Time to Unleash the Carbon Market?
What’s a ton of carbon (dioxide equivalent) worth? Not much if you ask the world’s carbon markets. The graph below summarizes prices and quantities covered by existing carbon emissions trading programs (green) and carbon taxes (blue). Nearly all carbon market prices are below $13/ton.
Source: State and trends of carbon pricing 2015
These low carbon prices have been making headlines, particularly in the context of the two largest carbon trading programs: Europe’s Emissions Trading Scheme (ETS) and California’s GHG emissions trading program. In California last month, the carbon allowance auction price hit the floor of $12.73, with only 11 percent of the 77.75 million allowances up for sale finding a willing buyer. European allowance prices have averaged around €6 in the first half of 2016, far below the €30 or more needed to encourage a shift away from coal-fired electricity generation.
The European lawmaker in charge of overhauling the EU ETS post-2020 has compared his carbon market without a price to “ a car without an engine”. I would spin this automotive analogy a little differently. As far as I can tell, these carbon markets have a working engine. We’re just not allowing them enough room to drive.
Here in California, the GHG emissions trading program (covering 85 percent of the state’s GHG emissions) has been cast in a supporting role. The updated scoping plan projects that over 70 percent of emissions abatement required under the 2020 target will be driven by “complementary measures” (e.g. mandated investments in low carbon technologies) rather than the permit price. Once you factor in offsets and the potential for emissions leakage and reshuffling, there’s not much work left for the carbon market to do.
In Europe the story is a little different. Because the ETS is less comprehensive (covering approximately 45 percent of emissions), many complementary measures are designed to tap abatement potential that lies outside the reach of the carbon market. But there are also important prescriptive measures mandating emissions reductions that fall within the scope of EU ETS. To put this into some perspective, the value of interventions (i.e. subsidies, feed in tariffs, etc. ) designed to accelerate investments in renewable energy has significantly exceeded the market value of emissions allowances in recent years (thanks Carolyn Fischer for highlighting this fact).
Prescriptive policies come at a cost
This preference for using prescriptive policies –rather than market mechanisms- to coordinate abatement helps explain why carbon prices are so low. Some simple graphs summarize the basics behind this cause and effect.
In the cartoon graph below, each colored block represents a different abatement activity (e.g. coal-to-gas fuel switching, renewable energy investments, energy efficiency improvements, etc.). Think Sesame Street meets the McKinsey curve. The width of the block measures achievable emissions reductions. The height of the blocks measures the cost per ton of emissions reduced.
In this cartoon cap-and-trade story, suppose baseline emissions are 200 and policy makers are seeking a 25% reduction. If we rely entirely on a permit market to get us there, we’d allocate 150 permits and let the market figure out where the 50 units of abatement will come from. An efficient market would drive investment in the lowest cost options: A + B + 1/2 C. The total abatement cost incurred to meet the target would be (20 x $10) + (20 x $20) + (10X$50) = $1100. The market clearing price (and the marginal abatement cost/ton) would be $50.
Now imagine that, in addition to the permit market, complimentary measures are introduced to mandate deployment of options D and E. These mandates take us 80% of the way towards meeting the emissions target. The role of the carbon market has been seriously diminished – we need only 10 more units of abatement to hit the target.
Under this scenario, the carbon market will incentivize investment in 10 units of A. The permit price drops to $10. The total cost of meeting the emissions target rises to 10 x $10 + 20 x $100 + 20 x $150 = $5100. And we wring our hands about low carbon prices and broken carbon markets.
Of course, this cartoon picture omits lots of real-world complexities (see this important EI paper for a more detailed analysis of California’s abatement supply and allowance demand). But it illustrates two real-world considerations. First, when complementary measures mandate relatively expensive abatement options, the carbon price we observe in the market will not reflect the marginal cost of reducing emissions. Second, a reliance on complementary measures to reduce emissions can significantly drive up the costs of hitting a given emissions target.
