Auction or Allocate


The great debate over emissions allowance distribution.

Fortnightly Magazine - February 2009

There may be no more sure way to increase interest in a cap-and-trade policy discussion than to broach the subject of distributing allowances. Frequently it’s one of the first issues stakeholders want to talk about, and one of the last resolved. The value of allowances makes distributing them one of the most challenging aspects of setting up a trading program, despite the fact that other program design features carry far more weight in determining the program’s ultimate success. In fact, the allowance distribution methodology has very little indeed to do with achieving the environmental and human health goals of cap-and-trade programs. However, given the level of importance understandably ascribed to allowance distribution, stakeholders should consider carefully the choices and the implications of alternatives, particularly in the context of a potential nationwide greenhouse gas (GHG) policy.

Cap-and-Trade Fundamentals

A cap-and-trade program first sets a cap, or maximum limit, on emissions that is below the business-as-usual level. Emission allowances, or authorizations to emit (usually 1 ton or 1 metric ton), are created by the government, with the total amount limited by the cap. Allowances function as the program currency, holding real value because of their scarcity and ensuring total emissions remain within the cap. Each regulated source can design its own compliance strategy to meet the overall reduction requirement, including sale or purchase of allowances, installation of pollution controls, fuel switching, or implementation of efficiency measures. Sources also completely and accurately measure and report all emissions. In order to comply, each emissions source must surrender allowances equal to its actual emissions to guarantee that the overall cap is achieved.1

There are two principal methods by which the allowances created under a cap-and-trade program can be distributed: They can be auctioned off, or they can be assigned free of charge to select entities (also known as “grandfathering” when given to those with compliance obligations). Of course, a hybrid of the two systems is another option. If allowances are auctioned, the government must decide how to design the auction and what to do with the proceeds; if allocated freely, the government must decide how to identify which sectors and entities should receive them, and how many each should receive.

Decisions relating to auction and allocation methodologies can be highly contentious, since they involve the distribution of valuable commodities and may affect some stakeholders differently than others. Governments often focus on the social cost of direct compliance, such as equipment installations and fuel switching. However, discussion of allowance-distribution methodologies also should consider the impacts related to compliance investments. Distribution methodologies that hinder the pass-through of allowance value (e.g., free allocation in cost-of-service states) or otherwise subsidize production (e.g., updating) may impact the program’s total cost by shifting the compliance response away from those able to achieve reductions at the lowest cost. Regardless, in cap-and-trade programs where initial reductions are far less than the remaining emissions (e.g., many climate proposals) the overall cost of actual mitigation measures initially is smaller than the face value of allowances.2

Importantly, how the allowances are given out doesn’t substantially impact a program’s results in human health and environmental protection; the emissions cap and banking provisions are the key program elements determining emissions levels over time.

Conventional Pollutants

The Environmental Protection Agency (EPA) has garnered a great deal of experience in allocating allowances for sulfur dioxide (SO2) and nitrogen oxide (NOx) emissions from the electric power sector, beginning with the acid rain program (ARP) and continuing through subsequent rulemakings on the NOx budget trading program (NBP) and the Clean Air Interstate Rule (CAIR).

Allowance distribution under the ARP was tackled by Congress in conjunction with the Clean Air Act Amendments (CAA) of 1990. These amendments created the ARP under a new CAA Title IV, producing a bold concept in reducing air emissions: the cap-and-trade program. Nearly 20 years later, cap and trade widely is recognized as an effective, efficient tool to reduce emissions significantly, but in 1990, it was a national experiment. The concept of emission allowances as a tradable commodity was new and the first exercise in allocation of these allowances was intense and political.

One important point to note at the outset is that allocation issues in Title IV were addressed after the cap was set. Further, the cap was held inviolable so distribution was a zero-sum game. That the allowances would be distributed at no cost to affected sources never was questioned seriously because it was vital to have the electric power industry on board with the new concept and desirable to achieve least-cost reductions. Also, due to the regulatory environment at the time of the legislation, the power sector faced a rate structure uniformly characterized as cost of service (unlike today’s environment, which is a hybrid of both restructured and traditionally regulated states) and savings were passed through to consumers. As a result, giving allowances away in 1990 meant smaller increases in energy bills, not creating potential windfall profits. The question facing Congress was more a matter of how to distribute the allowances most fairly.

