Addressing Leakage In Regional Regimes

Addressing leakage is important not only to protect the economic competitiveness of industries located within the regulated regime, but also to protect states outside the regulated region and the future trajectory of national and global GHG emissions reduction regimes. Regional GHG emission reduction regimes that promote leakage reduce their internally generated emissions but increase the emissions of neighboring, unregulated states. Increasing the emissions of neighboring states may make it difficult for those states to join a

regional regime later, since they increased their electricity production to increase exports to the neighboring regulated region. Entering the regional regime may put the late entrants at a disadvantage if their emissions reduction targets do not account for their increased production and export of electricity to serve the regulated region.

The failure to address leakage also makes it unlikely that non-member states will join the regional regime for other reasons. Non-member states that decide to join a leakage-prone regional regime will lose their ability to export electricity into the regional regime at a price advantage and will face the problem of cheaper electricity imports competing with domestic electricity generators. Finally, a leakage-prone regional regime, if designed as a model for a broader, coordinated emission reduction regime, can hinder the development of such a program by infusing doubt about the effectiveness of such a regime into the minds of legislators. Given the high stakes presented by emissions leakage, the RGGI states and the Western states, including California, have attempted a number of different approaches to minimize or eliminate leakage.

4.1 The Eastern Approach: Cost-Containment

Recognizing that leakage is a product of compliance cost differentials between entities within the RGGI region and surrounding jurisdictions, the RGGI states seek to address leakage by reducing the RGGI cost adder. RGGI's modeling analysis concluded that given the unique nature of CO2 pollution in the region and the various incentives that promote or mitigate leakage, RGGI will increase the cost of compliance by less than 5% of the average wholesale electricity price.58 Given the minimal existing price differences between RGGI-generated electricity and non-RGGI-generated electricity, this cost adder is not expected to promote significant leakage. Based on these data, the RGGI states concluded that they need not eliminate leakage completely, so long as they can keep it within acceptable levels. Their primary strategy to mitigate leakage is to contain the additional compliance cost faced by regulated entities.

4.1.1 Safety valve

The primary cost containment measure employed by the RGGI states is termed a 'safety valve.' A safety valve is a relief mechanism designed to minimize some of the cost implications of a regulation if the compliance costs become unacceptable or exceed a pre-determined threshold. RGGI builds on the Acid Rain Program's idea of a safety valve, which releases reserve

allowances into the market if the price of allowances exceeds a particular price threshold, with an offsets safety valve.59

A safety valve, while effective at keeping the cost of compliance within modeled expectations, may have unintended negative consequences. A safety valve, if not properly designed, can establish a 'price cap' on allowances. Indeed, the more effective the safety valve, the more likely it is to establish such a price cap and ensure that costs are contained below the trigger price level or cost threshold. While effective at containing costs, a price cap can keep the cost of compliance sufficiently low to discourage investment in technologies that might achieve sustained emissions reductions. Ultimately, RGGI rejected applying a safety valve in a way that would create a robust price cap.

The RGGI states took a slightly different approach than the Acid Rain Program when establishing a safety valve for the regime. Rather than increasing the number of available allowances when the price of allowances reaches a particular trigger level, the RGGI states created a safety valve that incrementally increases the availability of offsets if the price of allowances exceeds various trigger levels. Offset credits and allowance tracking are managed by a non-profit regional organization (RGGI, Inc.), which has no regulatory or enforcement authority, established by the RGGI states in 2007. This difference in design has substantial implications for the incentive structure of the regime.

The Acid Rain Program of the Clean Air Act essentially creates a price cap by increasing the supply of sulfur dioxide allowances available to regulated entities - and thereby reducing the cost of those allowances - if the price of allowances exceeds a particular threshold.60 This provides great clarity for regulated entities to determine whether the cost of investing in new emission-reducing technologies is economically advantageous, but it is not necessarily technology-forcing. It also does not ensure overall emissions reductions by a particular regulated entity, since it promotes a game of chicken, whereby all facilities seek to exceed their allowances, thereby driving up the allowance price, in the hope that the regime will release additional allowances. The Acid Rain Program's safety valve, therefore, undermines the Program's goal to ensure sulfur dioxide emission reductions.