In California and in Europe, there is growing evidence that low allowance prices in the carbon market belie much higher abatement costs associated with complimentary policies. For example, this paper estimates that the California Solar Initiative delivered emissions reductions at a cost of $130 – $196 per metric ton of CO2. California’s LCFS credit price (which reflects the marginal incentive to reduce a ton of MCO2e) is currently averaging around $120 per metric ton CO2. In Europe, researchers estimate that the implicit costs of renewable energy targets per metric ton of CO2 are on the order of hundreds of euros for solar (and wind in some locations).
Time to unleash the carbon markets?
Looking out past 2020, more ambitious targets are being set and the process of charting a course to meet these targets is now underway. This could be a turning point for carbon markets. How heavily are we going to lean on prescriptive policies versus carbon markets to meet these future emissions abatement goals? If the increased stringency of future emissions targets is met with increasingly aggressive mandates and measures, we may be signing up for another round of low carbon prices.
I’m not suggesting we should leave *all* of the driving to the carbon markets. There are good reasons for complementing carbon markets with some truly complementary policies and mandates (some of which are fleshed out here). But there are also costs associated with keeping the carbon market mechanism on a tight leash while chasing emissions reductions with prescriptive mandates and programs (see, for example, Jim’s recent post here). Right now, carbon markets are hamstrung by a growing medley /cacophony of policies that drive allowance prices down. If we want to see carbon markets really work, we need to give them more work to do.
Categories
Energy Institute Blog 
Markets are a not a fact of nature, but a social construct that is constrained by what people find acceptable. The political process allows for policies with an implicit carbon price of $200 per tonne to be enacted, but would likely not allow an explicit $200 per tonne price.
The cuts needed to forestall dangerous climate change are well in excess of where the cartoon abatement curve ends (i.e., the curve ends at 50% abatement; we actually may need close to 100% abatement). The marginal cost of such abatement is likely well in excess of $200.
I would argue that the policies mentioned (e.g., the LCSF and to a lesser extent the CSI) are attempts to create markets where none exist, and to incentivize technological innovation that pushes the abatement curve to the left. Given the hard-to-quantify benefits of those things, it may be alright that California is willing to spend $200 per tonne on them.
The problem is that some of the carve outs for these compliance measures are arbitrary, and more importantly, aren’t abandoned by policy makers when they run into severe headwinds. The LCFS is one example of a poorly designed program that needs to be reset, and probably would benefit from being folded into the CATP now that transportation fuels are included. The HSR is the top example of how this approach can fail badly and lead us down the wrong path. On the other hand, the CSI seems to be more successful than we envisioned, and it may end up being very low cost or even beneficial in the long run.
Which brings up the truly appropriate way to measure these costs, using long-term estimates that include technological innovation, and testing for vulnerabilities. This probably requires conducting a true portfolio study, which the ARB has not done to date.
For any items that CA spent $200/tonne on, they might have been able to spend it on 4 x (abatement + innovation) items that cost $50/tonne, or 20 x the items that cost $10/tonne. How much of the $150 or $190/tonne difference drives innovation and how much of it is just getting handed out as excess rents?
Thanks for the illustrated example. Broken link: “(see, for example, Jim’s recent post here)”
It seems for the carbon tax case, complementary policy means a lower tax level is required, which could be an advantage for initial perception and implementation [but more expensive in total cost]. The tax would then be ramped up if we decide to rely less on complementary policies.
What I take from this, is that the real problem with carbon markets is the emphasis on defining them as markets. A market is a medium where willing buyers and willing sellers meet to exchange finite goods free of coercion. Carbon markets fail every aspect of that definition. There is no willing seller or willing buyer only coercion to exchange a good nobody wants. The good itself is not even a good and it is infinite in nature. It’s a rube goldbergian tax for sure, but not a market.
I’ve said the same thing about healthcare markets…
More on point, markets don’t need “willing” buyers and sellers–only those who want to transact. CATP market participants can self supply–as they very well did in the latest auction round where only about 10% of the allowances were sold. You’re confusing the need to comply with emission reduction levels with the functioning of the market, which is separate. The CATP is a simply another compliance mechanism, like the RPS.
