Liability rules for international trading of greenhouse gas emissions quotas

Climate Policy 1 (2001) 85–108

Research

Liability rules for international trading of greenhouse gas emissions quotas Erik Haites a,∗ , Fanny Missfeldt b a

b

Margaree Consultants Inc., Toronto, Canada M5H 3X6 UNEP Collaborating Centre on Energy and the Environment, Risø National Laboratory, Roskilde, Denmark

Abstract To reduce the costs of mitigating greenhouse gas emissions in accordance with the Kyoto protocol, international trades of emissions quotas are allowed. The revenue from the sale of quotas may exceed the sanctions for non-compliance if these penalties are weak or poorly enforced. Under these circumstances emissions trading enables a country to benefit financially through non-compliance. To counter non-compliance due to trading a range of liability proposals have been suggested. Using a simple global model, we analyze the economic and environmental performance of these proposals for the first commitment period. In addition, the proposals are tested for their sensitivity to national circumstances and to market power. We find that penalties are sufficient to deter non-compliance if they are high enough and are effectively enforced. If the non-compliance penalties are weak or poorly enforced, the permanent reserve proposal is best able to ensure compliance by sellers at a cost similar to the competitive market equilibrium. While not sufficient to ensure compliance on their own, eligibility requirements and annual retirement of AAUs equal to actual emissions contribute to compliance at little or no cost. Hence, such provisions could complement other liability proposals. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Emissions trading; Liability rules; Kyoto protocol; Seller liability; Buyer liability

1. Introduction The Kyoto protocol establishes emissions limitation commitments for the period 2008–2012 for Annex I Parties that have ratified the United Nations Framework Convention on Climate Change

∗ Corresponding author. Tel.: +1-416-369-0900; fax: +1-416-369-0922. E-mail address: [email protected] (E. Haites).

1469-3062/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 1 4 6 9 - 3 0 6 2 ( 0 0 ) 0 0 0 0 6 - 1

86

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

(UNFCCC). 1 The commitments represent a reduction of approximately 5% from 1990 emissions or approximately 20% from projected emissions for 2008–2012. Each of these Parties receives an allocation of “assigned amount” equal to its commitment as listed in Annex B to the protocol. To comply with its commitment, a country must retire assigned amount units (AAUs) or other quotas equal to its actual emissions during the 2008–2012 period. 2 The protocol creates three mechanisms that Annex B countries can use to help meet their commitments at lower cost: joint implementation (JI), the clean development mechanism (CDM), and emissions trading. Emissions trading allows one Annex B country to sell some of its AAUs to another Annex B Party. If the seller subsequently does not have sufficient AAUs to cover its actual emissions it will be subject to the penalties for non-compliance. Revenue from the sale of AAUs may exceed the sanctions for non-compliance if these penalties are weak or poorly enforced. 3 Under these circumstances emissions trading enables an Annex B country to benefit financially through non-compliance. Liability proposals seek to ensure that non-compliance is not rewarded by limiting sales of AAUs to amounts surplus to the seller’s compliance needs. Liability proposals complement, but do not replace, sanctions for non-compliance. 4 Liability proposals focus on sales of AAUs, so they do not promote compliance by countries that are not sellers of AAUs. Sanctions for non-compliance apply to all Annex B countries, regardless of whether they have engaged in emissions trading. If an Annex B seller does not comply with its commitment despite the liability proposal adopted, it is still subject to the non-compliance sanctions. Over a dozen liability proposals have been suggested in the literature and international negotiations. 5 While many of these proposals have been subjected to critical review, there has been no quantitative analysis of the different proposals. The proposals are often complex and they affect the mixture of

1 Annex I Parties are the 41 industrialized Parties listed in Annex I of the United Nations Framework Convention on Climate Change (UNFCCC) that have endorsed the aim of stabilizing their year 2000 emissions at levels of 1990 (Article 4.2 UNFCCC). Under the 1997 Kyoto protocol 39 of the Annex I Parties would commit to the quantified emissions limits for the period 2008–2012 (Article 3 Kyoto protocol). Each country’s emissions limitation commitment is listed in Annex B of the Kyoto Protocol. Since the Parties listed in Annex B of the protocol are identical to the revised list of Annex I Parties to the Convention (Turkey and Belarus have not yet ratified the Convention) the terms ‘Annex B Parties’ and ‘Annex I Parties’ are used interchangeably. Although both lists include the European Community (as well as all of its member states), which is not a country, the terms Party and country are used interchangeably to apply to these Parties. 2 The term “quota” is used to refer to any or all of AAUs, ERUs issued for emission reductions achieved by Joint Implementation projects and CERs issued for emission reductions achieved by Clean Development Mechanism projects. 3 Chayes and Chayes, 1998 concludes that “sanctioning authority is rarely granted by treaty, rarely used when granted, and likely to be ineffective when used” (p. 32). 4 It has been argued that selling AAUs not surplus to compliance needs (overselling) is the same as failure to reduce emissions, so a liability provision to limit the overselling is not needed. The environmental impacts and economic rewards to the violator of one ton of excess emissions are similar whether this results from a failure to abate or from the sale of an AAU needed for compliance. However, the potential scale of these actions differs. Failure to abate is limited to the difference between a country’s “business-as-usual” emissions and its emissions limitation commitment. Under the Kyoto Protocol, this could be as much as 35% of the country’s commitment for some countries. Emissions trading allows a country to sell AAUs equivalent to 100% of its emissions limitation commitment. Liability proposals are intended to restrict the increased potential for non-compliance created by emissions trading if the penalties for non-compliance are weak. 5 Good reviews of the proposals are provided by Baron (1999), Hargrave et al. (1999) and Nordhaus et al. (2000a). Proposals submitted by countries to the international negotiating process are described in FCCC/SB/2000/4. Part Three. para 17, pp. 104–106.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