RGGI, on the other hand, does not have reserve allowances that are released into the market if allowances prices (e.g., the compliance cost) are deemed too high. Instead, if meeting the RGGI emission reduction targets becomes too expensive for regulated entities (i.e., exceeds the pre-determined allowance price triggers known as the 'Offsets Trigger Event' and the 'Safety Valve Trigger Event'), regulated entities can decide either to purchase

allowances on the market or offset their emissions in areas where cheaper emissions reductions are possible. This increased ability to use offset emissions helps keep the demand for, and therefore price of, allowances down.

Because RGGI limits the use of offsets to five or ten percent of a source's total allowances depending on the price level trigger, RGGI does not set an allowance price cap in the same way that the Acid Rain Program does. The allowance price under RGGI is variable and could increase fairly substantially depending on a variety of uncontrollable circumstances. Given this situation, RGGI encourages regulated entities to develop and install new technologies that will achieve lasting emissions reductions, which may result in total regional emissions below the overall target.

Given the potential volatility in the allowance market, RGGI's safety valve provision is only triggered when the price level exceeds the trigger level. When the allowance price exceeds the highest price triggering level ($10 (2005$)/ton CO2, as adjusted for inflation) for twelve consecutive months, then the compliance period may be extended for a year, for a maximum compliance period of four years.61 After a trigger of the various safety valve provisions, RGGI resets the compliance requirements and offset limitations at the beginning of the next three-year compliance period.62 Overall, RGGI's offsets safety valve serves to minimize potential leakage because it sets a rather low trigger price, though leakage can still be fairly significant if allowance price volatility is not controlled.

While the safety valve provisions related to the use of offsets are designed primarily to reduce the compliance cost to sources, RGGI also seeks to lessen the potential economic burden of the program suffered by consumers. RGGI requires that at least 25% of each state's CO2 allowances be awarded for a 'consumer benefit or strategic energy purpose.'63 These allowances can be used as a means of 'fostering renewable energy, offering consumer rebates, stimulating innovative carbon-reduction technologies, and funding the administration of the program.'64 The RGGI states have determined that the easiest way to allocate the allowances for consumer benefit or strategic purposes is through an auction. While the RGGI states have discretion how they will allocate their remaining allowances, most of the states (New York, Massachusetts, Vermont, Rhode Island, Connecticut, and Maine) have declared their inten

61 RGGI, MOU Amendment, n. 10, art. 1(a). This is similar to the safety valve provision in California's AB32, which allows the Governor to extend compliance deadlines up to one year in 'extraordinary circumstances.' California Health & Safety Code § 38599.

tions to auction 100% of their allowances, and use the proceeds of the auctions in programs designed to reduce consumer costs, such as demand-reduction strategies.

4.1.2 Early reduction credits

RGGI also seeks to reduce the compliance cost imposed on regulated entities by giving them an opportunity to earn additional allowances before the compliance period begins. Since RGGI uses 2009 as the baseline year against which future reductions are compared, the RGGI states recognized that the program may create perverse incentives for a state to over-pollute in 2009 to inflate its emissions baseline artificially. Inflating the baseline in this way would prevent the RGGI market from being 'short' on allowances - essentially creating a surplus of cheap allowances post-2009 - that would enable regulated entities to meet their GHG emission reduction targets cheaply.

To lessen this problem, RGGI enables sources to earn 'early reduction credits' for actions taken prior to 2009, but after the state hosting the sources entered into the RGGI Memorandum of Understanding, to reduce GHG emissions.65 By creating additional allowances, rather than carving them out of the existing budget, RGGI effectively increased the supply of allowances and helped to alleviate speculation during the 'market settling period' and start-up phases of the program.66 This approach will help to reduce price volatility and lower the allowance price, thereby reducing potential leakage.

4.1.3 Other cost containment measures

The RGGI states also employ a variety other mechanisms to help contain compliance costs and increase smooth price discovery of allowance prices during the 'market settling period.'67 First, RGGI uses a three-year compliance period unless a safety valve triggering event occurs to lengthen the

65 RGGI, MOU, n. 10, art. 2(H). Similarly, California's AB32 provides an opportunity for regulated entities to earn early action credits. Ca. Health & Safety Code § 38562(b)(3).