Some basics: the “willing” and “free of coercion” parts are not necessary conditions of “markets”. The good, CO2 abatement, is not a standard good, but a [global] public good. https://en.wikipedia.org/wiki/Market_(economics)
I have a question on assumptions re cost of GHG reductions using solar. My understanding is that GHG reductions from solar in actual use, including GHG for solar manufacture and GHG from running backup gas inefficiently, is that solar comes in around 1/4 of whatever the gas emissions are. Then the gas + solar (or wind) needs to be compared to the GHG emissions of the electricity supply at times the sun is on, or else to the gas which would have been used instead.
What are good assumptions made re GHG emissions of 1 kWh of solar? I saw that the paper you linked to compared to average emissions when the sun is on. However, unless backup is hydro, does it make sense to include only GHG emissions from manufacturing PV?
This is a very interesting post that raises important questions about how to balance a carbon market versus other policies. A few thoughts:
1. A key reason carbon prices are low is a lack of political will. The complementary policies certainly have a large impact in keeping prices low, particularly in California. But carbon prices in many markets (especially RGGI and the EU) are low primarily because the emissions cap is too high. Understanding the cause of low prices (low political will versus complementary policies) is empirically difficult but a necessary first step to understand how to balance a carbon price versus other policies.
2. Looking at the marginal cost of abatement is not only way to judge the carbon cost effectiveness of a policy. The overall policy cost is also important. In the example from the article, the total carbon price cost would actually be relatively high: $1,100 (the abatement cost for the reduced emissions) plus $50/ton * 150 tons = $7,500 (the cost for non-reduced CO2). At $8,600, this total policy cost is actually higher than the ‘complementary policies approach’ at $6,600 ($5,100 for the approach plus $10/ton*150 tons =$1,500 for the price). There is a reason that policy makers so heavily favor ‘command and control’ regulations – marginal abatement costs may be higher, but the overall policy cost for getting a certain amount of emissions reductions could actually be much lower. Of course, there is a lot more complexity behind this (particularly around permit allocations) but carbon pricing is not necessarily the “cheapest” way to reduce emissions.
3. Comparing the cost effectiveness of different policy approaches is difficult because of changing technology and costs over time. Sure, the ITC seems a really inefficient way to reduce carbon emissions when compared with the lowest hanging fruit. But the ITC causes the solar industry to grow and reduce costs, making it possible for solar to be cost competitive without subsidies in the future. These types of market effects are not captured in a simple marginal abatement cost comparison but are critical for long-term policy making. Marginal abatement costs can change significantly over time (heavily influenced by policy) – analysis of carbon pricing versus alternatives should reflect these and other direct policy impacts.
The math in your example in 2. is incorrect. There is not an added $50/ton charge on top of the inframarginal $1100 expenditures. So carbon pricing always is the cheapest solution.
And having been in on these discussions, none of the regulators are trying to claim that command and control is cheaper than the market solution.
Your right about subsidizing new innovations to get to being market ready. But that hasn’t really been the prime focus of these policies. It’s largely about the “guarantees” that the lawyers and environmental scientists want–they are simply uncomfortable leaving it the ambiguity of “market mechanisms.”
I should have been more clear in the example in 2 because it is a critically important point that is often left out of the discussion. The example used in this article is focusing solely on the marginal abatement costs of the two potential policy solutions. The marginal abatement cost is only the actual money spent to reduce emissions. It does not necessarily include costs resulting from the functioning of the policy itself. Simply put, carbon prices (either through taxes or C&T) work by levying a price on the cost of emitting carbon. Accounting for the policy cost should include the costs on carbon itself.
