A critique of “Do Environmental Markets Cause Environmental Injustice? Evidence from California’s Carbon Market,” a 2020 NBER working paper by Danae Hernández-Cortés and Kyle C. Meng

Joint post by Danny Cullenward and Katie Valenzuela

---

Danny Cullenward a Lecturer and Affiliate Fellow at Stanford Law School, where he teaches courses on energy and climate law. He is also the Policy Director at CarbonPlan and the California Senate’s appointee to the Independent Emissions Market Advisory Committee, which is charged with reviewing California’s cap-and-trade program. Danny holds a JD from Stanford Law School and a PhD in Environment and Resources from Stanford University. He can be reached at dcullenward@ghgpolicy.org.

Katie Valenzuela is an environmental justice advocate based in Sacramento. She previously served as Policy & Political Director for the California Environmental Justice Alliance, Principal Consultant for the Joint Legislative Committee on Climate Change Policies, and Co-Chair of California’s AB 32 Environmental Justice Advisory Committee. She can be reached at kbvale@gmail.com.

---

When this working paper was released, it was our ardent hope that the technical flaws and clear research bias would prevent it from being used to inform deliberations on climate policy. As time has passed, however, and it has become apparent that this paper is being used to undermine legitimate critiques of carbon pricing programs like California’s cap-and-trade program, we feel it is our responsibility to highlight the errors in method and assumptions present in this paper.

Below, we review five key errors in the Hernández-Cortés and Meng analysis. These errors demonstrate that the paper’s findings cannot be relied upon in the California context nor as an indicator of likely outcomes in other jurisdictions that are considering similar climate policy strategies. 

First, a caveat: we fully understand that greenhouse gas emissions trends do not necessarily correlate perfectly with the emission of other co-pollutants and that optimal strategies for reducing climate and local air pollution might be very different from one another. However, no one can argue that greenhouse gases and the co-pollutants that harm public health do not come from the same sources. This is why investments that reduce greenhouse gas emissions hold significant potential to reduce co-pollutants and why environmental justice movements focus on reducing emissions at the source as much as feasible.

Some will argue that, while this particular paper may not be accurate, there is still no evidence that cap-and-trade harms environmental justice communities. We would argue, however, that simply not harming environmental justice communities was never the goal of California’s landmark climate law, AB 32. The goal of this law, as clearly outlined in the text and in subsequent legislation, was to “[c]onsider the potential for direct, indirect, and cumulative emission impacts from these mechanisms, including localized impacts in communities that are already adversely impacted by air pollution.” In that respect, the cap-and-trade program has fundamentally failed to protect communities burdened by pollution by failing to ensure that facilities located in environmental justice communities reduce emissions. California can — and must — do better. Any state or country seeking to follow our example should heed this valuable lesson: low carbon prices are not sufficient to make a large impact on either greenhouse gas emissions or local air quality.

Error 1: Inaccurate definition of environmental justice communities

The working paper identifies environmental justice communities by zip code, not by census tract — a critical mistake that creates a fundamental flaw for the entire analysis that follows. For example, in CalEnviroScreen 3.0 — the state tool used to identify environmental justice communities — Sacramento zip code 95818 includes three census tracts: 6067002200, 6067002300, and 6067002400. Census tract 6067002200 (shown in red) is in the highest 10% of CalEnviroScreen scores in the state, representing significant pollution burden and social risks. The other two census tracts (shown in green) are in the lowest 30-40% brackets of scores, meaning they have significantly lower environmental burdens and social risks. 

Source: CalEnviroScreen 3.0

Analysis of what community is an environmental justice community at the zip code level blunts the disparities that exist at times within zip codes, which is why California opted to use census tracts rather than zip codes for analyzing environmental justice impacts.

Error 2: Data quality issues

There are important but poorly understood technical problems with the paper’s underlying pollution data, which come from 35 different regional air districts that use 35 different methods for data collection. Some of those methods are not good or reliable, which is part of the reason a major statewide effort is underway to enhance and standardize monitoring, particularly in environmental justice communities. The data set used for this working paper comes with explicit government caveats about the process of air pollution permit updating and potential incompatibility between districts' reported data and facilities' reported data. 

At the same time, Table S2 shows that the authors' modeled pollution concentrations do not correlate well with EPA's observed data. Across four different local pollutants the correlation between modeled “HYSPLIT” and observed ambient pollution exposure data is very weak — statistically different from zero, but hardly a reliable predictor of what actual pollution monitoring data show.

Source: Hernández-Cortés and Meng (2020)

It should also be noted that industrial sector greenhouse gas emissions have been roughly constant since 2006 (the year AB 32 became law). As illustrated in the chart below, nearly all of the state’s emission reductions have come from the electrical power sector, not the industrial sector. 

It is curious that the authors claim progress in industrial emissions under cap-and-trade when California’s own data shows that notion to be false with respect to greenhouse gases. It is possible that industrial greenhouse gas emissions would be higher (counterfactually) without the cap-and-trade program and with no other policy alternatives. But even if the cap-and-trade program has prevented an increase in emissions relative to having no program at all, it hasn’t substantially reduced total emissions. Any suggestion that maintaining air pollution levels in some of the country’s most polluted communities is an environmental justice success story strains credulity.

