The California Air Resources Board is finalizing its Scoping Plan for implementation of the state’s global warming law, AB 32, which could establish a precedent for federal legislation by the 111th Congress. Barbara Boxer recently announced plans for a cap-and-trade initiative to be introduced in January, and she earlier indicated that the next go-round on federal legislation could cede much of its design to California’s AB 32 legislation. As Congressional leaders look to California for guidance in fashioning federal policy, they should be cognizant of what AB 32 actually requires and how those requirements are — or are not — being translated into regulatory action.

The Scoping Plan relies on cap-and-trade in the tradition of the U.S. acid rain program, in combination with existing programs and traditional command-and-control mechanisms, to achieve most of the requisite emission reductions. As noted in the plan documents (Appendix C [PDF]), “The acid rain program has successfully achieved the environmental goal of the cap at a cost of several billion dollars less than originally expected.” CARB’s cost projections for the Scoping Plan are similarly biased toward cost conservatism (Appendix G [PDF], Sect. 1.2), so CO2 trading prices in 2020 could be well below the projected level of $10/MT (i.e., $10 per CO2-equivalent metric ton).

Extrapolating from prior experience with the acid rain program, the E.U. ETS, and RGGI, there is a distinct possibility that under the proposed plan emission allowances could be trading at something like $3/MT in 2020 — even after the last remnants of the arctic ice cap have vanished into an empty sea. Statewide emissions will have returned to their 1990 level, and cap-and-trade will have succeeded at achieving the “environmental certainty” of the cap at minimum cost; yet the impending threat of catastrophic global climate change may nevertheless have become an inevitable certainty.

The specter of CO2 allowances selling for $3/MT in 2020 does not comport with the legislative policy of AB 32. The statute does not identify the state’s 1990 emission level as a “target” to be hit in 2020. It specifies a “maximum allowable level of statewide greenhouse gas emissions in 2020,” which is equivalent to the 1990 level. This is an upper limit (not a target), and the statute further directs CARB to maximize emission reductions within certain stated limitations:

38560. The state board shall adopt rules and regulations in an open public process to achieve the maximum technologically feasible and cost-effective greenhouse gas emission reductions from sources or categories of sources, subject to the criteria and schedules set forth in this part.

The requirement for maximum emission reductions makes it clear that the “target” is the limit of feasibility and cost-effectiveness; it is not just the 1990 level.

In essence, the statute requires a reasonable best effort to reduce emissions. By contrast, cap-and-trade, as conventionally practiced, is fundamentally a “least-effort” policy. Emissions are capped at a predetermined, unsustainable level (perhaps declining according to a predetermined schedule, but without basis in climate science), and the system provides no long-term incentive for further emission reductions even if allowances are trading at just $3/MT.

So how is CARB able to reconcile AB 32 with the possible prospect of a $3/MT emission price in 2020?


(Warning to the reader: This discussion is about to descend into a labyrinth of linguistic and legalistic abstraction, but it actually has significant and practical implications for what California and other states and governments are going to do about climate change.)

The answer to the above question can be found in the way that CARB interprets the concept of “cost effectiveness.” The statute provides the following formal definition:

38505(d). “Cost-effective” or “cost-effectiveness” means the cost per unit of reduced emissions of greenhouse gases adjusted for its global warming potential.

This appears reasonably concise, except that the term “cost-effective” is used in the statute as an adjective, not as a noun. Taking the above definition literally, the statute would read something like “… achieve the maximum technologically feasible and ‘dollars-per-ton’ greenhouse gas emission reductions …” (??) The legislative intent is clear: “Cost-effective” (in the adjective sense) is clearly a threshold condition subject to which “greenhouse gas emission reductions” (not cost reductions) are to be maximized; and the above-quoted definition suggests that the threshold is intended to be a dollar-per-ton limit. However, the statute does not specify the threshold or provide a criterion for assessing whether a regulation is or is not cost-effective. Apparently, CARB is free to establish any such criterion that is consistent with the legislative intent (and with the 2020 limit).

The Proposed Scoping Plan [PDF] outlines an “approach” for evaluating cost effectiveness (Sect. III.C), although the description obfuscates CARB’s understanding of what “cost-effective” actually means. Following is an excerpt:

… In the meantime, the current estimates [of costs for CARB’s proposed regulatory measures] provide a range illustrating the cost per ton of the mix of measures that collectively meet the 2020 target. This range will assist the Board in evaluating the cost-effectiveness of individual measures when considering adoption of regulations. The range of acceptable cost-effectiveness may change if effective lower-cost measures and options are identified … the approach must provide flexibility to pursue measures that simultaneously achieve policy objectives other than greenhouse gas emissions reduction (such as energy diversity).