67 Some claim that RGGI uses a circuit-breaker to help contain costs. This is technically inaccurate. A circuit-breaker cost containment provision operates in a manner similar to a safety valve provision in that if an allowance price trigger is exceeded, the schedule providing for a declining overall emissions cap is frozen until the spike in the allowance price abates. RGGI has no such direct cost containment provision, despite the request of some commenters for such a provision. Nevertheless, one might view the possible extension of the compliance period for a year on the occurrence of a Safety Valve Trigger Event as a type of circuit-breaker cost containment provision, since the compliance deadline extension may prevent a lower cap from becoming effective during that year.

compliance period. This multi-year compliance period smooths out CO2 emissions spikes relating to unforeseeable events, such as heat waves that result in more air conditioning use than expected. The multi-year compliance period, then, reduces the cost of compliance for regulated entities. The first three-year compliance period begins on 1 January 2009.

Second, RGGI enables participating states and regulated entities to 'bank' their surplus credits into future years of the program, thereby using surplus credits from one year to offset increased emissions in subsequent years. This temporal flexibility appears scientifically permissible in a GHG emission reduction regime given the long-time horizon of GHG pollution and its expected impacts. Banking credits helps smooth out weather-related compliance crunches, thereby reducing the cost of compliance for regulated entities.

Finally, most of the RGGI states plan to auction all of their allowances. Auctioning allowances can help ensure smooth discovery of allowance prices, and thereby minimize price volatility. Reducing volatility of the market price for allowances can help reduce costs to regulated entities, and thereby reduce a potential cause of emissions leakage.

4.2 The Western Approach: Load-based Emissions Caps Plus

In the western United States, there is greater diversity of primary energy fuels and environmental controls among states, which results in states having highly divergent per kilowatt hour GHG emissions. California, relying heavily on natural gas, generally emits low levels of GHGs per kilowatt hour - at a relatively high price - while its neighboring states either have an abundant supply of cheap hydroelectric power (Pacific Northwest) or rely on cheaper, but GHG-emitting, coal-fired power plants. This diversity results in a substantial price difference between electricity generated in California and its surrounding states - a price differential that spans from a minimum of over three cents per kilowatt hour to almost eight cents per kilowatt hour (nearly two-thirds of California's retail electricity price of 12.820/kWh in 2006).

Given the large electricity price differentials between California and the surrounding WCI and non-WCI states identified in Table 9.2, and California's need to import electricity to satisfy its insatiable energy demands, emissions leakage is a significant concern for California. Accordingly, California has taken a variety of steps to try to minimize and eliminate emissions leakage.

4.2.1 Load-based emissions caps

Beginning in 2005, California began discussing use of a load-based emission cap to control emission leakage. California's Public Utility Commission ultimately determined that it prefers to use a load-based emission cap to implement the cap-and-trade regime authorized by AB32. This load-based emission

Table 9.2 Western Interconnection Generation and Sales in 200668













Fuel Source)



(% of

(1000 tons)




New Jersey







Arizona (Coal)













New Mexico



































Idaho (Hydro)













Nevada (Gas)













cap is the first of its kind and is designed to capture the GHG emissions from electricity imports in California's regulatory scheme.

California's load-based emission cap imposes GHG emissions caps on load-serving entities, which are private companies that sell electricity to end users after purchasing that electricity from sources. This load-based regime targets GHG emissions related to delivery of electricity into California, in contrast to a generator-based regime, which, if adopted, would target the GHG emissions resulting from the production of electricity within California.

By imposing emissions caps on LSEs, California's load-based regime creates an incentive for LSEs to purchase electricity provided to California from low GHG-emitting generators. LSEs are not bound by geography in their provision or purchase of electricity, so a load-based cap provides emissions

68 United States Energy Information Administration, n. 7, pp. 261-2, tbl. A1 (Selected Electric Industry Summary Statistics by State, 2006).

caps for all LSEs seeking to deliver electricity to the California grid, irrespective of whether they are located in California or purchasing electricity generated within California.69 This load-based cap regulates the GHG emissions of almost all electricity consumed within California.70 While California is more prone to problems with leakage than RGGI and other states, most of the other members of the WCI have declared their intention to establish a similar load-based cap, which is specifically sanctioned by WCI.