Source: California greenhouse gas emissions inventory

Error 3: Control group bias

One of the more remarkable methodological problems with this paper is that the authors’ "treatment" and synthetic "control" groups are not at all similar. This is a major red flag because the authors use a "difference in differences" statistical model, which assumes that that the two groups experience identical trends before the policy applies and therefore that the observed trend in the synthetic "control" group (non-C&T regulated facilities) can predict the counterfactual baseline for the treatment group (C&T regulated facilities). If constructed correctly, the counterfactual baseline can then be used to estimate the impacts of the cap-and-trade policy on the treatment group. 

Whether a facility is included or excluded in cap-and-trade is based on a standard emissions threshold (25,000 tCO2e/year) that is intended to pick up nearly all of the sources in some sectors while covering only some of the sources in others. Meanwhile, the control group includes all facilities that are large enough to be required to report emissions (greater than 10,000 tCO2e/year) but smaller than the cap-and-trade coverage threshold. As a result, the control group has almost no representation from the refining and hydrogen sectors, most of which are individually large sources (>25,000 tCO2e) and which collectively represent a major driver of local air pollution in many environmental justice communities. If any of the sectors that are not well distributed between the "treatment" and synthetic "control" groups and experience divergent trends, then the "differences in differences" method could wrongly attribute the effects of sector-specific differences to the cap-and-trade policy. 

You can see from Table S1 that the authors’ data approach has grouped nearly all of the state’s hydrogen and refinery sources into the treatment group (C&T regulated facilities), without any significant representation from these sectors in the control group (non-C&T regulated facilities). This means that any policies or exogenous trends that have an important impact on the hydrogen and refining sectors will be wrongly interpreted by the authors’ model as an impact caused by the cap-and-trade program.

Source: Hernández-Cortés and Meng (2020)

Error 4: Omitted variable bias

The control group bias problem compounds the authors' assumption that no other policies affect their treatment group, when in fact at least one major policy overlaps the time period and treatment groups the authors address in their model. This is called omitted variable bias. The authors incorrectly assume that no other policies disproportionately affect their treatment group:

We use the facility eligibility criteria and the 2013 introduction of the GHG C&T program to isolate stationary, facility-level, local air pollution emissions driven by the program.[5]

[5] CARB also developed other regulations to meet the AB 32 target, such as a Low Carbon Fuel Standard and Advanced Clean Car Standards. These programs, however, were not market-based, were mostly introduced prior to 2012, and did not have the same facility-level eligibility criteria as the cap-and-trade program. Thus, while our estimated C&T effects take place in the presence of these other regulations, it is unlikely that C&T effects are conflated with the effects of these other regulations.

That is factually incorrect. The Low Carbon Fuel Standard (LCFS) is a "market-based" program — it's quite literally another cap-and-trade program that applies specifically to transportation fuels, and includes fuels' associated emissions from refining and hydrogen production. The LCFS began to affect refinery and hydrogen sector emissions at almost the same time as cap-and-trade came into operation, and LCFS prices ramped up significantly to more than 10 times the cap-and-trade carbon price during the authors' study period. 

Source: CARB LCFS dashboard, CARB cap-and-trade auction data

Readers who have heard success stories about California’s cap-and-trade program might wonder at how a system that is so widely celebrated has produced such a modest carbon price signal. The answer is not at all surprising: the program has far too many allowances and carbon offsets in circulation to restrict pollution levels. This finding is now widely documented in the academic literature (e.g. Cullenward et al., 2019; Mastrandrea et al., 2020; Inman et al, 2020), but the working paper doesn’t cite any of these peer-reviewed articles.

Because the refinery and hydrogen sectors are nearly 100% clustered in the "treatment" group, a much stronger "signal" that applies to these same sources from the LCFS program is picked up by the authors' statistical model as being attributed to the statewide cap-and-trade program. (Again, there are very few small refinery players whose operations are affected by the LCFS program, so the LCFS effects manifest almost entirely at facilities that are subject to both cap-and-trade and LCFS—hence the effects of the LCFS are strongly correlated with the effects from cap-and-trade, contrary to the authors' assumptions.)

The working paper also failed to control for the effect of other overlapping policies with more direct emissions reductions than cap-and-trade. An example is the Renewable Portfolio Standard, which drove early procurement of zero emission power projects in solar and wind and dramatically reduced emissions from fossil fuel power plants through reduced demand. Attributing all reductions observed at refineries and power plants to the cap-and-trade program ignores the impacts of those stronger policies on emissions trends, giving the reader the false impression that the cap-and-trade program drove those reductions.

Error 5: The question is not cap-and-trade versus nothing

Finally, the paper asks whether cap-and-trade has an impact relative only to the modeled absence of cap-and-trade, but that's not the policy and political question in California — nor is it the right question to ask anywhere else. The real question is whether the program California has — which is massively oversupplied and therefore insufficient to drive GHG pollution down in line with California's climate laws (see here and here for an overview) — is enough. The better alternatives are (1) a stronger cap-and-trade program, (2) more direct regulations on polluters, and (3) more of both kinds of policy. It's not news that a flawed program is better than no program at all, but that's all the paper looks at.