Its easier to illustrate with a tax. In a tax scheme, all carbon emissions have a cost equal to the carbon price. In this case, that is $50/ton. Companies will only reduce emissions if it is cheaper to do so than pay the $50/ton tax. The $1,100 estimate in this article is based only on the costs from reducing emissions and thus ignores the costs incurred from paying to emit CO2. The $7,500 in my example refers specifically to the costs of emitting CO2 under a carbon price. Excluding these policy costs (which do not directly factor into marginal abatement costs) paints a misleading picture of carbon pricing.
All of that said, things get weird with C&T systems because the permit allocation and price systems become important. There are different policy costs (separate from marginal abatement) depending on whether permits are given away freely, whether they are auctioned, and whether large amounts of trading is occurring (and at what price). Nevertheless, I would hazard that the non-abatement policy costs in a C&T system will be much closer to the carbon price than to $0.
I would certainly doubt that regulators believe that but I would not be surprised if politicians at least implicitly do. Carbon pricing goes nowhere because the policy cost required to unlock the cheapest marginal abatement sources could be astronomical. Why increase energy prices by X% for everyone to induce a relatively small amount of CO2 emissions, when a subsidy can do the same thing for a lower absolute amount (even if it is by pursuing higher marginal abatement cost opportunities)?
I still don’t think your example works. This is the supply curve, not the market solution. The question in that calculation is whether all transactions clear at the CATP price or carbon tax, or if bilateral transactions set the price. The curve doesn’t address those assumptions. It simply illustrates how the supply curve functions.
I wrote about issues of CATP market design here, and reference a paper I wrote in 1996 on these issues: https://mcubedecon.wordpress.com/2015/10/28/cap-trade-and-market-design/
Regarding 1., the above examples assumed a given emissions cap and therefore a fixed required amount of reductions regardless of complementary policy. Complementary policy will reduce the price since their effect loosens the cap.
So carbon prices being “too low” due to a “high cap” or extra credits etc due to political will seems like a question that can be separated.
The linkage would be that if when setting the cap, policymakers anticipated the impact of complementary policy and included it in their baselines and cap calculations.
Regarding 2., if you do want to cut all the remaining CO2, you need to vary the price for all the remaining CO2 according to the abatement supply curve. The total costs, in that case, seem to be equal (but we are nowhere near reductions of 100%). You might be referring to “policy cost” being lower e.g. the costs of regulation like CAFE are invisible , but economically, the costs are there.
Regarding 3., I agree such effects exist and should be reflected in analysis, but with careful and conservative assumptions.
Put on a carbon tax to raise the value of CO2.
so in case 1, without complementary policies, a carbon tax of $50/tonne is needed to reduce 50 tonnes.
in case 2, with the complementary policies, a carbon tax of $10/tonne is needed to reduce the remaining 10 tonnes (but with a high total policy cost)
“The updated scoping plan projects that over 70 percent of emissions abatement required under the 2020 target will be driven by “complementary measures” (e.g. mandated investments in low carbon technologies) rather than the permit price. Once you factor in offsets and the potential for emissions leakage and reshuffling, there’s not much work left for the carbon market to do.”
We need to stop humoring the euphemism of calling command & control climate policies “complimentary” to carbon markets. A more accurate description would be “redundant.” Unless the policy is addressing a market failure other than unpriced externalities, or a addressing a “behavioral failure”, there’s absolutely zero justification for these redundant policies.
I agree. In California, both the AB 32 and RECLAIM programs came up with “market solutions” only as add ons when the command and control programs fell short of feasible ideas.
We pointed out this problem in our report on behalf of EDF in the 2008 Scoping Plan proceeding. (EDF’s press release and letter doesn’t quite get right the content of our report.) It’s a bit misleading saying “Here in California, the GHG emissions trading program (covering 85 percent of the state’s GHG emissions)…” In fact, it was only 15% of emission reductions in the 2008 plan. And the CTP was added simply because the ARB staff couldn’t come up with more prescribed reductions on their own. Unfortunately, government employees are where they are because they are risk averse, including to markets, so they are prejudiced against solutions that leave decisions to individuals. (I’ve been inside enough to know that this is universally true in my professional experience.) The result that we are handed some of the most expensive solutions rather than the most effective.