A load-based emission cap has a number of advantages over a source- or generator-based emission cap. First, a load-based regime is most effective at reducing or eliminating leakage.71 By focusing on consumer demand, it incorporates industrial process-related emissions as well as generation-related emissions. As with a source-based emission cap, however, it does not address leakage that results from the export of energy-intensive industries (and resultant consumption of electricity) outside the regulated region as a result of increasing electricity costs. Second, by focusing on consumer demand, rather than generator supply, a load-based cap inherently values demand reduction strategies and efficiency improvements.

A load-based emission cap is also preferable to a source-based cap because it places the decision-making power in the hands of LSEs, which have the ability to prioritize low GHG-emitting technologies. A load-based cap creates incentives for LSEs to invest in cleaner technologies and creates a price signal that low-emitting generation is a valuable commodity. This price signal means that the design of a load-based emission cap is simpler than a source-based cap when it comes to incorporating renewable energy strategies, since renewables have inherent value in a load-based system. In a source-based system, on the other hand, creating value and incentives for production of renewable energy requires treating different categories of sources differently or creating a second layer of regulation, in terms of emission portfolio standards.

Finally, a load-based cap ensures greater flexibility than a source-based cap. Many generators are coal-fired power plants that cannot suddenly shift to producing renewable electricity without substantial economic loss.

69 California Public Utilities Commission (08/02/2008), Proposed Decision on Rulemaking 06-04-009, Interim Opinion on Greenhouse Gas Regulatory Strategies (adopting the first-seller rule as recommended by the Market Advisory Committee, with a slight modification). For a discussion of the first-seller rule, see generally Market Advisory Committee (2007).

70 California Public Utilities Commission (16/02/2006), Decision 06-02-032 in Regulation 04-04-003, p. 17 ('LSEs would be subject to an emissions cap for all resources procured to serve their load, no matter what the source, including imports.').

71 See Bird et al. (2007, p. 39); RGGI, Initial Leakage Report, n. 10, pp. ES-11, 41; Climate Action Team (March 2006, p. 69), Report to Governor Schwarzenegger and the Legislature.

Accordingly, a generator-based emission cap, while effective at reducing per kilowatt hour GHG emissions for each type of generating source, does not create incentives to shift electricity generation to different types of generating sources with lower GHG emissions. A load-based cap does. Under a load-based cap, LSEs can meet their load-based allocation by shifting their portfolio to reduce their purchase of high GHG-emitting electricity in favor of low-GHG electricity. The flexibility of a load-based emission cap makes it less likely to increase costs to end-use consumers as significantly as a generator-based cap might.

While a load-based emission cap is desirable for a variety of reasons, it also suffers some potential design problems. A load-based cap may suffer from a problem known as 'contract shuffling.'72 LSEs purchase electricity from different types of electricity generators in a number of states to provide electricity to a variety of states. Given this dynamic, LSEs can decide to allocate all the low GHG-emitting electricity in its generator portfolio to California and the WCI states and allocate all the highest GHG-emitting electricity to the other states served by the Western Interconnection electricity grid that are not within the load-based regional regime. This contract shuffling may mean that while the electricity consumed in California meets California's GHG targets, no actual emissions reductions were achieved by the regulation.73

Governor Schwarzenegger's Climate Action Team concluded that contract shuffling is likely to be a one-time problem that is unlikely to persist given California's declining emission cap.74 It concluded that 'once all the existing low-emitting units are spoken for, additional emission reduction would need to be achieved through other means.'75 This conclusion has merit since the states with excess electricity capacity derived from low GHG-emitting sources - Oregon and Washington - have determined that they will join WCI and implement a load-based cap. Accordingly, there is unlikely to be substantial contract shuffling of low GHG-emitting electricity out of those states to California, since each of the states joining a regional load-based regime will need low GHG-emitting electricity to satisfy their emission reduction targets. This means that only the excess low GHG-electricity capacity will likely be sold to California. That excess capacity would only cover about half of

72 Market Advisory Committee (2007, p. 44; RGGI, Initial Leakage Report, n. 10, p. 41); Cap and Trade Subgroup (2006, pp. 23-4).