If California had chosen a path of more prescriptive, direct emissions reductions in accordance with state law, we would likely be seeing far more emissions reductions at the source, and far more improvements for environmental justice communities than we're seeing under the cap-and-trade program today. To say it's better than nothing ignores the fact that adopting a weak cap-and-trade program has led to prolonged and higher emissions in environmental justice communities than if California had adopted a stringent carbon pricing policy or relied instead on non-market mechanisms that would have been targeted at the pollution reductions our communities need.

CalMatters oped: It’s time to fix California’s cap-and-trade – again

CalMatters published my oped on last week’s problematic cap-and-trade auction:

Successful auctions normally bring in between $600 million and $800 million per quarter to support a wide variety of air quality, climate and fire protection programs, but now – in the middle a massive budget crisis – those funds are gone. So how did a program touted as the leading approach to climate policy produce yet another crisis?

Simple: the program has too many allowances. Every year to date, the Air Resources Board has made more allowances available to market participants than there are climate emissions. That has led to a growing private “bank” of unused allowances that depress market prices and risk revenue shortfalls at quarterly auctions.

KQED: California’s Cap-and-Trade Program Generates Severely Reduced Revenue

KQED’s Kevin Stark spoke with me in connection with the dismal cap-and-trade auction results announced earlier today and their implications for the state’s climate budget. Here’s Kevin’s story and my quote:

But an undersold market could also signal that the state’s market is bloated with more allowances than are necessary, given the paucity of economic activity in the state.

“It's basically the worst-case result,” said Danny Cullenward, an energy economist and lecturer at Stanford University, referring to the latest auction. “Almost no revenue is coming into the state.”

“The bad outcome will get blamed on COVID but is much more the product of a weak program design that was vulnerable to surprises, with way too many allowances,” he said.

Kevin also quotes Stanford’s Chris Field, who suggests that the auction shortfall is just a COVID-related blip and that policymakers should wait and see before considering any program reforms.

It’s disappointing to see my colleague deflect blame for the bad auction because smart cap-and-trade programs with good market designs aren’t experiencing revenue crises—just look to the European Union or Regional Greenhouse Gas Initiative to see how policymakers can design programs that respond to macroeconomic surprises and produce predictable outcomes. Both the EU and RGGI programs have rule-based adjustments that automatically adjust market supplies to make the market tighter or looser in response to macroeconomic uncertainty. California does not. That’s the story.

The COVID argument doesn’t hold any water in numerical terms, either.

A new paper from the Global Carbon Project team in Nature Climate Change estimates that a national quarantine in the United States through July followed by prohibitions on non-essential travel and large gatherings through December would produce an annual GHG reduction of about 11.5% (5.4% to 18.5% CI) from 2019 to 2020 (Le Quéré et al., 2020: Table S14). California’s cap-and-trade emissions were 319.9 MMtCO2e in 2018, so a reduction of 11.5% normalizes to about 37 MMtCO2e (with 5.4% equal to 17 MMtCO2e and 18.5% equal to 59 MMtCO2e). Meanwhile, California is lifting its stay-at-home orders earlier than this scenario assumes, so these numbers are likely to be an upper-bound, not a central estimate.

In contrast, cap-and-trade oversupply clocked in at about 227 MMtCO2e in 2018, the year with the most recently available data (Cullenward et al. 2019). When we get 2019 data this November, we’ll see that number go substantially higher, likely heading toward 300 MMtCO2e. So I’m not buying the argument that a shift in demand on the order of 37 MMtCO2e (17 to 59 MMtCO2e) was enough to destabilize an otherwise sound market when that market was already long by 227+ MMtCO2e and counting. The regulator’s own pre-COVID assumptions about future market conditions would have produced a slack and oversupplied market all the way through 2030 (Inman et al., 2020).

So if COVID-related emission reductions are temporary, then any transient effects are small in relation to the structural oversupply problems that have plagued the California market since its origins. (How quickly we forget that oversupply caused another revenue crisis just four years ago; see Cullenward & Coghlan, 2016.)

Alternatively, it’s entirely possible that long-term emissions trajectories drop as part of a deep or protracted recession—a possibility that smart economists have been warning California climate policymakers about for some time (Borenstein et al., 2019). But that’s an argument in favor of rule-based program adjustments, not a wait-and-see attitude.

The advantage of rule-based program adjustments is twofold. First, these approaches automatically apply if and only if conditions meet predetermined triggering conditions, so policymakers don’t have to wait for problems to get worse before they implement a solution. Second, their preferred adjustments won’t take effect until they want they want them to—policymakers can set the trigger however they like.

It’s high time California started putting these basic climate governance lessons to use. Long-term market prices remain below the price floor, raising the possibility that future cap-and-trade auctions will meet a similar fate. After all, the alternative to making reforms is putting California’s climate, air quality, water, and fire prevention programs at risk of ongoing budget shortfalls.