The criteria for judging cost-effectiveness will be updated as additional technological data and strategies become available …

This “approach” raises more questions than it answers: It asserts that CARB’s proposed regulatory measures can provide a reference basis for evaluating cost-effectiveness; but by what standard (if any) is the cost effectiveness of those measures established? Regarding changes in the “range of acceptable cost-effectiveness,” why would the cost effectiveness of a particular measure be affected by unanticipated cost reductions in an unrelated industry? How and why would the criteria for judging cost-effectiveness be updated in response to “additional technological data and strategies”?

The public documents underlying the Scoping Plan provide insight into these questions. On June 3, 2008 CARB held an Economic Analysis Technical Stakeholder Work Group Meeting on the subject of cost effectiveness, in which Professor James Sweeney of Stanford University, at CARB’s invitation, gave a presentation [PDF] outlining the economic foundations of “cost effectiveness.” The clarity and mathematical precision of his proposed definition stands in contrast to the Scoping Plan’s opacity:


A set of greenhouse gas mitigation measures is cost-effective under a given target emission reduction if and only if the set of measures together imposes the minimum cost to society (among all feasible measures) of meeting the target emission reduction.

  • This is defined in relation to a particular target.
  • We are referring to the social cost here.
  • Cost must include ancillary costs/benefits, e.g. non greenhouse gas environmental impacts.

In a nutshell, “cost-effective” translates to “least-cost” under this definition.

Its conciseness notwithstanding, the above-defined criterion is highly discretionary because consideration of “ancillary costs/benefits” — which are not necessarily quantifiable — can override the cost minimization objective. In particular, the Scoping Plan includes adoption of a 33 percent Renewable Portfolio Standard by 2020, which would impose projected incremental net costs of $133/MT (relative to the established 20 percent RPS) even while alternative, GHG-equivalent compliance measures would be available through the trading system at a marginal cost of only $10/MT. (Moreover, incremental reductions in passenger vehicle emissions would yield projected net savings of $262/MT.) But the $133/MT RPS cost is justified by the ancillary benefit of “energy diversity.” (The governor just recently ordered implementation of the 33 percent RPS.)

Considering CARB’s discretionary latitude in accommodating ancillary costs and benefits, the only really firm requirement in Sweeney’s definition is that of meeting the target. This can explain how the cost effectiveness of CARB’s proposed regulatory measures is established.

The above-noted RPS cost projection has significant implications for what could be achieved with a best-effort regulatory strategy. If the benefits of the 33 percent RPS (including energy diversity) would justify the $133/MT incremental cost, then those benefits could justify an even greater expansion of renewable power if costs turn out to be much lower than expected, as will likely be the case. (Solar power is expected to reach “grid parity” in about eight years, and a stable price incentive on the order of $133/MT might accelerate commercialization of solar well beyond the 33 percent target.) Furthermore, energy diversity considerations would apply equally well to petroleum dependence in the transportation sector, and could justify commensurate incentives for vehicle efficiency technologies such as lightweighting and hybridization. (Actually, considering that one ton of vehicle CO2 emissions equates to about 100 gallons of gasoline, an incentive consistent with zero net cost might suffice to induce commercialization of even the most advanced vehicle technologies.)

Sweeney’s definition of cost effectiveness is inconsistent with a best-effort regulatory strategy because it essentially requires least-effort (i.e., lowest-cost) emission reductions. It does not recognize any policy rationale for exceeding the reduction target, no matter how low the cost. This explains why, under CARB’s approach, the “cost effectiveness” of a particular measure could be affected by unanticipated cost reductions in an unrelated industry: If a new, low-cost technology or strategy is adopted, then some other measure would have to be discontinued to avoid significantly overshooting the 2020 target and incurring the added cost of “overcompliance.” CARB’s “criteria for judging cost-effectiveness” are that the 2020 target is achieved at minimum cost (or subject to other policy objectives imposed by CARB).

The Sweeney definition provides insights into CARB’s interpretation of cost effectiveness, but would create legal complications if it were adopted literally in the Scoping Plan. The predication of the definition on “a particular target” conflicts with the statutory language requiring “the maximum technologically feasible and cost-effective greenhouse gas emission reductions” because emission reductions cannot be maximized if the emission target is predetermined. Furthermore, the Sweeney interpretation of “cost-effective” as meaning “least-cost” is problematic because CARB’s adopted measures do not appear, in any objective or quantitative sense, to actually minimize costs.

The concept of “cost effectiveness” portrayed in the Scoping Plan is nebulous and ill-defined, perhaps purposefully so, because the ambiguity circumvents legal complications while allowing CARB to apply a regulatory approach that qualifies adopted measures as “cost effective” if they simply achieve the 2020 target along with any other discretionary criteria that CARB might impose (e.g. accommodation of the governor’s directives or other political influences).