After much debate, legislators agreed to allocate SO2 allowances on a permanent basis, primarily based on historic fossil-fuel utilization. Each source’s share of total allowances under the cap generally was determined by the product of an SO2 performance standard and average heat input. Ultimately, though, there were over 20 formulas used to account for a variety of different situations and political compromises, making it critical that specific allocations were spelled out in the legislation to provide much-needed certainty. Title IV also provided several set-asides of allowances under the cap to reward specific behavior. New sources did not receive any allocations in the statute, but annual government auctions of approximately three percent of the total allowance pool ensured at least one source of allowances outside the secondary market. That Congress made key decisions about allocations and other program details proved to be instrumental in getting the program off the ground without major delays.

Several years later, EPA had the opportunity to revisit the allocations process and examine alternatives under the NOx SIP call, a regulatory program in which states were required to revise their state air quality implementation plans (SIPs) to reduce summer time regional NOx emissions. To cost effectively achieve the emissions reductions required under the SIP call, EPA developed a model cap-and-trade program (NBP) to implement jointly with participating states in the eastern United States. States participating in the NBP were responsible for allocating allowances to their sources within certain parameters established by EPA, including the requirement that allowances be allocated at least three years in advance of their first-use date. EPA provided an optional allocation methodology that distributed allowances based on recent actual heat input data, and included a small new source set-aside. While each state adopted its own methodology, most opted for heat input-based allocations, while three states chose electricity output-based. Several states also included some small degree of auctioning.

EPA’s latest experience in allocating allowances came during the development of CAIR, which sought to further reduce SO2 and NOx. EPA continued Congress’s distribution approach for SO2 allowances under Title IV of 1990 amendments of the Clean Air Act. However, NOx allowance allocations were again the purview of the states within general parameters set forth in the regulation.

As different as these three experiences with allocations appear in some respects, the processes similarly were contentious with no clear right answer emerging for all of the stakeholders.

Distribution Schemes

Looking forward, while the shift in focus to a possible nationwide GHG-trading program substantially amplifies the total value of allowances, many aspects of the program remain similar for the electric power sector. EPA’s extensive experience with this sector offers insights that can inform the allowance distribution discussion. As in the past, various methods are available to distribute allowances: free allocations, auctions, or a combination of the two. Distributional issues remain a key consideration, particularly since GHG reductions will depend primarily on the generation mix, which varies substantially by company and geographic region (and likely will change over time).

It is important to keep in mind, particularly with the electric power sector, that the allowance market does not function in a vacuum. The status of price regulation in electricity markets plays a significant role in the treatment of allowances by firms. In traditionally price-regulated markets, history suggests that most generators have to pass on costs and savings from allowance transactions to their customers, so allowances that are freely given help to lower customer prices and don’t directly benefit the companies. However, generators in restructured markets bid their full marginal costs, including the opportunity costs of allowances, regardless of how they were obtained, into dispatch bids. The market price for all successful bidders is the marginal cost of producing the last kilowatt hour, including the opportunity cost of allowances. Thus, in each load segment, consumers in restructured states face the market clearing price that is higher with cost of allowances figured in and pay more for all of their generation even when some portion of it comes from sources not producing carbon (e.g., nuclear). Revenues obtained based on rates that include the opportunity cost of a free allowance have been characterized, in a similar vein to oil price run ups, as windfall profits.

Although we are focusing on the electric power sector, the basic options available for allowance distribution—allocations, auctions, and a hybrid approach—apply more broadly across the economy.

Allocating Allowances

When opting to allocate allowances, three major elements must be considered: allowance recipients, allocation basis, and how to address future changes. For simplicity, we assume fossil electric generators are responsible for compliance.

Determining the allowance recipient is the most elemental decision in developing an allocation scheme. Previous EPA programs allocated either to generators or to states (who reallocated to generators and at times, others). Another alternative is to distribute allowances to other entities not responsible for compliance. Similar to an auction, this third-party approach effectively requires emitting generators to purchase all allowances from a third-party, removing the distinction in allowance treatment between regulated and restructured markets.3 The end result is the full passage of allowance costs to consumers, unless the revenue collected by the third-party allocation recipients is used to offset consumer price increases (e.g., through rebates or energy efficiency improvements).4