References

Borenstein et al. (2019), Expecting the Unexpected: Emissions Uncertainty and Environmental Market Design, American Economic Review 109(11): 3953-77, doi.org/10.1257/aer.20161218.

Cullenward & Coghlan (2016), Structural oversupply and credibility in California’s carbon market, Electricity Journal 29: 7-14, dx.doi.org/10.1016/j.tej.2016.06.006.

Cullenward et al. (2019), Tracking banking in the Western Climate Initiative cap-and-trade program, Environmental Research Letters 14: 124037, doi.org/10.1088/1748-9326/ab50df.

Inman et al. (2020), An open-source model of the Western Climate Initiative cap-and-trade programme with supply-demand scenarios to 2030, Climate Policy, doi.org/10.1080/14693062.2020.1760774.

Le Quéré et al. (2020), Temporary reduction in daily global CO2 emissions during the COVID-19 forced confinement, Nature Climate Change, doi.org/10.1038/s41558-020-0797-x.

Climatewire: Virus threatens to weaken California climate programs

E&E’s Anne Mulkern quotes me in a Climatewire story about current conditions in California’s carbon market, where prices have dropped significantly below the price floor. That’s a problem for state budgets because the government won’t sell new allowances at the upcoming May auction for less than the price floor, currently $16.68 per allowance in 2020. If secondary markets continue to trade at lower prices, then we should expect companies not to buy all of the allowances on offer in May’s auction—and that, in turn, could mean the state’s climate budget funds will take a significant hit.

We’ve seen this happen before:

Danny Cullenward, a Stanford researcher, said cap-and-trade auctions faltered in 2016 not long after spot market prices fell. In May that year, 90% of available allowances failed to sell, leaving at least $880 million in revenue out of reach, according to a paper he wrote for The Electricity Journal.

"A penny below the floor is significant," Cullenward said. "That is strongly indicative for the demand at the quarterly auctions to fall short."

The non-partisan Legislative Analyst’s Office has a great figure showing volatility in state climate proceeds from quarterly cap-and-trade auctions (see here):

The problem in today’s market is that the California Air Resources Board chose to keep the program oversupplied—possibly all the way through 2030—which means that prices were stuck near the price floor prior to the economic slowdown. Under those conditions, the wasn’t any room for markets to adjust to macroeconomic changes without risking demand destruction at future auctions.

For those interested in the data on the 2016 market crisis, my article with Andy Coghlan in The Electricity Journal is available here. For more on the oversupply outlook through 2030, see this preprint of a paper currently under peer review.

NPR: A Check-Up On California's Efforts To Combat Climate Change

NPR’s Hannah Hagemann spoke with me as part of a Weekend Edition Sunday story on California’s climate policies. As I told Hannah, all the good news is in the electricity sector, where the state’s clean energy policies have cut emissions dramatically over the last fifteen years. But California has struggled to cut emissions the industrial and transportation sectors. Not only are the relevant technologies lagging the cost declines in the wind and solar industries, but the intersection of transportation, housing, and affordability issues makes the challenge of controlling pollution from cars and trucks even more difficult. (It doesn’t help that the Trump Administration revoked California’s longstanding right to set its own pollution standards in this area, either—the outcome of that attack that will be resolved in the courtroom, if not overtaken by the 2020 election first.)

For a deeper dive into the state’s experience reaching its 2020 climate goal early, check out this new paper in Energy Policy, which I co-authored with Mike Mastrandrea and Mason Inman.

Climatewire: Cap-and-trade feud may chill cross-border pacts

E&E Climatewire reporter Jennifer Hijazi quotes me in a story on the federal government’s lawsuit against the link between cap-and-trade programs in California and Québec:

Cullenward said he expects that even if courts rule against the agreement, the case would not have a major impact on the individual cap-and-trade markets in California and Quebec. He noted that the individual programs can stand on their own — even in the absence of an agreement linking the markets.

"Setting aside the disruption and the uncertainty and the political vindictiveness of this lawsuit, there's not a lot about the markets that doesn't work in isolation if each of them were operating on their own," he said.

That’s not an accident, but rather a reflection of the fact that subnational governments can’t bind one another to legal commitments in linked markets. Under those conditions, the only realistic market links involve programs that are fairly similar to one another to begin with. Anything else risks major political and financial investments on the back of a legally unenforceable form of cooperation.

MIT Technology Review: Here’s how we could go carbon neutral in 25 years

James Temple quotes me in a story on a new LLNL/ClimateWorks report on California’s 2045 carbon neutrality goals. Although I like a lot in the report, I also have some concerns:

For one, Carnegie Institution researcher Danny Cullenward warns that the three major levers in the plan—land use, direct air capture, and hydrogen from biomass—are all “untested at scale.” He adds that the study may overestimate how cheap direct air capture will become and how much carbon the state’s forests and soils are likely to store. After all, in some recent years, including the major wildfire season of 2018, California’s forests have produced more carbon dioxide than they absorbed.