87

domestic action and greenhouse gas quotas purchased. Hence, it is difficult to assess the performance of the proposals without the aid of a model. A model is developed and used to analyze the different liability proposals. We adopt a global perspective, rather than the perspective of an individual seller, to the analysis of the liability proposals. This allows the proposals to be assessed on the basis of their impact on Annex B non-compliance and compliance costs. The model is based on the emissions projection and policy analysis (EPPA) model of the Massachusetts Institute of Technology (MIT) (Babiker et al., 2000). The EPPA model was used because it was relatively easy to develop a simplified version with the features needed to analyze the liability proposals. EPPA results tend to fall in the middle of the range for models that estimate the economic consequences of limiting greenhouse gas emissions (Weyant and Hill, 1999). Any analysis that uses a model is necessarily limited by the structure and assumptions of the model. The results presented should be interpreted with caution and should be only one input to the evaluation of the different liability proposals. A liability proposal may increase compliance costs for Annex B countries by restricting sales of AAUs below the full compliance level and thereby causing the buyer to use more costly actions to meet its commitment. Alternatively, a liability proposal may allow a seller to benefit financially through non-compliance, leading to non-compliance by the seller and lower compliance costs for the buyer. To be effective, a liability provision should deter deliberate efforts to benefit financially through non-compliance. Yet it should not increase compliance costs for participants that meet their commitments and adhere to the rules. To assess the effectiveness of the liability proposals, we, therefore, assess their performance under “worst” case assumptions — assumptions that lead to the highest excess emissions or the largest increase in compliance costs. Given these anticipated impacts, we assess the performance of the different proposals on the basis of • Excess emissions by the Annex B seller. The purpose of a liability provision is to limit sales of AAUs to amounts surplus to the seller’s compliance needs. Proposals that keep the seller’s emissions within 2% of the least-cost full-compliance level are preferred. 6 • Compliance cost for the Annex B buyer. The Annex B buyer bears a net cost, while the sellers earn net income from their quota sales, so we focus on the buyer’s compliance cost. Proposals that keep the buyer’s compliance cost within 5% of the least-cost full-compliance level are preferred. • Sensitivity of the liability proposal to national circumstances. The appropriate operational specification of a proposal may be sensitive to a country’s national circumstances and it may not be possible to determine the appropriate specification for each Annex B country accurately in advance. Thus, a liability proposal that is sensitive to national circumstances is not desirable. • Sensitivity to seller behavior. One or more of the Annex B sellers may be able to exercise market power by virtue of their share of the market for AAUs. A liability proposal that is not sensitive to market power is preferred. • Temporal impacts. A liability proposal may affect the date when trading of AAUs can begin or the quantity of AAUs available at different times during the commitment period. Proposals that allow trading to begin in 2008 and continue throughout the period are preferred.

6

Full compliance is assumed to be achieved voluntarily. This would be the case if the sanctions for non-compliance were judged to be serious by each Annex B country. Liability proposals are being considered precisely because this assumption may not be valid.

88

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

• The distribution of net income across regions. The mix of domestic action and quota purchases used by the buyer affects the compliance cost to the Annex B buyer and the distribution of net income. The compliance cost for Annex B buyers and net income for Annex B and non-Annex B sellers may affect the support for different liability proposals. Our analysis indicates that penalties can deter non-compliance through emissions trading if they are sufficiently high and are effectively enforced. The model suggests that the penalty should be at least 2.3 times the expected price. If the non-compliance penalties are weak or poorly enforced, several of the liability proposals can achieve results essentially equivalent to the least-cost, full-compliance equilibrium even assuming the worst behavior by the Annex B seller. The permanent reserve proposal is the one that best meets the criteria adopted. It requires Annex B sellers to establish a reserve equivalent to an estimate of total emissions during the commitment period. Trading is limited to AAUs surplus to that reserve. In practice, the size of the reserve should be capped at some percentage, say 95%, of each Annex B country’s assigned amount so that countries that are net buyers can also sell AAUs. A permanent reserve rule would mean that Annex B countries could not use AAUs as the allowances for domestic emissions trading programs. Annex B countries could, however, establish domestic emissions trading programs that use domestic allowances which can be exchanged for surplus AAUs. In Section 2 we present an overview of the model. Section 3 describes the different liability proposals analyzed and how they have been specified in the model. The performance of the different proposals is evaluated in Section 4 using the criteria defined above. The performance of different proposals is compared in Section 5. Policy implications and conclusions are presented in Section 6.