The next major consideration is the metric upon which allocations are based. Commonly discussed metrics in the power sector include input (fuel use, fuel type), emissions, and output (kilowatt hours). EPA’s programs generally have adopted an input-based approach. The ARP allocates based on heat input, while the CAIR NOx program recommends allocations based on fuel-adjusted heat input. An input-based scheme essentially allocates based on emissions potential. Adjusting for fuel type directs more allowances towards generators with higher emissions—those needing to make the most reductions. Another approach is to allocate based on historical emissions. If generators receive the allocations, an emissions-based scheme would direct allowances towards those responsible for compliance. Alternatively, an output-based scheme considers only production and ignores emissions. While this approach rewards cleaner generation, a sector-wide application delivers allowances to sources that do not need them for compliance (e.g., nuclear, renewables). This outcome could be avoided by allocating to fossil fuel-fired sources only.

The third major consideration stems from the reality that input, output, and emissions behavior likely will change over time for all participants. Some will reduce emissions more than others, some will cease operations, and new generators will enter the market. These changes create a new landscape, where the objectives of the initial allocation scheme no longer may be met. One approach to address these changes is to update allocations at specified intervals (e.g., every four years). Updating provides a way to allocate to entities consistent with their current or more recent operations. Updating also provides an opportunity to allocate to new sources and discontinue allocations to retired sources. However, updating may influence the decisions of market participants by awarding allowances for, or subsidizing, the actions that the program is intended to reduce (e.g., emissions, fuel use, or production, depending on the allocation basis). The ensuing over-production may result in a lower electricity price and higher allowance price. Updating also can be resource-intensive, requiring substantial data collection and review by the government.

Allowance Auctions

Under an auction, anyone who meets a predetermined set of criteria can purchase an allowance. Auctions are attractive because they present a number of advantages, including their relative simplicity and transparency, the immediate setting of a price signal, the internalization of costs by emitting sources, and the provision of equal access for all market participants. Auctioning also removes any need to choose an allocation method, create special provisions for prior emissions reductions activities or other desirable behavior, or to account for the commissioning or retiring of regulated facilities. However, auctioning raises compliance costs for emitters, the same group that will bear a large share of the cost burden to meet the program cap. In addition, policymakers must decide how to use auction proceeds. Some economists argue that using these revenues to cut other distortionary taxes (e.g., income, investment) leads to the most efficient and least costly possible outcome from a cap-and-trade system (note this does not mean the cost of the program is reduced, but that the way costs are distributed can be used to reduce costs elsewhere in the economy). This notion that by raising revenue from emission allowances, the economy will benefit from a reduction in distortionary taxes and the environment will benefit from the introduction of a price signal that discourages pollution, commonly is referred to as the “double dividend.” On the other hand, various interests will be competing for a share of the revenues regardless of efficiency or ability to correct distortions. Revenue use can be progressive or regressive depending on how it is implemented.

Auction designs can vary widely, and selecting the auction format is one of the most critical decisions affecting the auction results. With the use of available technology and improved information systems, there is a great deal of flexibility in how auctions are conducted, and few geographical or physical limitations on participants. Prices can go up or down, bidding can occur all at once or over multiple rounds, bidders may or may not be able to see their competitors’ bids, and winning bidders all can end up paying the same price or the price at which they bid. There also can be a combination of grandfathered allowances and auctioned allowances and the mix can change over time.

There can be many different objectives favoring different auction designs, and priorities may include any of the following: simplicity and cost minimization, transparency, efficiency, price discovery, limitation of market volatility, adequate liquidity in the secondary market, and raising revenue. Appropriate design occasionally will require a tradeoff between competing goals: greater transparency, for example, can increase opportunities for collusion.5 In the United States, we have had less experience with auctioning allowances than with allocating them, although a great deal has been written on the subject, particularly in the context of a potential GHG program.6,7 Regardless of the auction format or percentage of total allowances, there will be a secondary market where allowances will be available for purchase or sale.

