Cost Containment in a Cap-and-Trade System“] — has demonstrated that a “price collar” approach has the additional benefit of reducing long-term emission abatement costs relative to expected long-term abatement costs with a price ceiling alone. This is because the policy provides more consistent financial incentives for sustained investment in low-carbon technologies that can reduce compliance costs in the long run: Rather than being subject to boom-bust cycles when allowance prices fall, new low-carbon technologies would be assured a certain level of market stability. This would allow them to develop in a more orderly and ultimately cost-effective way.

But NCEP also explains the value of the reserve:

… a robust, well-designed reserve auction mechanism could be extremely useful for increasing public confidence in the nascent greenhouse gas market. If true costs are much higher than projected, the reserve would provide a “cushion” while Congress considers whether further program adjustments are needed. On the other hand, if allowance prices are in line with, or modestly above expectations, the allowance reserve auction would never be triggered.

Price collar plus should be attractive to both sides

Fence-sitting Senators and industries can legitimately see it as achieving stronger cost-containment protection than their analysis suggests the House bill now provides, including protection against speculators running the permit price up, while progressives can legitimately see it as achieving better environmental outcomes than their analysis suggests the House bill now provides.  Win-win.


I would keep the W-M provision that “the annual limit on the number of emission allowances from the strategic reserve account that may be auctioned is an amount equal to 5 percent of the emission allowances established for that calendar year.”  It is hard to see how one would need more than 5% in any given year, especially when there are so many domestic and international offsets available for emitters to purchase — and of course so many strategies emitters can use to reduce their emissions and hence their need to purchase permits.

BUT I would change how Waxman-Markey refills the reserve once the initial reserve is auctioned out.  W-M fills the reserve with “international offset credits from reduced deforestation.”

Bad idea.  Reduced deforestation should be utterly separate and additional.  We have no hope whatsoever of averting catastrophic global warming if we don’t sharply cut fossil fuel emissions here (and abroad) while simultaneously stopping deforestation [see “How the world can (and will) stabilize at 350 to 450 ppm: The full global warming solution“].  And one of the best things in the House bill is that it already devotes substantial funds generated from the allowances to stopping deforestation — achieving some 720 million tons of emissions reductions in 2020, equal to 10% of total current US greenhouse gases — all of which are additional to the domestic GHG reductions.  The notion that deforestation tons should be separate and additional should be be an inviolate principle of U.S. action.

No, I would fill the reserve with domestic offsets.  I’m not really expecting the initial reserve to sell out until well past 2020.  And I know the businesses who signed onto this deal wanted a large pool to refill the reserve — but at the likely trigger or ceiling price post-2020 (more than $40 a ton of CO2e), there would in fact be a lot of domestic offsets.  And I have more confidence in our ability to ensure the quality of domestic offsets than I do of our ability to ensure the quality of international offsets (though I do expect the quality of the latter to get better).  Moreover, if CBO is right, then half of the domestic offsets are going to be genuine emissions reductions in uncapped sectors.  And the other half will be soil/forestry/agricultural sequestration, which should make certain politically powerful domestic groups happy.

So this strikes me as both better environmentally and more attractive politically to US Senators.

Finally, I’d like to re-offer my suggestion of how to “rein in offsets,” as Brookings suggests.  I consider all of the following cost containment measures a major concession by those who want the strongest possible environmental integrity for the bill:

  • Price collar plus
  • Too weak of a 2020 target.
  • Up to 1 billion domestic offsets in place of emissions allowances.
  • Up to 1 billion international allowances (a number that can be potentially revised upward to 1.5 billion if the domestic number is revised down).
  • Allowances distributed to regulated utilities and other entities to directly mitigate cost impacts on the public and businesses.
  • Tremendous energy efficiency efforts that will also directly mitigate cost impacts on the public and businesses.

So my final recommended change is one I have been proposing for a while —sunset the offsets. My more politically palatable version is to apply the same reduction to the offsets that you are applying to emissions in the bill:

  • a 17 percent cut by 2020 (to 1.66 billion tons)
  • a 42 percent reduction by 2030 (to 1.16 billion)
  • an 83% cut in 2050 (to 0.34 billion)

I am aware that the domestic offsets are probably too popular to sunset — so the sunsetting could be applied simply to international offsets.  The other advantage of that, as one economist told me, is that it would provide extra motivation to developing countries to engage in the process early, since they’d know that the U.S. wasn’t going to keep purchasing international offsets forever.