2. Overview of the model As shown in Fig. 1, the model includes a single Annex B buyer who seeks to minimize compliance costs, a single Annex B seller of AAUs and emissions reduction units (ERUs) and a single non-Annex B seller of certified emission reduction credits (CERs) both of whom seek to maximize their net income. The buyer represents all Annex B buyers and the sellers represent all Annex B and non-Annex B sellers.

Fig. 1. Model structure.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

89

As a result, the compliance costs and excess emissions are the global totals rather than those for a single buyer or seller. The model is limited to the first commitment period, 2008–2012. The buyer is assumed to be a firm located in an Annex B country that is likely to be a net importer of quotas (Annex II countries 7 ) faced with domestic obligation to limit its emissions. It must demonstrate compliance with its domestic obligation annually 8 and faces stringent domestic penalties for non-compliance. 9 The firm can implement domestic reductions or purchase international quotas (AAUs, ERUs and CERs) to achieve compliance. However, only AAUs that have no risk of being invalidated are accepted by the government for compliance purposes. 10 The supply of CERs from CDM projects is determined by the cost of emission reduction actions in non-Annex B countries, taking into account transaction costs and adjusted for banked CERs from projects begun prior to 2008. 11 Because CERs are subject to an international review process, they are not subject to the liability provisions. The marginal abatement cost curves are based on energy-related CO2 emissions only. Incorporating additional emissions sources and sinks into the analysis will reduce compliance costs (Reilly et al., 1999).

7

The 25 Annex I Parties listed in Annex II of the UNFCCC. These are the wealthier industrialized countries that were members of the OECD in 1990. Membership of the OECD has subsequently expanded to include Poland, Hungary and the Czech Republic that are part of the “rest of Annex B” region as well as Korea and Mexico that are part of the “non-Annex B” region. 8 Annual compliance is well established for environmental and other regulations. A longer compliance period increases enforcement problems due to turnover in the population of participating firms. A longer compliance period offers no advantages over annual compliance with banking, except for the possibility of borrowing within the period. Borrowing creates a risk of non-compliance by the government due to borrowing by sources that cease operation before the end of the period. With annual compliance, borrowing can be arranged through commercial transactions, which means the risk of non-compliance is borne by the participants rather than the government. 9 The penalties for non-compliance, like those under the United States acid rain program, consist of the loss of quota in the next year equal to the excess emissions plus a financial penalty. The financial penalty is the higher of US$ 100/tC (1995 US dollars) or three times the market price during the year the excess emissions occurred. 10 This assumption has significant implications for the timing of transactions. It means that under user (buyer) liability proposals the purchased quotas cannot be used until 2012 or later when the compliance status of the issuers is known. In practice governments might allow firms to use AAUs from specified “low risk” sellers, such as members of the same Article 4 “bubble”. However, the model is used to test the performance of the liability proposals under the most restrictive conditions. Less restrictive assumptions will lead to results that lie between those reported and the reference case. 11 The inventory of CDM quota from early projects is assumed to be proportional to the CDM reductions during each year of the commitment period. This means that the size of the CDM inventory changes with each case. It also means that the inventory is distributed over the 5 years, rather than being available in full at the beginning of the period. It could be argued that the analyses of the different proposals are not strictly comparable because of the differences in the CERs inventory. On the other hand, if the liability proposal affects the mix of quotas used for compliance, this is likely to affect CDM activity prior to 2008 and the inventory of CERs. If the CDM generated the same quantity of emission reductions each year from 2000 to 2012, the inventory available at the beginning of 2008 would be 7 years’ supply. When this inventory is spread over the 5 years from 2008 to 2012, the inventory represents 7/5 = 1.4 years’ supply for each year of the commitment period. Thus, to add the inventory to the supply created in a given year, the latter must be multiplied by 2.4. The rate at which CDM projects generate emission reductions is likely to increase over time, so the inventory adjustment is assumed to be <2.4. Since the model is limited to the period 2008–2012, the cost of the CERs generated prior to 2008 is not calculated. Since the Annex B buyer purchases the CERs during the commitment period, the cost to the buyer and the revenue to the non-Annex B seller are included. But the net revenue to the non-Annex B seller is overstated because the pre-2008 costs of generating CERs have not been deducted from the revenue it receives.