As previously mentioned, when allowances are sold at auction, the government captures the value of the allowances in the form of revenue raised. The government then faces the task of deciding how that revenue is to be used: Whether it will be returned to businesses or consumers or spent on government programs. The Congressional Budget Office looked at the effects of a 15 percent cut in CO2 emissions on various segments of the population under three different revenue recycling schemes.8 While the CBO analysis is informative, there are some limitations to the findings. Notably, the report is not specific to the power sector and does not address important differences in traditionally regulated and restructured states. Looking economy-wide, the study found that a lump-sum rebate given out directly to households was the most progressive policy in that it made lower-income families better off (because the size of the rebate was greater than the increased energy cost). However, this approach also had a high overall cost to the economy and less effect on incentives to invest in emissions reductions. Using revenues from auctions to cut payroll taxes was found to be regressive because it favored higher wage earners. Finally, if allowances were given out freely to regulated entities, it would result in the same cost to the economy as lump-sum rebates, but be the most regressive because the scheme largely benefits corporate shareholders while still increasing energy costs for consumers (with the exception of electricity costs inside cost-of-service areas). Other uses for revenue include funding R&D in new low-emitting energy technologies, providing aid to low income households for energy efficiency improvements, and reducing the federal deficit.

Allocation-Auction Hybrid

It also is possible to allocate freely a static or dynamic share of the allowances while auctioning the remainder. This hybrid approach permits policymakers to meet various objectives as they shift over time (e.g., near vs. long term). Most proposals that end up adopting this approach tend to allocate a decreasing share of allowances over time while auctioning an increasing share. The main advantage of this type of approach is the comparatively more measured impact on electricity consumers in cost-of-service states (which EPA believes is more than two-thirds of electricity customers) when a cap-and-trade program goes into effect. On the downside, there is no assurance of optimized compliance actions achieving reductions at the lowest cost.

In the short term, emitting generators may need to adjust their investment decisions to comply with new regulations. During this time period, free allocations to generators might assist with this transition. In the long term, the majority of the burden likely will shift to consumers who can be compensated through the use of auction revenue—suggesting the benefit of an increasing share of auctioned allowances over time. Alternatively, allocations or auction revenue could be directed toward a third party (e.g., federal agencies or states) to decide how to best use revenues from the sale of allowances.

Achieving Policy Goals

Both allocations and auctions can be used to achieve specific policy goals. Policymakers can direct assistance to regulated entities or others via an allocation scheme or use revenues from auctions in any number of ways. These two basic approaches to handling the allowances created by a cap-and-trade program can work alone or in combination.

When making decisions about how to distribute allowances in a cap-and-trade program, it is imperative to remember that the impacts of allowance allocation are distributional. Inevitably, there will be the perception of winners and losers in financial terms based upon which distribution scheme is implemented, but the overall environmental benefit generally is unaffected. In a cap-and-trade program, the emission cap and banking, not the distribution scheme, determines the amount of reductions that will be made. However, this doesn’t mean that the distribution method isn’t tremendously important or controversial. With tens of billions of dollars at stake, allowance distribution in any future nationwide GHG cap-and-trade program certainly will have significant and wide-ranging effects on different regions of the country and different stakeholders. As demonstrated in the past under Title IV and given the magnitude of value and the range of interests involved in any GHG scheme, specific direction from Congress can be invaluable for launching a new program. Then the most important work of reducing emissions can begin.



1. A cap-and-trade program conceivably could shift the point of regulation away from emitters (e.g., to fuel producers or distributors, who must surrender allowances equal to the emissions that would result from use of the fuel).

2. National Commission on Energy Policy. Allocating Emissions in a Greenhouse Gas Trading System, March, 2007. The figure on page 6 of the report illustrates why there is likely to be a large discrepancy between CO2 allowance value and mitigation costs at the program’s beginning when reductions are relatively small. Note that the shape of the marginal cost curve dictates the degree to which emissions reductions must exceed remaining emissions before the total mitigation cost surpasses allowance value. See report for details.

3. One notable distinction is that auction revenue goes to the government, whereas proceeds from sales of third-party allocation go to allocation recipients.

4. This concept extends beyond the electric power sector only to the extent that producers are able to pass costs to consumers, which may not be possible in some sectors.

5. Burtraw and Palmer. RGGI Workshop Summary.

6 . The U.S. government does have considerable experience using complex auction procedures for other products it regulates such as electromagnetic spectrum portions, treasury bills, and access to the electricity grid.

7. The first greenhouse gas allowance auction in the United States was completed by the Regional Greenhouse Gas Initiative (RGGI) in September 2008. More information on the auction proceedings and results is available through RGGI.

8. Congressional Budget Office. Issues in Designing a Cap-and-Trade Program for Carbon Dioxide Emissions, Sept. 18, 2008.