Cullenward also says reaching carbon neutrality presumes that California will achieve its goals in cutting emissions from electricity, transportation, and other sectors we know how to clean up. At the current rate of reductions, however, the state will miss its 2050 decarbonization target by a century.

There’s a lot to like in the new report, and I’m in complete agreement with lead author and LLNL senior scientist Roger Aines, who says now is the time to start working on negative emissions strategies.

But the report’s discussion of forest carbon isn’t sufficiently rigorous. It relies on an extremely optimistic study from researchers at The Nature Conservancy (Cameron et al., 2017, PNAS). My colleagues in the ecology community have repeatedly questioned how realistic these forest carbon scenarios are, and in any case the LLNL report doesn’t include climate feedbacks like drought, fire, and disease that make forest carbon storage a particularly challenging prospect in the Western U.S.

As a result, I’m inclined to write that part of the report off and focus on its main moving parts—negative emission technologies using biomass wastes from forest and agricultural systems, as well as novel direct air capture technologies. After all, these strategies produce about 100 of the 125 million ton CO2 per year negative emissions potential the report identifies for 2045, so they constitute the bulk of the study’s findings.

On these fronts the report provides a helpful roadmap for discussion. Not only will we need some or even all of these negative emissions technologies to work, but unlike most of the offsets-oriented discussion of forest carbon, these technologies could be deployed in California to deliver local jobs and economic benefits. Technology strategies with these features are much more realistic and robust—both in terms of the domestic prospects and the potential to inspire climate followers—so there’s a lot to like here.

Congrats to the LLNL team for an important contribution.

CalMatters: Clean up cars and cement industry to reach California’s climate goals, report says

CalMatters’ Rachel Becker quotes me in a story about a new report from Chris Busch and Robbie Orvis at Energy Innovation LLC on how to strengthen California’s 2030 climate strategy. In addition to calling for efforts to reform the struggling cap-and-trade program, the new report highlights potential for policy innovation in the often-overlooked industrial sector:

Cullenward called the proposals “illustrative of the kinds of new and complicated things that climate policy makers need to do to begin tackling the industrial sector.”

Policies that drive technologies capable of generating heat high enough for oil extraction, or producing steel and cement, could be useful globally, Cullenward said: “If we could successfully tackle that, it would show the way on a new chunk of emissions — where we could be a leader for others to follow.”

Perhaps the most impressive thing about the new report is that it’s the first major study to offer both a deep dive on the mitigation potential for individual regulatory policies and carbon pricing in a single, comprehensive analysis. By contrast, California’s 2017 Scoping Plan relies exclusively on analysis of regulatory policies and doesn’t incorporate economic feedbacks from the state’s carbon price, despite designating cap-and-trade to do nearly half of the work in getting to the state’s 2030 climate target.

Analyzing regulatory and carbon pricing policies is essential to developing and implementing a clear plan; Chris and Robbie deserve a lot of credit for putting them together in a single effort. The Energy Innovation report should set the methodological standard for the next state Scoping Plan, which is due in 2022.

CalMatters: California’s renewable energy targets slashed carbon pollution — now there’s a proposal to pause them

CalMatters’ Rachel Becker quotes me in a story on a great new report from Dr. Ross Brown at the California Legislative Analyst’s Office:

“Understanding why emissions fell — as well as what we don’t know — is critically important to the task of informing California’s pursuit of future climate targets,” Danny Cullenward, policy director at the climate change think-tank Near Zero, said in an email to CalMatters. He called this week’s report “the gold standard.” 

LAO’s annual climate policy report series is a must-read. This year’s report focused on electricity sector emissions, with the 2018 reports providing an overview of state climate policy and a deep dive on the transportation sector.

LA Times: Why California’s climate solution isn’t cutting it

Los Angeles Times contributing writer Jacques Leslie quotes me in an op-ed on the oil industry’s influence over California’s cap-and-trade program:

In the face of repeated requests by experts and journalists to show how it will deliver on the state’s 2030 climate law, CARB hasn’t provided a convincing road map or even acknowledged that a problem exists, according to Danny Cullenward, an energy economist and policy director of the climate policy think tank Near Zero. Meanwhile, the Western States Petroleum Assn., the oil industry trade group, has called on CARB to maintain its current course, as it stated in comments to the board on program amendments in 2018.

It’s time for state policymakers to have a serious conversation about whether cap-and-trade is going to drive California’s progress to its 2030 emissions limit, and if not, what they plan to do instead.

Politico: How California and Quebec are trying to Trump-proof their climate effort

Politico’s Jeremy B. White and Andy Blatchford quote me in a story about Québécois Premier François Legault’s visit with Governor Newsom to discuss the two governments’ cap-and-trade partnership, which the U.S. Department of Justice has challenged in court:

“What really matters is how you get others to follow, and cooperation between subnational governments like California and Quebec is often seen as a way to do that,” said Danny Cullenward, a Stanford economist who helps oversee California’s cap-and-trade program.

But that effort to generate some momentum doesn’t guarantee a durable result, Cullenward said, given that it’s unenforceable.

