I founded BAN CARBON, a nonprofit that stops American power plant emissions by taking away their rights to pollute. I’m an economist (UChicago PhD, 2022). Before BAN CARBON, I researched mechanisms that collect private information to cost-effectively fund public goods.
Paco del Villar
Karma: 15
Hi Ville, thank you. I agree this is a central concern. If retiring 100 allowances leads the government to issue an extra 20, the climate effect is just 80%.
I also agree that politicians benefit from raising the cap in response to high allowance prices (ultimately driven by large-scale retirements). Raising the cap
pleases the polluting industries, which can reward politicians (e.g. via fundraising).
lowers the polluting industry’s (private) costs, expanding supply and lowering consumer prices. And lower prices = more votes.
My point is that politicians also face significant costs of raising the cap:
Political capital. Changing caps requires legislative or regulatory amendments (such as EU directives or RGGI state regulations), which are politically costly and resource-intensive.
Judicial risks. Financial institutions hold allowances as assets. They would initiate lawsuits against arbitrary cap increases, which lower allowance prices. A notable example is the Acid Rain Program (the world’s first cap-and-trade), where states and polluters sued over an arbitrary cap reduction and won in 2008, ultimately contributing to the end of the program (Schmalensee and Stavins, 2013).
Revenue impact. Governments generate revenue by auctioning allowances. A policy of raising the cap when allowance prices are high would trigger investor sell-offs, depressing prices and government revenue. Rather than being a threat, third-parties who retire allowances are useful to politicians: they raise taxes for them.
Historically, the political problem in cap-and-trade has been lowering, not raising, the cap. For instance:
The federal government’s attempt to tighten the cap in the Acid Rain Program failed.
In RGGI, caps have been systematically lowered by removing “banked” (unused) allowances. Retired allowances are considered banked allowances, so the cap reductions have amplified allowance retirements. For example, retiring one allowance in 2013 would have resulted in the effective retirement of three after adjustments: a tripling of the effect (see the second and third cap adjustments for banked allowances).
Your point that the cap is endogenous to allowance prices is well-taken. Historically, what politicians want is lower caps and higher prices—exactly what retiring allowances delivers. For current and moderately higher allowance prices, my sense is that the risk of politically motivated cap increases remains modest.
Lastly, you can retire allowances without causing cap increases in cap-and-trade programs where the relationship between the cap is an increasing and publicly known step function of the price of allowances. For example, RGGI employs a uniform-price auction and releases additional allowances only if the clearing price exceeds a publicly-known trigger (it’s called the cost containment reserve). To make sure that your bidding does not inadvertently raise the clearing price and raise the cap, you just bid below the trigger price. If the auction clearing price surpasses the trigger, your bid is below the lowest winning bid, so you don’t affect the outcome.
That report is super valuable, thank you for sharing it. I thought there were no better climate actions hehe. Please let me know what’s on your mind.
The point of my post was to clarify what retiring allowances accomplishes under what circumstances, and it is my fiduciary duty to address the criticisms ;) I’ll keep Ultra Civic on the sidelines, as I’d prefer to discuss the mechanics of allowance retirements. I’ll bring up inflationary taxation via crypto issuance as an alternative to finance public goods in another post.
I’m arguing that skepticism is not warranted in certain programs (like RGGI).
The idea that additionality requires polluters not to have banked allowances is wrong. Granted: if every year they used all issued allowances, then retiring an allowance causes a one-ton reduction in emissions this year. If they were to use all issued allowances over a longer time period, the retirement would cause a one-ton reduction in emissions over that longer period. We can never know if polluters will use all issued allowances, but we can know if polluters plan to use all available allowances, and the answer is positive if the current market price of allowances is bounded away from zero.
Retiring allowances in a small cap-and-trade program like NZ is not super scalable—at best you’d correct the negative externality of emissions in tiny NZ. But the programs I listed cover 2GtCO2/year. I’d say there’s potential for scale?
(RGGI-specific) It’s good that RGGI imports power, not bad—imported power has a strong Canadian hydro component and is 40% cleaner than in-RGGI power. It would be bad if RGGI’s power imports were mostly coal-fueled.
Thanks again for bringing all this up (and sorry for taking so long to respond—I’ve been travelling a lot)
Hi ethai, thanks for reaching out! I learned about Climate Vault through a friend and longtime Greenstone RA, who suggested the allowance retirement idea to kick off a crypto project that crowdfunds public goods. I loved the idea. I wouldn’t say that Climate Vault and Ultra Civic are super different. As far as I’ve seen, they differ in that:
Climate Vault purchases allowances with the intent to sell them later on and fund carbon removal. Ultra Civic permanently retires allowances.