There it is — price collar plus.

9 Responses to How the Senate can fix cost containment in the climate bill with ‘price collar plus’

  1. Ken Johnson says:

    If I understand it correctly, the “strategic reserve” isn’t really a price ceiling — it’s more like a form of banking. Allowances that are “banked” in the reserve are sold at the trigger price (e.g. $28), but if the reserve is depleted then there is nothing to prevent allowance prices from hitting $100 or $1000. But if prices are expected to remain low, then a real price ceiling (even one starting below $28) might not be a bad compromise in exchange for a higher price floor (e.g. starting at $15). Maybe a ceiling higher than $28 would be politically acceptable if it is really a hard ceiling.

    [JR: You do not understand the strategic reserve correctly — you are not alone here, but you haven’t even understood my description of it, for which presumably I bear some blame. It is NOT a form of banking. It is, to start, a form of borrowing from the targets from 2012-2050. But then it is refilled, poorly, I have argued above.]

    If the price floor results in fewer than the authorized number of allowances being sold, then the unsold balance could be put in the strategic reserve and held for sale at the ceiling price — that’s how banking normally operates. If the reserve is depleted, it could effectively become a debit account (i.e. the reserve would be negative), and the ceiling would become the floor price until the debit is cleared. In this way, the reserve would operate alternatively as a banking or borrowing mechanism. Without some form of borrowing, there would be no guarantee that allowance prices will not hit $100 or $1000.

    The problem with a price collar is that it leaves cap-and-trade without a coherent policy foundation. The fundamental policy objective of a “textbook” cap-and-trade system is to achieve a predetermined cap a minimum cost. But with a price ceiling the cap is no longer guaranteed, and with a price floor regulatory costs may be higher than the prescribed minimum.

    [JR: Again, you have misread W-M and hence my proposal. I assert that my proposal achieves better environmental outcomes than W-M from the perspective of progressive analysis of the bill, while at the same time achieving better cost containment than from moderate/industry analysis of the bill.]

    In view of political and climatological realities, a more appropriate policy goal would be to cap costs and minimize emissions. That’s not what cap-and-trade is designed to do, but a price collar is a step in that direction.

  2. Peter Wood says:

    Good post Joe.

    Much of the literature suggests that a price collar (what is usually described as price ceilings or price caps and price floors) approach has significantly lower expected costs than a purely cap-and-trade approach. Two of the most important papers are:

    Roberts, M. J., Spence, M., 1976, ‘Effluent Charges and Licenses under Uncertainty’, Journal of Public Economics 5, 193-208. and

    Philibert, C. (2008) ‘Price Caps and Price Floors in Climate Policy: A Quantitative Assessment’, International Energy Agency, Paris, France.

    Philibert’s results suggest that the expected cost of globally reducing emissions by 50% compared to 2005 by 2050 using pure cap-and-trade would be similar to the expected cost of of an approach with price ceilings and floors that has a cap reducing global emissions by 75% by 2050.

    Philibert also finds that:

    A climate policy with price caps set below best guess marginal abatement costs will not achieve its stated objectives, but may remain largely preferable to the absence of any policy. Price caps should be set higher than expected marginal costs. Price floors would further reduce the expected costs of achieving a given environmental result.

    For this reason I think that a starting price of $28 per tonne is too low, but I think applying a 5% discount rate is a good idea. The 5% discount rate provides a very strong long term-signal that can help guide investment.

    [JR: I agree that $28 is too low, but it is what the House agreed to, so I doubt the Senate would go much higher. With the 5% real increase per year, though, it is acceptable.]

    The standard approach to having a price ceiling (including the approach in the Brookings proposal) is to sell an unlimited amount of permits at the fixed price. An objection to ceiling prices for greenhouse gas emissions is that they would make it less likely that countries meet international commitments to reduce their emissions by a certain target. If a price ceiling is not a strict ceiling, because the amount of extra permits is still limited, this is no longer a problem. This is something I like about the strategic reserve proposal, and variant thereof.

    A simpler alternative to the strategic reserve would be to auction most of the permits with an auction reserve price that is the floor price ($10 + 5% pa), and auction the rest of the permits with the ceiling price as a reserve price, if not all are sold, they can be saved for later.