90

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

The supply of ERUs from JI projects is limited to a share of the domestic reduction actions available to the Annex B seller taking into account transactions costs. 12 ERUs are assumed to be subject to an international review process similar to that for the CDM and hence are not subject to the liability provision. 13 The supply of AAUs available is the total amount issued to the Annex B seller less any ERUs and AAUs previously sold, unless constrained by the liability proposal. The Annex B seller implements no domestic reductions, other than JI projects, unless this increases the quantity of AAUs that can be sold under the liability proposal. 14 Supply and demand are equated by comparing the buyer’s demand with the supply (domestic reductions by the buyer plus CERs, ERUs and AAUs) starting at the lowest price (US$ 1/tC) and continuing in US$ 1/tC increments until the supply is sufficient to meet the demand. Actual emissions for a given year are the “business-as-usual” emissions of the buyer or seller as projected by the model less any domestic emission reductions. Actual emissions of the sellers are only known with a 2-year lag. Compliance is assessed by comparing quota (CERs, ERUs and AAUs) holdings with actual emissions. The compliance cost for the buyer is the cost of domestic actions implemented plus the market price of the quotas (CERs, ERUs and AAUs) purchased plus any penalties for non-compliance less the value of any banked quotas at the end of the period. 15 The net income of a seller is the revenue from the sale of quotas less the cost of domestic actions implemented less any penalties for non-compliance plus the value of any banked quotas at the end of the period. The compliance cost (net income) is expressed as the present value in 2008 of the costs (net income) during the period discounted at 5% net of inflation. All monetary values are 1995 US dollars. The least-cost full-compliance reference case of the model, which serves as a base case for comparison with the liability proposals, assumes • Full compliance by the Annex B buyer and seller annually. • Interpolated “business as usual” emissions for the years 2008–2012. 16 • Transactions costs 15% for JI and 25% for CDM. • An inventory of CDM quota due to actions implemented prior to 2008 equivalent to 80% of the reductions achieved each year during the commitment period. • A total of 25% of the potential emission reductions in Annex B seller countries can be implemented as JI projects. • No restrictions on international quota trade. The reference case price rises from US$ 26 in 2008 to 42/tC in 2012, which is substantially lower than under the price of US$ 47.70/tC in 2010 under the EPPA assumptions. The reason for the lower prices is availability of early CDM quotas in our reference case. The reference case includes transaction costs on JI and CDM and limits on JI supply, which tend to increase prices. However, the inventory of CERs due 12

When JI quotas are available they are always valid and so are at least as attractive to buyers as AAUs. This will encourage the Annex B seller to structure as much of the available emission reductions as JI projects. Thus, JI quotas could account for a relatively large share of the available emissions reductions. 13 The implications of leaving the review process entirely to the host government are discussed in Section 5. 14 For example, with a non-compliance penalty of US$ x/tC, the Annex B seller can increase its net revenue by implementing domestic reductions with a cost of
E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

91

to reductions prior to 2008 reduces prices by an even larger amount. So the net effect is lower prices in our reference case than in the EPPA global trading case. 17 3. Description of the liability proposals The proposals use several different strategies for limiting sales of AAUs to amounts surplus to the seller’s compliance needs. We group the proposals into the following six categories: • Sanctions for non-compliance Penalties for non-compliance Eligibility requirements • Limits on sales Annual retirement of AAUs equal to actual emissions Limit on sales • Trading limited to surplus quota Sales prohibited until compliance established Permanent reserve Swiss proposal • Restoration of default Compliance reserve Compulsory insurance Escrow account • User liability User liability Shared liability Double liability • Progressive response to probability of default Traffic light The proposals in the first four categories assume issuer (seller) liability — the Annex B government that first issues the AAUs bears the consequences for selling quotas and then not achieving compliance. 18 The proposals in the first category rely on ex-post penalties or eligibility requirements to encourage compliance by the issuer. The proposals in the second category seek to keep the issuer in compliance by limiting the quantity of quota it can sell. Proposals in the third category define the quota likely to be surplus to the seller’s compliance needs and allow that surplus to be traded. The fourth category includes various proposals to bring the issuer back into compliance if a default occurs. 17 If the inventory of CDM quota is eliminated from the reference case, the prices rise from US$ 39 to 61/tC. If all of the transaction costs are zero and all reductions in the non-OECD Annex B countries can generate JI quota, the prices range from US$ 23 to 37/tC. 18 In the context of this model, issuer and seller liability are the same, because there is only one Annex B seller. Therefore, there is no possibility for secondary trading. The same holds for user/buyer liability.