A $3/MT trading price would be compatible with this approach, but how does CARB reconcile its concept of cost effectiveness with the clearly-stated legislative policy of maximizing emission reductions? CARB’s thinking is clearly revealed by its response to the following question (which I posed at the November 21, 2008 Board meeting [PDF]):

… Suppose that the statute had been written with the word “maximum” omitted from that language [“maximum technologically feasible and cost effective emission reductions”] … Is there anything in what you’ve proposed that is significantly different from what it would be if the word “maximum” were omitted …?

CARB staff (Chief Counsel Robert Jenne) responded as follows:

Regarding Mr. Johnson’s question, which was really about statutory interpretation, we designed the plan to achieve the 2020 target by using the most cost-effective measures we could identify. If we are able, as we learn more, to identify additional measures that are within the same range of cost effectiveness, and they’re technologically feasible, the Board would be required to adopt such measures to fulfill the statutory mandate in AB 32 to adopt the maximum technologically feasible and cost-effective measures.

This response paraphrases the statute, but with “measures” substituted for “greenhouse gas emission reductions.” It is not clear exactly what quantity is being maximized in this context. “Maximum” appears to qualify “measures,” but that doesn’t make sense semantically. (What are “maximum measures”?) The reference to “most cost-effective measures” suggests that CARB reads the word “maximum” as applying to “feasible and cost-effective” in the above context and in the statute (as in “maximally feasible and cost-effective”). But that does not make grammatical sense because “maximum” is not an adverb. In any case, there appears to be no recognition that “maximum” qualifies the noun “reductions,” which does not appear in the above statement.

The staff response continues:

We would also note that the plan already provides for that in some degree. We have a margin of safety in the plan to achieve additional reductions beyond the 2020 target to make sure we actually get there.

This paragraph does mention “reductions” but omits the qualifier “maximum.” The “additional reductions” amount to about one percent of the 2020 emission limit, and this small overshoot is only intended to accommodate regulatory uncertainty — not to maximize reductions.

CARB’s Chief Counsel does not appear to recognize any statutory requirement or authorization for emission reductions significantly beyond the 2020 limit, even at $3/MT. The statutory language relating to maximum reductions has apparently not affected reduction levels under the Proposed Scoping Plan, and the Plan would not preclude a $3/MT emission price in 2020.

CARB’s statutory interpretation has important implications for post-2020 regulations, which will need to achieve much greater reductions in statewide emissions (e.g. at a 4 percent annual reduction rate between 2020 and 2030, according to the Proposed Scoping Plan, Sect. V.C). Inadequate incentivization of maximum emission reductions prior to 2020 could result in exceedingly costly and economically disruptive regulations after 2020 — particularly if manifestations of climate change such as disintegration of the arctic ice cap are approaching crisis proportions. Thus, from the perspective of the statute’s long-term legislative policy objectives the regulations will not have complied with the legislation’s stated requirements for early action and cost minimization:

38562(b) … the state board shall … (1) Design the regulations … in a manner that is equitable, seeks to minimize costs and maximize the total benefits to California, and encourages early action to reduce greenhouse gas emissions.

(Regarding cost minimization, Sweeney’s definition of cost-effectiveness as “least-cost” would be compatible with the statute if the “target” were defined as climate stabilization.)

Most of the institutional stakeholder and advisory groups involved with the AB 32 process have failed to recognize and understand the significance of the qualifier “maximum” in the statute, as it applies to emission reductions. (Two exceptions: Earthjustice and the Western Environmental Law Center have submitted legal briefs to CARB challenging the legality of its plan.)

The maximum reduction requirement would not necessarily require more stringent caps or standards. A clear lesson of the U.S. acid rain program and other regulatory systems is that emission caps and standards cannot reliably achieve “maximum technologically feasible and cost-effective” emission reductions because it is not possible to predict the limitations of feasibility and cost-effectiveness years or decades into the future. However, such measures can be complemented with incentive-type regulations, such as a price floor in the context of cap-and-trade, or vehicle feebates, to better capture the full potential of cost-effective reductions.

Regarding federal legislation, Congressional leaders must decide whether to adopt a “best-effort” or a “least-effort” approach to greenhouse gas regulation. If a best-effort approach is taken, then Congress can follow the precedent set by the AB 32 mandate requiring “the maximum technologically feasible and cost-effective greenhouse gas emission reductions.” But the legislation should give EPA clear guidance on the meaning of “cost-effective”; simply defining it as “dollars per ton” would leave too much to regulators’ imagination (or lack thereof).