    In their appendix, Roberts and Spence suggest that by issuing an arbitrary number of different kinds of permit, it would be possible to approximate any convex damage function arbitrarily closely. This could further reduce expected costs. This would be easy to implement: auction most permits at the floor price; auction some more at a higher price (e.g. $28 + 5% pa); and more at an even higher price (e.g. $50 + 5% pa).

    There are significant advantages to these sorts of approaches in terms of international cooperation. International environmental agreements on issues such as acid rain and pollution of the North Sea have had weak binding commitments, but also have had ministerial level nonbinding commitments that are significantly stronger (Victor, 2007). An approach could be introduced where there are two or more different types of permit, the total number of permits is based on a binding commitment, and the number of permits with a low reserve price is based on the non-binding commitment.

    All of the above mechanisms could be done (in a superior way) by not having a reserve price, and instead having firms pay an “extra fee” when they surrender their permit. There could be different extra fees for different types of permit. This may work better with international trade of permits.

  3. Peter Wood says:

    It would make sense for the House to pass the bill with a $28 reserve price if the price quickly shifts to 160% of the price over the previous 36 months — it makes sense to phase these things in. I don’t think that the 5% real increase makes it much better, but it stops it from getting less bad. Marginal damages of emissions are expected to increase, and marginal costs of emissions reductions are expected to increase at the same time. The auction price could exceed $28 plus 5% pa of course, because the reserve is not infinite — a lower reserve price increases the likelihood of the reserve fund being depleted.

    I think I would prefer a reserve price of 160% of the price of the previous 36 months to 5% per year above a previously low value, but accept there may need to be compromises to get the bill through the Senate.

    The problem with the $28 reserve price is that it translates into only $41.37 (2012 dollars) by 2020. But in 2020 the amount of permits going into the Strategic Reserve Fund will increase by an extra percentage point (as it also will in 2030). It would be quite easy to put this in a new strategic reserve fund (call it Strategic Reserve Fund 2) which is auctioned at a higher reserve price (twice as much could be appropriate, or 160% of the price of the previous 36 months).

    This would be slightly more complicated on paper (maybe an extra page out of 1440 or so), but would have significant benefits, as suggested by the work of Roberts and Spence.

    One thing I like about the strategic reserve is that if the cost of emission reductions is sufficiently low, U.S. emissions could be 1-3% lower than specified by the cap (lower still if the price floor comes into play). If the U.S. wanted to tighten its emissions reductions at a later date, one way it could do so is by increasing the percentage of permits that go into the reserve.

    I have a couple of comments on sunsetting the offsets. One way that this could be done is instead of decreasing the amount of available offsets, one could increase the amount of offsets that you need to account for a tonne of greenhouse gasses. Also, it may be politically easier to sunset offsets if some of the permit auction revenue was used to fund biosequestration of carbon dioxide in ecosystems. This way the biosequestration would be additional.

    [JR: I confess I don’t follow you. The allowance price isn’t going to be anywhere close to $41 in 2020, as I’ve shown. So my proposal is better than W-M, from a progressive/environmental perspective. But for oderates and industry who don’t believe my analysis, my proposal is also better. But the 160% is just a loser — for reasons I explained.]

  4. Ken Johnson says:


    I’m still a little unclear on the banking-versus-borrowing distinction. Could you please correct or clarify the following points?:

    (1) The main differences between your proposal and W-M is (a) the reserve trigger price (5% annual compounding), and (b) the way the reserve is filled (using domestic offsets).

    (2) In either case, reserve allowances would only be released at the trigger price, and only up to a limit of 5% of the total allocation in any calendar year.

    (3) In either case, allowance prices would not be limited if the reserve is depleted.

    [It’s a little unclear how 1(b) would work, since domestic offsets are creditable toward compliance obligations anyway. Is the idea that domestic offsets in excess of the 1 billion ton limit could be used to replenish the reserve? Also, (3) seems to be incompatible with “borrowing”, which would allow emissions to at least temporarily exceed the cap.]

    [JR: Yes to #1. Yes to #2 — same as W-M. Not sure I understand #3. The point is to keep refilling the reserve. Of course, it probably starts with 3,000,000,000 tons (I haven’t done the math) and I don’t see it being needed until 2030ish at the soonest (though obviously industry models would suggest otherwise). I can’t see it being depleted for a long, long time. In any case, you use the proceeds from the reserve auction to buy domestic offsets at whatever price they are available for — but you can’t sell them below the trigger price. I’d think domestic providers of offsets would like that.]