Ultra Civic encourages DIY: enter the cap-and-trade program yourself, retire allowances, and earn tokens. I think it’s fair to say that Climate Vault keeps the cap-and-trade mechanics a “secret sauce”.
Hi Ulf, thank you for bringing up Equal Right: I was actually not familiar with it!
From their cap-and-share proposal, I gather that they advocate for a cap on emissions with allowances that aren’t tradeable. An argument in favor of trading allowances is that polluters can freely redistribute allowances towards those who value them more (i.e. emit to produce more valuable things), resulting in lower pollution abatement costs. Cap-and-share involves direct government control over individual polluters, which makes climate mitigation costlier: Greenstone et al. (2025) found that cap-and-trade reduced pollution abatement costs by 11% relative to the traditional command-and-control approach in India.
Thank you for commenting! I believe you’re saying that the cap could be large enough that there’s excess allowances, so retiring one doesn’t affect polluters’ emissions.
The market price of allowances reflects whether the cap is binding or not. An allowance that polluters don’t plan to use is worthless. If polluters did not plan to use them all, their market price would have to be near zero, as was the case in the EU ETS in 2007 (~$0.10/allowance) or in the Acid Rain Program today ($0.01/allowance in the 2025 auction).
For RGGI, a question is: do power plants plan to use all allowances given that they were willing to buy every last one for $20 in the last auction? I would think so.
Thanks for the guidance!
Hi ethai! Thanks again. I had the chance to read Making Climate Policy Work and found it super insightful.
Regarding perfect foresight
You’re right: I glossed over the extent of imperfect foresight when discussing uncertainties. For RGGI specifically, here are reasons why polluters likely have accurate expectations about their future allowance usage:
Actual RGGI caps have gone according to plan, with the exceptions of three adjustments explicitly made to reduce excess banked allowances (see my response to Ville’s comment below, adjustments 1-2, and adjustment 3).
RGGI covers power generation, arguably the sector with the clearest and most predictable pathways toward decarbonization (Cullenward and Victor, 2021, Fig. 1.2).
Recent forecasts indicate higher power demand in 2026--2030 than previously anticipated. If anything, this suggests that power plants may have underestimated their allowance needs, making full usage more likely.
There are also valid reasons for uncertainty:
RGGI is inherently political and dependent on voter preferences, which can change unpredictably over a 5-year horizon.
Low-carbon tech exists (solar, wind, nuclear), but breakthroughs in fusion or storage could unexpectedly render fossil fuels obsolete.
Whatever the foresight of polluters is right now, it should improve as third-parties retire allowances. Right now, a power plant’s cost of purchasing a CO2 allowance it won’t use is $20. The higher the price of allowances, the greater the cost of making mistaken purchases, so you’d expect their foresight to improve as more allowances are retired and prices increase.
Regarding carbon credits and Tradewater
I think it’s useful to distinguish between carbon credits and the climate actions they incentivize. Any climate action has to answer: “How much does it reduce atmospheric CO2?” But carbon credits must also answer: “How much future climate action does purchasing this credit incentivize?” That’s another hard question, and the impact of carbon credits is often unclear.
For example, Tradewater burns refrigerants and issues carbon credits, which they later sell. Even if burning refrigerants were perfectly additional, it’s unclear how much more refrigerant they will burn because you paid them $21 for a carbon credit. Similarly, someone might retire CO2 allowances and issue carbon credits to sell later. Even if retiring allowances were perfectly additional, you can’t know how many more allowances that person will retire because you bought a carbon credit from them.
Regarding policy, systems change, and Making Climate Policy Work
Cullenward and Victor emphasize the critical role of political feasibility. Carbon pricing programs (CO2 taxes + cap-and-trade) are politically costly and consequently haven’t been very useful, whereas targeted regulations (e.g. car emissions standards) have quietly done the most to decarbonize society.
They do acknowledge that a global cap-and-trade program with allowance prices matching the social cost of carbon would efficiently eliminate harmful carbon emissions. It’s just that it’s infeasible.
There is a role for cap-and-trade programs, only smaller than previously thought. To be effective in driving decarbonization, they need higher allowance prices. Their policy recommendation is to stop overallocating allowances.
You can view retiring allowances as implementing this policy recommendation, only without the politics.
Thanks again for engaging. I will really appreciate hearing your further thoughts.
Paco