“It’s really just a political commitment that’s been written down,” Cullenward said. “It’s not a commitment that a party could take to a court in another jurisdiction and enforce.”

ProPublica: Cap and Trade Is Supposed to Solve Climate Change, but Oil and Gas Company Emissions Are Up

ProPublica’s Lisa Song quotes me in a comprehensive story about the challenges with California’s cap-and-trade program:

Countries have called California’s cap-and-trade program the answer to climate change. But it is just as vulnerable to lobbying as any other legislation. The result: The state’s biggest oil and gas companies have actually polluted more since it started.

[….]

“There’s no question a well-designed regulation on oil and gas can have an effect,” said Danny Cullenward, a Stanford researcher and policy director at Near Zero, a climate policy think tank. “And that was traded away for a weak cap-and-trade program.”

[….]

Cullenward said the reality of society’s reliance on fossil fuels leaves regulators in a bind — stuck between knowing what it will take to manage climate change but adopting a market solution that’s too weak by design. Any attempt to seriously strengthen it risks a consumer backlash that makes it “politically toxic,” he said.

“Essentially, what we’re doing is kicking the can,” Cullenward said. “We’re saying, keep prices low, let’s not do a lot, and later, we’ll hope for a miracle.”

Julie Cart (CALmatters): Big polluters get help from the state, renewing doubts about California’s climate goals

CALmatters’ Julie Cart reports on the December 2018 Air Resources Board hearing, where the Board adopted its flawed post-2020 cap-and-trade regulation without confronting the program’s obvious lack of stringency or even analyzing its sufficiency under state climate law:

Danny Cullenward, an economist with the climate-change think tank Near Zero and a member of the advisory committee, said the air board’s position is based on a math error.

“The analytical integrity of what the staff has put forward does not meet the standard set by state law,” Cullenward said.

[Air Resources Board Chair Mary] Nichols acknowledged Thursday that the issue “sticks in the craw” of those who continue to question whether the board has sufficiently addressed it.

“It hasn’t been put to rest; let’s put it that way,” she said.

Although the Board did not respond to the public comments on this issue at its December meeting, the Board decided to instruct staff to review the issue of allowance oversupply in 2019 and to consider potential adjustments to the market design as needed. No public text of the relevant document (Board Resolution 18-51) is yet available, but I grabbed a printed copy of the draft resolution at the hearing. [January update: the text of Resolution 18-51 is now online and matches the proposed text below.]

Here’s the key instruction from the draft resolution:

BE IT FURTHER RESOLVED that the Board directs the Executive Officer to quantify and report back to the Board, by no later than December 31, 2021, the volume of unused allowances from 2013 through 2020, including volumes held in private accounts, and the potential for unused allowances to hinder the ability of the program to help achieve the SB 32 target. The Executive Officer shall hold a public workshop in 2019 to discuss potential methodologies to evaluate this topic.

This is an important step forward, but it’s worth emphasizing that the evidence of a problem is crystal clear and part of the Board’s own record (see, for example, the nearly 500 pages of analysis Near Zero submitted into the formal 45-day comment period). Despite an extensive formal record, however, the Board refused to analyze the stringency of its post-2020 program caps or even acknowledge independent analysts’ concerns. So while I’m very glad that the Board has agreed to a process for revisiting its questionable analysis, the ARB staff’s willingness to consider facts and evidence will need to improve if this process is to deliver meaningful progress.

Dave Roberts (Vox): California's cap-and-trade system may be too weak to do its job

Vox’s Dave Roberts has a long story up on the failure of the California Air Resources Board to address allowance oversupply in the state’s cap-and-trade program:

Concerns about oversupply in the California system have been raised for years, dating back at least to May 2016, when allowance prices collapsed and 90 percent of allowances went unsold. Later that year, Danny Cullenward, policy director at the think tank Near Zero, and Andy Coghlan of UC Berkeley published a paper in The Electricity Journal warning of “structural oversupply.”

The Carbon Market Supply Association warned CARB in September 2016 that oversupply was “as high as 300 million tons.” Warnings were echoed by the LAO (in February 2017 and again in December 2017), Chris Busch at the energy policy firm Energy Innovation in December 2017, the Environmental Commissioner of Ontario (in a January 2018 report contemplating that province’s own carbon pricing system), Near Zero in March 2018 (and many times since; see here for a synthesis), and the Independent Emissions Market Advisory Committee in October 2018.

The entire story is well worth a read.

California cap-and-trade research updates

Near Zero Research Notes (with Mason Inman and Michael Mastrandrea):

• An open-source model of supply and demand in the Western Climate Initiative cap-and-trade program (Oct. 10, 2018)

• Tracking banking in the Western Climate Initiative cap-and-trade program (Sept. 12, 2018)

• Ontario’s exit exacerbates allowance overallocation in the Western Climate Initiative cap-and-trade program (July 16, 2018)

• California’s “self-correcting” cap-and-trade auction mechanism does not eliminate market overallocation (May 23, 2018)

• Ready, fire, aim: ARB’s overallocation report misses its target (May 7, 2018)

• Implementing AB 398: ARB’s initial market design and “allowance pool” concepts (Mar. 16, 2018)

• Interpreting AB 398’s carbon offsets limits (Mar. 15, 2018).