92

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

The proposals in the fifth category rely on user (buyer) liability to keep the Annex B government that first issues the AAUs in compliance. User liability proposals return some, or all, of the quotas to the issuer as necessary to bring the issuer into compliance. This provides buyers with an incentive to purchase quotas from issuers likely to achieve compliance. Finally, the traffic light proposal in the last category allows trading on an issuer liability basis until a defined condition, indicating a risk of non-compliance, is reached. Thereafter, further sales are made on a user liability basis. Many of the proposals have not been defined in operational terms. We analyzed alternative operational specifications for each proposal. The operational specifications for each liability proposal that yield the highest level of compliance by the Annex B seller, assuming a competitive market, is listed in Table 1. More complex proposals, that include a combination of two or more of the above proposals, have not been evaluated. 19 3.1. Penalties for non-compliance Compliance by the issuer can be increased by sanctions for non-compliance and minimum eligibility criteria to participate in emissions trading. The penalty acts as a deterrent for non-compliance, while eligibility criteria attempt to limit trading to participants that exhibit ‘sound’ behavior. The procedures for determining non-compliance and the consequences for non-compliance with the emissions limitation commitments by Annex B Parties are not yet defined (UNFCCC, 2000a). Article 18 of the protocol states that the consequences should reflect the “cause, type, degree and frequency of non-compliance” (UNFCCC, 1997). Proposed consequences include financial penalties, reduction of the Party’s assigned amount for the subsequent commitment period, loss of access to the mechanisms, and suspension of rights and privileges. 20 But since participation is voluntary, the penalties can not be too onerous or the offending Party may simply withdraw from the protocol. The model is limited to the first commitment period, so measures that apply during the subsequent commitment period cannot be modeled explicitly. 21 To analyze the effect of sanctions for non-compliance, we model monetary penalties from US$ 1/tC to 100/tC representing the expected value to the Annex B seller of the sanctions imposed. Eligibility criteria proposed for participation in emissions trading include • Compliance with Article 5 of the Kyoto protocol, maintaining an acceptable national emissions inventory. • Compliance with Article 7 of the Kyoto protocol, reporting actual emissions and other information. • Being a Party to a compliance regime adopted under Article 18 of the Kyoto protocol. • Maintaining a registry to track AAUs and ERUs that meets specified standards (UNFCCC, 2000b). Since the proposed eligibility conditions relate, at least in part, to the adequacy of the emissions inventory we assume a 2-year lag between the event that triggers loss of eligibility and the date when participation 19

See Climate Action Network (2000), for example. The assigned amount for the next commitment period could be reduced by (100 + X%) where X% (say 5–50%) represents a penalty for non-compliance. This requires that the commitments for the subsequent period are negotiated well in advance of the end of the current period. Otherwise a Party that expects that it might be subject to such a penalty could seek to negotiate a higher commitment to reduce the impact of the anticipated penalty. Article 3.9 of the protocol specifies that consideration of commitments for the second commitment period will begin no later than 2005, so it is possible that agreement on the commitments for the subsequent period could be reached prior to the start of the first commitment period. 21 CCAP (2000) proposes annual retirement plus a series of penalties to take effect during the second commitment period reflecting the severity of the non-compliance during the first commitment period. 20

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

93

Table 1 Operational specifications of the liability proposals analyzed Penalty for non-compliance • Penalty > US$ 40/tC Eligibility requirements • Eligibility established prior to 2008 but lost at the end of 2009 • Eligibility not established until the beginning of 2012 Annual retirement of AAUs equal to actual emissions • Annual retirement of AAUs equal to actual emissions with a 2-year lag Limit on sales • Sales of AAUs during a given year limited to 20% of the average assigned amount Sales prohibited until compliance established • No sales of AAUs prior to 2014 Permanent reserve Sales are limited to AAUs surplus to a permanent reserve equivalent to projected emissions during the commitment period. Projected emissions are calculated as • Option 1. After 5 years of emissions calculated from the most recent data available adjusted for actual emissions as the data become available during the commitment period, i.e. five times the 2006 emissions in 2008 and 2009, four times 2006 emissions plus 2008 emissions when the latter become available in 2010 • Option 2. The sum of projected emissions for the years 2008–2012 as estimated from a regression equation fitted to actual emissions for the years 2000 through 2006 Swiss proposal Sales limited to AAUs surplus to a compliance plan defined by the seller • Regular start — AAU sales begin in 2010 based on a comparison of actual emissions with the compliance plan for 2008. The compliance plan is defined as average assigned amount +33% in 2008 declining to average assigned amount −33% in 2012 • Prompt start — AAU sales begin in 2008 based on a comparison of 2006 emissions with the compliance plan for 2008 and so on. The compliance plan is defined as average assigned amount for each year Compliance reserve With each sale of AAUs, the seller deposits a specified quantity of its remaining AAUs into a reserve that can be used only for the seller’s compliance needs • Option 1. A reserve requirement of 300% with sales limited to the average assigned amount • Option 2. A reserve requirement of 1600% with sales limited to the remaining assigned amount Compulsory insurance • AAU premium. The Annex B seller pays a “premium” of one AAU to the insurance company for each AAU sold • Insurer purchases AAUs. The insurer purchases enough quota each year so that the Annex B seller’s remaining assigned amount and the insurer’s quota purchases are equal to the seller’s projected emissions. To help achieve compliance, no emissions trading is allowed in 2012 Escrow account Revenue from the initial sale of AAUs is deposited into an account where it is held until the issuer establishes compliance. If the seller does not achieve compliance, the funds are used to purchase the amount of quota needed to bring the seller into compliance or the amount that can be purchased with the available funds, whichever is lower. The seller sets a minimum price of US$ 20/tC for AAUs User liability The buyer purchases AAUs for compliance use at its risk. If the Annex B seller does not achieve compliance, some, or all, of the AAUs purchased are returned to the issuer. It is assumed that the buyer can only use the AAUs in 2012 after the seller has established compliance