73 Similarly, a load-based emission cap may create an incentive for high GHG-emitting generators within the regulated region to sell their electricity outside the regulated region, resulting in 'reverse leakage.' Cowart (2004). This problem is of little concern in California, given other environmental protections and performance standards that discourage the development of high GHG-emitting sources.

75 Ibid.

California's import needs. Low-GHG emitting sources in other states simply do not exist in sufficient quantity to enable significant contract shuffling in a declining cap scenario.76

Nevertheless, given the potential problem of contract shuffling, a load-based cap must be able to identify the GHG emissions associated with imported power, which may be difficult given existing GHG monitoring and reporting (despite advances made through The Climate Registry). Some have suggested that one way to avoid contract shuffling would be to assign an average GHG emission rate to imported power.77 This ostensibly would avoid the contract shuffling incentives, but could present problems under the United States Constitution by treating imported power differently from domestic power.

Irrespective of whether such an approach is ultimately adopted, addressing the problem of contract shuffling in a load-based cap-and-trade scheme presents a number of implementation difficulties.78 One such difficulty arises because a load-based cap requires an emissions tracking system. A tracking system is problematic for long-term electricity supply contracts that do not specify the generating source, as well as spot market purchases, which generally do not identify the generating source. In 2007, the Oregon legislature proposed House Bill 3545 to address this problem by specifying a 'residual emission rate' to unspecified power purchases based on an average emission rate of all generation within the power pool not accounted for in existing contracts, but the bill has not yet gone to vote. The California Energy Commission is considering a similar proposal, but California addressed this concern another way: by prohibiting LSEs from entering long-term power purchase agreements unless the contracting generator meets a GHG emission performance standard.79

While solving one problem, such proposals create a complex incentive structure that may require additional study prior to adoption. On the one hand, a 'residual emission rate' could inadvertently create an incentive for dirty

76 For a discussion of contract shuffling in this regard, see Davis (2005, p. 14).

78 For a discussion of some of these difficulties, see RGGI, Initial Leakage Report, n. 10, p. ES-11.

79 California Public Utilities Code § 8341. For more information about emission performance standards, see Davis (2005, pp. 8-11). It is well-recognized that efficiency improvements can reduce leakage by reducing the cost of compliance to generators and LSEs. E.g., Prindle (2006). Accordingly, California employs both emission performance standards and 'decoupling,' which rewards generators for selling less electricity than expected. Bryk & Snyder (2007, p. 99). The effectiveness of providing 'decoupling' rewards to high GHG-emitting generators within a load-based emission cap is yet to be determined, but appears to risk rewarding dirty generators.

generators to provide their electricity through the unspecified spot market or long-term power purchase agreements to receive a more favorable GHG emission rate.80 This unintended consequence, however, may be offset by a similar incentive for low GHG-emitting generators to avoid the spot market and unspecified generator contracts so they might receive the full value of their low rate of GHG emissions. Since LSEs are price takers in the spot market, a 'residual emission rate' may ultimately provide a clear decision rule for LSEs deciding whether to purchase electricity on the spot market, which may discourage carbon-intensive spot market purchases, but more information is needed about the incentives created by a 'residual emission rate' to understand its full implications.

Another concern with a load-based emission cap is that LSEs may currently lack sufficient information to make informed decisions about the GHG emissions from their existing contracts. As a result, it may be difficult for LSEs to discover the price of allowances on the market, which could result in significant allowance price volatility in the early years of operation. To overcome this problem, some have suggested that allowances be sub-distributed to generators based on LSE emission allocations, since it is presumed that generators will have greater information about their GHG emissions than LSEs. This solution is unsatisfactory, however, because this would enable LSEs to earn windfall profits from 'flipping' their allowances to generators, and it would undermine a core benefit of the load-based cap: providing flexibility in meeting the cap through demand-reduction strategies and portfolio adjustments.