  5. Ken Johnson says:


    Re “you use the proceeds from the reserve auction …”: What does W-M say about the use of reserve auction proceeds? Using the proceeds to buy domestic offsets would seem to be the same as just buying allowances because offsets are interchangeable with allowances (up to the 1 billion ton limit). Also, there wouldn’t be any proceeds unless the trigger price is triggered, in which case no one will want to sell at or below the trigger price. If offsets/allowances are repurchased at the trigger price, then the trigger price would become a floor price until the revenue is all spent, which would occur when the reserve is sold out. If the reserve is empty, it would seem that there would be no cost containment (?)

    [JR: No. Rethink what you’ve written. Offsets are not the same as allowances.]

    The prospect of an empty reserve and unlimited allowance prices may be exceedingly unlikely, in which case a hard price ceiling would do no harm. Without a guaranteed price ceiling, the perception that prices might go “through the roof” could necessitate unnecessary political compromise (e.g. weak emission targets).

  6. Ken Johnson says:

    Hmmm … Let me think about this.

    HR2454 says “Covered entities collectively may … use offset credits to demonstrate compliance … by holding 1 domestic offset credit or 1.25 international offset credits in lieu of an emission allowance.” (p 740-741) If allowances are selling for $28 but domestic offsets are available for less than $28, then covered entities would logically prefer to use the cheaper offsets to meet their compliance obligation. The demand for cheaper offsets would tend to bring the price up to parity with allowances. If the $28 price trigger is triggered, it’s unlikely that anyone would be willing to sell either allowances or offsets for less than $28 because the market is willing to pay $28. On the other hand, if you use reserve auction revenue to buy domestic offsets or allowances at $28 or higher, then the market price would go up because you would be reducing the supply of offsets and allowances — which kind of defeats the purpose of the strategic reserve.

    [JR: No. The strategic reserve is not designed to continuously replace the auction, but to occasionally be tapped when the price peak or blips up — if you designed it right. So you must have a reasonable starting price and annual increase, like I propose. It is all but inconceivable to me that if my proposal were adopted as part of Senate bill that was roughly the same as the house bill in other respects, that the reserve would be seriously drawn down before 2030, if then.]

  7. Ken Johnson says:

    The reserve could be used to sell allowances at the predetermined trigger price (e.g. $28), with revenue used to repurchase allowances and/or offsets at a predetermined floor price (e.g. $10). “Buy low, sell high.”

    If it is “all but inconceivable” that the trigger price would be significantly exceeded, why not make it a hard price ceiling (in exchange for some negotiating concessions, of course, like a higher floor or more stringent cap)?

  8. R says:

    Starting from Ken’s points 1) and 2) from comment #4, which JR agreed with, do I understand correctly that if 105% of a yearly cap’s allowances have been given out (the original 100% plus the yearly reserve allowance of 5%) and the market-clearing allowance price is still above the official ceiling, then there will be no further action to cap the allowance price in that year? I.e. the price ceiling is not absolute because we are only allowed to release up to 5% of the yearly quota from the strategic reserve in any given year.

    (Perhaps this seems improbable given JR’s analysis, but I’m trying to understand how the system functions at the extremes)

    [JR: The reserve isn’t “given out” so your comment doesn’t make sense.]

  9. R says:

    Please help me understand where I go wrong in this hypothetical scenario:

    1) assume for argument’s sake that it is 2015, the cap is 100 Gt CO2e (wrong, I know, let’s just use 100 to make the math easy), 100 Gt worth of allowances have been either given out or auctioned, and the market-clearing price is above the defined ceiling (let’s say $30/ton)

    2) as a result, extra allowances – up to 5 Gt worth – are auctioned from the “strategic reserve”, and those who buy them can use them in the year 2015

    3a) if the extra allowances (so now 105 Gt of allowances for 2015 are on the market in total) bring the price below $30/ton, great, the price collar has worked

    3b) if even with the extra 5 Gt of allowances, the market-clearing price is still above $30/ton, no further allowances are auctioned in 2015 and the price is allowed to trade above the ceiling of $30/ton

    Is this how “price collar plus” would be designed to function in the case that the market-clearing price for allowances is higher than expected?