94

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

Table 1 (Continued) Shared liability If the Annex B issuer (seller) does not achieve compliance, part of the AAUs purchased are returned. Three alternatives with liability shared 50%, but different penalties and seller responses are analyzed Double liability If the Annex B issuer (seller) does not achieve compliance, AAU sales are invalidated as necessary to bring the issuer into compliance in the same manner as under user liability. In addition, the seller is subject to penalties for non-compliance Traffic light Sales are made on an issuer liability basis until a specified a risk of non-compliance is reached. Thereafter, sales are on a user liability basis. The conditions that trigger the switch from issuer to user liability are • Limit of 14%. AAU sales of 14% of assigned amount • Limit of 2%. Cumulative actual emissions plus AAU sales Ω [(assigned amount/5) × years of data available + 2% of assigned amount/years of data available]

in emissions trading is restricted. Rather than model the eligibility criteria explicitly, we consider the effect of non-compliance with eligibility requirements. Specifically, we test the effect of losing eligibility to trade after 2009 and of first gaining eligibility to trade in 2012. 3.2. Limits on sales of AAUs The rationale for limiting sales of AAUs to deter non-compliance is that the less you sell, the smaller the extent of the non-compliance, if any. Two ways of limiting sales have been discussed: ‘annual retirement’ of quotas equal to actual emissions (CCAP, 2000) and limiting sales based on a specified formula. The latter stems from a proposal made by the European Union in the context of ensuring that the use of flexible mechanisms would be supplemental to domestic action (UNFCCC, 1999). 22 Under an annual retirement scheme, the quantity of quota available for sale is the seller’s total assigned amount for the commitment period less cumulative emissions and cumulative sales of AAUs and ERUs. Sales of AAUs and ERUs reduce the quantity of AAUs available for sale immediately. The seller’s actual emissions reduce the quantity of AAUs available for sale with a 2-year lag. The formula for limiting sales determines the maximum quantity of AAUs that can be sold each year. The limit is specified as a percentage of the “average” assigned amount available for each of the 5 years of the commitment period. The EU proposal constitutes a limit of approximately 5%. A limit of 20% keeps the Annex B seller in compliance. 3.3. Trading limited to surplus quota The proposals in this category establish a budget equal to the quotas likely to be needed to achieve compliance and then allow the balance to be traded. Three proposals of this type are investigated. The most restrictive is the first proposal, which prohibits any trading until compliance is established and then limits sales to surplus quota. We assume a 2-year lag in reporting actual emissions by the 22 The precise European Union proposal to limit transfers under the supplementarity clause is: sales of AAUs during a given year are limited to a maximum of 5% of (1990 base year emissions + (assigned amount/5))/2.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

95

Annex B seller, so compliance cannot be established until 2014. This means no sales of AAUs during the commitment period. A second proposal is to establish a ‘permanent reserve’ of quotas that may not be traded until they are shown to be surplus to the country’s compliance needs after the end of the commitment period. The permanent reserve includes sufficient quotas equal to 5 years of projected/actual emissions. Two options for calculating the size of the reserve are modeled: • A permanent reserve equal to five times the 2006 emissions initially with adjustments based on actual emissions during the commitment period as the data become available — e.g. four times 2006 emissions plus 2008 emissions when the latter become available in 2010. • A permanent reserve equal to the sum of the projected emissions for the years 2008–2012 as estimated from a regression equation fitted to actual emissions for the years 2000–2006. 23 The third proposal in this category is the ‘Swiss Proposal’ which was proposed by the Swiss government (UNFCCC, 1999). Under this proposal, an Annex B country that wishes to sell AAUs must submit a compliance plan to the Climate Change Secretariat before the start of the commitment period. The compliance plan specifies how the country’s assigned amount is to be allocated over the 5 years of the commitment period. 24 When the country’s actual emissions for a given year are submitted to the Secretariat, they are verified. If the compliance plan exceeds the verified actual emissions on a cumulative basis, the difference can be sold. 25 The compliance plan has the following structure for the years 2008–2012     AA Y AA AA Y AA AA + Y %, + % , , − % , − Y% (1) 5 5 2 5 5 2 5 where AA is the country’s assigned amount for the commitment period and Y is a deviation from the average assigned amount. Switzerland has suggested a range of ±20%; i.e. Y = 20. We find that the best performance for the Annex B seller is achieved with Y = 33 and present our results for that value. Since the seller’s actual emissions are assumed to be known only with a 2-year lag, sales of AAUs cannot begin until 2010. We also test the performance of a prompt start version of the Swiss proposal where the assigned amount allocated to 2008 is compared with actual emissions for 2006, etc. The prompt start version performs best with Y = 0, meaning that the assigned amount is allocated equally over the 5 years. 3.4. Restoration of default Three of the proposals tested here focus on the restoration of damages from defaulting quota sales: a compliance reserve, a compulsory insurance, and an escrow account. 23 A regression equation is fitted to randomly adjusted actual emissions of the Annex B seller for the years 2000–2006 and used to project the values for 2008–2012. 24 Countries are allowed to change their compliance plan during the course of the commitment period subject to the restriction that the compliance plan cannot be less than the actual emissions plus AAUs and ERUs sold prior to the revision. 25 In 2010 when actual emissions for 2008 are known they are compared with the compliance plan for 2008. Any surplus AAUs can be sold. In 2011 emissions for 2008 plus 2009 are compared with the compliance plan for 2008 and 2009. If under the cumulative compliance plan emissions exceed the cumulative emissions, any AAUs sold during 2008 are subtracted to get the AAUs available for sale in 2009, and so on.