• Removing excess cap-and-trade allowances will reduce greenhouse gas emissions (Jan. 1, 2018)

• California’s climate emissions are falling, but cap-and-trade is not the cause (Nov. 10, 2017)

Will FERC uphold state support for clean energy?

My friend Shelley Welton and I have a guest op-ed at Utility Dive: 

The energy law community is abuzz with the news that the Federal Energy Regulatory Commission (FERC) weighed in on a high-profile case challenging a state-level support program for nuclear generators. In a May 29 amicus brief joined by the U.S. Government, FERC argues that the Seventh Circuit should not resort to the “extraordinary and blunt” remedy of preempting Illinois’ nuclear subsidy program.

Some are now celebrating the brief’s apparent defense of state environmental policies, but that reaction is premature. Rather than defend state programs, FERC has asked the court to let it decide what’s right. If the Commission follows its recent practice in approving capacity market reforms that shift costs to generators that receive state environmental policy support, that could spell bad news for state-supported clean energy.

Update: Shelley and I published our longer analysis at the Yale Journal on Regulation Bulletin.

New comment letter: dormant commerce clause risks and regional GHG accounting in CAISO EIM

CAISO continues to refine its approach to greenhouse gas accounting in its regional Energy Imbalance Market. In a revised February 2018 proposal, CAISO announced its decision to move away from the pilot two-pass approach to calculating net GHG emissions after "secondary dispatch" (i.e., resource shuffling to serve California load) and suggested a new approach for attributing net GHG emissions in the regional electricity market. 

Although the challenges CAISO identified with the two-pass solution are serious and indicate the need for fresh thinking, CAISO's new proposal would put the costs of managing leakage only on out-of-state zero-carbon resources (wind, solar, and hydro). As I point out in a new comment letter, not only does this approach ignore the leakage risks from replacing high-carbon fossil resources with low-carbon fossil resources (i.e., replacing coal-fired electricity with power from natural gas), but the outcome would mean that the GHG costs attributable to zero-carbon resources will depend on whether those resources are located in California (in which case they face zero GHG costs) or outside of California (in which case they would face a positive GHG cost under the CAISO proposal).

In my view, CAISO's proposal increases the risks of a dormant commerce clause challenge. This should cause some concern as the state government of Utah is actively contemplating litigation behalf of its coal power plants (E&E News subscription req'd). While CAISO's proposal to exempt fossil resources from carrying the costs of GHG leakage might avoid potential legal challenges from coal resources, it wouldn't protect California's climate policies from a challenge brought by zero-carbon resources (such as wind, solar, or hydro) that want to sell their power to CAISO territory on equal terms with in-state competitors. 

The good news is that CAISO has other options for incorporating California's climate policies in the EIM tariff design without increasing legal risks. For example, CAISO and the Air Resources Board could update and refine the situation in place today. Currently, ARB intends to retire a certain number of allowances based on a coarse estimate of the regional GHG leakage from "secondary dispatch" (as CAISO calls it, also known as resource shuffling in other discussions). These leakage estimates could be improved by CAISO's rigorous calculation of the net GHG emissions attributable to EIM deliveries to California territory. ARB could then account for leakage in the cap-and-trade program by retiring an equal number of allowances to compensate for the leakage CAISO observes. This approach would preserve the environmental integrity of California's carbon market without favoring in-state resources over comparable out-of-state facilities because retiring allowances places raises carbon prices equally for all entities subject to the cap-and-trade program, independent of the location of the power plant that exports electricity to serve California load.  

CAISO has been a leader on integrating state-level carbon pricing into regional electricity market design, successfully implementing the country's first FERC-approved tariff to include an explicit carbon price in the EIM. I hope CAISO will continue to demonstrate climate leadership by carefully evaluating the dormant commerce clause risks of any modifications to the current EIM tariff structure and incorporating these issues into their discussions with ARB.

Harvard Electricity Law Initiative workshop report

Ari Peskoe just released a workshop report from a recent meeting hosted by the Harvard Electricity Law Initiative called Regulatory Paths Forward for a Cleaner Grid. I was glad to join Ari and the other workshop participants to talk about California's approach to resource adequacy and some of the jurisdictional tensions that arise in regional energy markets. 

Media: CALmatters on cap-and-trade spending

Julie Cart at CALmatters has a follow-up story on how the issue of extra free allocation to industry is playing out in the state legislature. Last week at a Senate Budget Subcommittee hearing on cap-and-trade program revenue spending, Senator Bob Wieckowski raised concerns about ARB's decision to increase free allocation to industry above and beyond what is required under AB 398. CALmatters reports that the Senator indicated he would consider putting language in a cap-and-trade appropriations bill that would reverse ARB's directive in Resolution 17-21 to increase free allocation to industry in the market's third compliance period, 2018-2020. 