To address this concern, California adopted an innovative approach to ensuring smooth price discovery in a load-based emission cap without an existing GHG emission tracking system. California imposed a carbon procurement adder as a near-term bridge to a load-based cap. A carbon procurement adder requires LSEs to consider the 'shadow price' of carbon (assumed to be $8/CO2) in their planning decisions.81 While a carbon procurement adder may

80 Gillenwater and Breidenich (2007, p. 5). In the PJM, the spot market accounted for about 40% of the total electricity load. RGGI, Initial Leakage Report, n. 10, p. 7 n.13.

81 California Public Utilities Commission (07/04/2005), Decision 05-04-025, 'Order Instituting Rulemaking to Promote Consistency in Methodology and Input Assumptions in Commission Applications of Short-Run and Long-Run Avoided Costs, Including Pricing for Qualifying Facilities,' p. 44; California Public Utilities Commission (16/12/ 2004), Decision 04-12-048. In addition to the carbon procurement adder, the California Public Utilities Commission is considering charging a fee for GHG pollution. California Public Utilities Commission (31/01/ 2008), BAQMD Regulation 3-334 & Schedule T (draft). It is unclear, however, how such a fee will interact with other GHG pollution abatement requirements.

not in and of itself minimize leakage, as a near-term bridge to a load-based cap, it does help LSEs discover the price of allowances in the cap-and-trade system, which will help minimize price volatility.

4.2.2 Emission portfolio standards

California and other western states also seek to reduce leakage by imposing emission portfolio standards (EPS) - essentially a performance requirement -requiring that LSEs meet an output-based GHG emissions standard per kilowatt hour.82 Such EPSs encourage investment in low GHG-emitting technologies to reduce the overall GHG emission rate of the LSE's mix of generated electricity.

An EPS will improve the per kilowatt hour emission of GHGs, but without an emissions cap, absolute GHG emissions can continue to grow as demand grows. Accordingly, an EPS is insufficient, standing alone, to ensure GHG emission reductions.

Even with an emissions cap, however, an EPS will only marginally impact emission leakage by reducing the supply of GHG-intensive electricity. In combination with a load-based emission cap, however, a GHG EPS could promote investment in renewable energy technologies and would capture GHG emissions from sources outside the regulated region. It could also promote energy efficiency initiatives by crediting such initiatives as zero emission generation.

Similar to the contract shuffling problem faced by a load-based emissions cap, an EPS may suffer from 'attribute shuffling.'83 An EPS would identify environmental attributes of electricity generation and separate those attributes from the underlying commodity. Since WCI does not yet cover all the sources serving the Western Interconnection, it is possible for LSEs to shuffle their low GHG-emitting attributes to ensure compliance with the EPS within the WCI region or California, while sending their high GHG-emitting attributes outside the regulated region.

As with contract shuffling, however, there is unlikely to be a sufficient surplus of low GHG-emission attributes available to meet California's demand, so attribute shuffling is not expected to be a significant problem in California. Nevertheless, California established a carbon procurement emis

82 Another type of EPS is the renewable portfolio standard (RPS), which requires that a certain percentage of an LSE's electricity comes from renewable electricity. California, Oregon, Washington, Arizona, Nevada, Colorado, Montana, and New Mexico have mandatory RPSs of varying degrees of stringency, while Idaho, Utah, and Wyoming do not. Pew Center on Global Climate Change (August 2007), 'States with Renewable Portfolio Standards,' being_done/in_the_states/rps.cfm.

83 RGGI, Initial Leakage Report, n. 10, pp. ES-10, 37.

sion rate - requiring long-term contracting generators providing electricity to California to meet a specified CO2 emission rate, in an effort to avoid attribute shuffling.84 California's carbon procurement emission rate is based on the CO2 emission rate of a combined cycle natural gas turbine (approximately 1,000 pounds of CO2/MWh).85 By tying the emission rate to the bundled commodity, California has successfully avoided the attribute-shuffling problem inherent with an EPS that is not contract-driven.86 This attempted solution, however, may create a disincentive for establishing long-term contracts, resulting in an increase in spot market purchases. As discussed above, the potential for dirty generators to hide behind cleaner generators in the spot market creates a potential regulatory problem, though one that is solvable.

0 0

Post a comment