96

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

The compliance reserve proposal requires the Annex B seller to deposit AAUs equal to a multiple of the quantity sold into a compliance reserve. Only the difference between the country’s available assigned amount and the compliance reserve can be sold. We tested two options: • The available assigned amount each year is the total assigned amount divided by 5. A compliance reserve equal to 300% of the AAUs sold works best in this case. • The available assigned amount is the assigned amount remaining after previous sales of ERUs and AAUs and the associated compliance reserve. A compliance reserve equal to 1600% of the AAUs sold works best in this case. Under the compulsory insurance proposal the buyer purchases AAUs at its risk, but is required to purchase insurance that replaces some or all of the AAUs purchased if the Annex B seller does not achieve compliance. 26 The insurance premium should reflect the risk of non-compliance by the seller. 27 The seller would bear the cost of the insurance coverage. 28 The seller can reduce the insurance premiums for sales of its AAUs, and hence increase its revenue, by implementing policies to reduce the risk of non-compliance. One such policy is to limit sales of AAUs to quantities surplus to its compliance needs. Since the model does not include a separate insurer, two cases were analyzed as different approximations of the impact of compulsory insurance. The first requires the seller to pay the insurer a premium in the form of AAUs equal to the quantity sold. Under the second proposal the insurer purchases enough quotas so that the Annex B seller’s remaining AAUs and the insurer’s quota purchases are equal to the seller’s projected emissions. To help achieve compliance in this case, no trade is allowed in 2012. The escrow account proposal requires that the revenue from the initial sale of AAUs be deposited into an account where they are held until the issuer establishes compliance. 29 Interest is earned by the funds in the account. If the issuer achieves compliance at the end of the commitment period, it receives the funds in the account. If the seller does not achieve compliance, the funds are used to purchase the amount of quota needed to bring the seller into compliance or the amount that can be purchased with the available funds, whichever is lower. Any remaining funds are paid to the seller. 3.5. User liability User liability motivates the buyer to purchase only from sellers likely to achieve compliance. Under the user liability proposal, the buyer purchases AAUs for compliance use at its risk. If the Annex B seller does not achieve compliance, some, or all, of the AAUs purchased are returned to the issuer. 30 The issuer 26

See Haites (1998) for a discussion of this proposal. This is feasible only if a last-in-first-out (LIFO) approach is used to invalidate transactions in the case of non-compliance. This allows the insurer to assess the risk of non-compliance at the time of the sale. Proportional invalidation of sales would mean that subsequent sales would affect the risk of invalidation of the current sale so it is not possible to establish a premium for the insurance coverage. On the other hand, a LIFO approach means that the risk of invalidation of the current sale is not affected by subsequent sales. 28 Buyers are able to purchase CERs and ERUs with no risk of invalidation. Hence the buyer will only pay the equivalent price for fully insured AAUs. This means that the price received by the seller is reduced by the cost of the insurance coverage regardless of whether the insurance is actually purchased by the buyer or the seller. 29 See Haites (1998) for a discussion of this proposal. 30 The sales agreement between the buyer and the seller would include provisions for reimbursement, and possibly penalties, in the event the transaction was invalidated. Those arrangements are not central to the liability proposal, which would be implemented by making all transfers subject to final approval by the secretariat once compliance had been established. However, if buyers insist on penalties in the event a transaction is invalidated, it could reduce the volume of AAU sales significantly. 27