I want to take a step back and reflect on the arguments from industry, which broadly supports the direction the Air Resources Board gave its staff in Resolution 17-21. Julie Cart quotes Catherine Reheis-Boyd, president of the Western States Petroleum Association, who responded to Senator Wieckoswki's remarks by noting that “the way the allowances are distributed has no impact on how the cap-and-trade program ultimately reduces emissions.” In essence, Ms. Reheis-Boyd is saying we shouldn't worry about how much any individual group pays under the program; instead, we should focus on the bottom line for greenhouse gas emissions. Her argument is textbook environmental economics, and it's fair as far as that goes. But it's also an unsatisfying answer for two reasons. 

First, neither ARB nor WSPA has provided a justification for why WSPA members should receive extra benefits from the public. Free allocation is a transfer to special interest groups from the general public, which pays for the costs of climate policy. Sometimes these transfers are justified in order to protect against the risk of emissions leakage—the undesirable outcome in which some companies are put at a competitive disadvantage compared to their competitors who don't face a carbon price. Other times the transfers are pure politics that have nothing to do with leakage. In this case ARB directed staff to reverse course and increase, rather than reduce, free allocation. Some industries might very well need additional support, but the oil and gas industry gets the lion's share—72% of free allowances offered to industry in 2016, with refiners getting 50% of the total (see ARB data here). So as far as I'm concerned, there is plenty of room to ask how effectively a political solution that primarily benefits the oil and gas industry addresses leakage risks—especially given WSPA's central role in the AB 398 negotiations. 

Second, if we focus only on greenhouse gas emissions reductions and not whether consumers bear program costs that aren't shared equally with industries, Ms. Reheis-Boyd's argument is still incomplete. As she suggests, the level of free allocation doesn't affect the total number of allowances in circulation; in turn, the total number of allowances is what determines cumulative emissions under the cap-and-trade program. I agree that is important to avoid confusion over these two issues, but Ms. Reheis-Boyd's suggestion that cap-and-trade will deliver emission reductions whether or not any specific industry receives special treatment is unfounded. The program's effects on total emissions will depend on how ARB implements AB 398. 

Even if one is concerned only with the cap-and-trade program's effect on cumulative greenhouse gas emissions, industry's influence over the free allocation question is relevant because ARB will decide determine the program's stringency. Right now there are too many allowances in circulation, so the cap-and-trade program is weak and producing little if any effect on emissions. ARB adopted placeholder post-2020 regulations in Resolution 17-21 that do not address the market's oversupply problem. Unless ARB acts to tighten the program during the AB 398 implementation process, cap-and-trade will fall well short of California's climate goals.

From the beginning of the debate over ARB's Resolution 17-21, I have argued that it is fair to ask about the treatment special interests are receiving. Above and beyond those concerns, it is also extremely relevant to observe the fact of political side deals on free allocation that illustrate the influence powerful industries are exercising over the AB 398 implementation process. That influence is cause for concern with respect to the cap-and-trade program's ultimate stringency. ARB can and should do the right thing in addressing market oversupply, reforming its cap-and-trade regulations to deliver a stringent program that helps the state meet its 2030 climate target. But if ARB stays the course and listens to industry during this implementation process, the program outcome is likely to be weak and the state is likely to miss its 2030 climate target.

So whether one cares about the equity impacts of climate policy, the bottom line for emissions, or both, it's time to pay close attention to the Board's actions. 

Media: LA Times on carbon market auction

Chris Megerian quotes me in a Los Angeles Times story on the most recent carbon market auction results, which featured strong demand for allowances at modest prices. While the auction-clearing prices were the highest yet, they are a long way from where prices likely need to be to deliver on California's 2030 climate target:

However, some analysts are concerned about how the state will meet its ambitious goal for slashing emissions to 40% below 1990 levels by 2030.

Even though the cost of permits has been rising, it may not happen fast enough to provide an adequate financial incentive for companies to clean up their operations, said Danny Cullenward, a Stanford University energy economist.

State regulators should take steps to ensure they’re auctioning the right number of permits because making too many available could allow too many emissions, he said.

"Lurking in the background is a much more serious issue,” he said.

With AB 398's passage on 2/3 vote basis, there is no longer any legal uncertainty over the program's future, so it'e heartening to see auction demand pick up. Nevertheless, AB 398 delegates broad discretion to the market regulator, CARB, to determine the program's stringency going forward. CARB faces enormous pressure to keep the market in an oversupplied condition, but could do the right thing and clear up the oversupply problem with the legal authority the legislature provided. 

Without a clear signal that CARB is going to increase program stringency to deal with the oversupply problem, it's no wonder the new auction results were modest. Auction prices cleared a little more than a dollar over the minimum offer price, which is both the highest price in market history and far short of what California needs to get on track for its SB 32 target. This is exactly what one should expect given that AB 398, the market re-authorization bill, only requires CARB to analyze, not eliminate, oversupply. Absent a strong signal from CARB that it will do the right thing and design a program that binds, we should expect low market prices. 

So for all the good news, the auction results also indicate that those who want to see this program deliver on the state's ambitious goals have a lot more work to do.