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

97

uses the returned AAUs to achieve compliance. 31 The buyer may need to acquire other quota to achieve compliance. It is assumed that a grace period would be provided for this purpose. Pure user liability is modeled by allowing the buyer to purchase AAUs up to its projected needs for 2012 during earlier years. The price for such quota purchases is discounted to reflect the risk that the transaction will be invalidated. To make a purchase decision, the buyer projects the seller’s actual emissions on the basis of the available (with a 2-year lag) data and ERU sales (with no lag). The price discount is applied to the market price during the year in which the purchase is being considered. The non-compliant quota may be invalidated on the basis of last-in-first-out (LIFO) or on a proportional basis. This indirectly affects the price at which these quotas are offered during the commitment period. When liability is shared, the risk of non-compliance by the issuer is shared between the buyer (user) and the issuer. If the Annex B issuer does not achieve compliance, some of the AAUs purchased are returned. It is assumed that sufficient transactions are invalidated to bring the issuer into compliance. The buyer keeps the balance of the AAUs purchased, but may need to acquire additional quotas to achieve compliance. It is assumed that a grace period would be provided for this purpose. To model the shared liability proposal the user liability share of AAUs purchased is returned to the Annex B seller if the issuer does not achieve compliance. The balance of the AAUs can be used by the buyer for compliance in 2012. 32 Four alternative specifications of these proposals, which included specifications with minor financial penalties, were tested (Haites and Missfeldt, 2000). To model double liability the Annex B issuer’s AAU sales are invalidated as necessary to bring the issuer into compliance in the same manner as under user liability. In addition, the seller is subject to penalties for non-compliance. It is not clear whether the issuer is subject to sanctions based on the extent of the non-compliance before the transactions are invalidated or the extent of the non-compliance after the transactions have been invalidated. We interpret double liability to mean that the issuer will be subject to sanctions for the extent of the non-compliance before the transactions are invalidated because Annex B Parties will be subject to sanctions for non-compliance under Article 18 regardless of whether they participate in emissions trading. 33 3.6. Progressive response to probability of default The ‘traffic light’ proposal is based on a progressive response to the performance of the issuer within the trading regime. It allows sales to be made on an issuer liability basis until a defined condition, indicating a risk of non-compliance, is reached. Thereafter, further sales are made on a user liability basis. If a condition that suggests a high risk of non-compliance is reached, further sales are prohibited. The working of the ‘traffic light’ approach has been discussed in a number of papers, including CIEL (1999). 31

If the AAUs are returned to the seller, the transaction is invalidated, the seller is brought into compliance to the extent possible, and the buyer may need to purchase other quotas to achieve compliance. It is also possible to invalidate the AAUs. Then, both the buyer and the seller need to purchase replacement quotas to achieve compliance. 32 Ideally, the buyer should be able to use the issuer liability share of the AAUs purchased for compliance during the year they are purchased rather than 2012. When the proposal was programmed in this way it led to a “loop” with the use of the AAUs affecting the level of non-compliance, which in turn affected the fraction of AAUs that could be used for compliance, which affected the level of non-compliance. 33 Imposing the sanctions on the Annex B seller based on the extent of its non-compliance after invalidation of the transactions is identical to user liability. To create a stronger incentive for the seller to comply, double liability must impose sanctions based on the extent of its non-compliance before the transactions are invalidated.

98

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

To model this proposal two different triggers were tested as the switch from issuer to user liability. • AAU sales limited to 14% of assigned amount. • Cumulative actual emissions plus AAU sales, i.e. ((assigned amount/5) × years of data available + 2% of assigned amount/years of data available]. The percentages chosen, 14 and 2%, provide the best results in terms of keeping the Annex B seller in compliance. Until the applicable limit is reached, sales proceed without restriction on an issuer liability basis. Beyond the applicable limit AAU sales are subject to user liability with LIFO invalidation. All AAUs sold on an issuer liability basis can be used by the buyer during the year they are purchased. AAUs purchased under user liability can only be used in 2012 if the seller achieves compliance.

4. Proposal by proposal evaluation of liability In this section, we evaluate the performance of each of the liability proposals using the criteria discussed in Section 1. The first two columns of Table 2 indicate whether the proposal is able to keep the Annex B seller’s excess emissions to within 2% of the maximum possible non-compliance of 4029.3 MtC (80 MtC) and to within 5% of the buyer’s reference case compliance cost of US$ 141.16 billion. 34 The third column of Table 2 indicates whether the operational specification of the proposal needed to ensure good environmental and economic performance is sensitive to national circumstances. National circumstances are modeled as differences in the ratio of the Annex B seller’s emissions limitation commitment to its projected “business-as-usual” emissions. The fourth column of Table 2 indicates whether the effectiveness of the proposal is sensitive to the ability of the Annex B seller to exercise market power. Proposals that are not sensitive to market power by the seller are preferred. The last column of Table 2 indicates whether the proposal allows AAUs to be traded during the entire commitment period. This could affect compliance costs for firms required to comply on an annual basis. The sixth criterion, the compliance cost and distribution of net income among sellers, is discussed in Section 5. 4.1. Penalties for non-compliance The Annex B seller adjusts its behavior in response to the penalty. If the penalty is below the market price, it becomes the minimum price at which AAUs and ERUs are sold. If the penalty is above the market price, the Annex B seller calculates the quantity needed for compliance and sells only AAUs and ERUs in excess of this amount. A penalty of US$ 40/tC, or more, leads to full compliance. This behavior yields the widely accepted result (see, for example, Shavell, 1987) — that sanctions lead to compliance if the penalty faced by the seller is higher than the market price and is enforced. In that case, compliance is less costly than the penalty. A penalty above the market price does not increase compliance costs because compliance remains the less costly option. 34

The cost range of 5% lies within the range of uncertainty for JI and CDM transactions costs (−5.4% if transactions costs for both JI and the CDM are 10% and +8.3% if the transactions costs for JI and the CDM are 35 and 50%, respectively), and is much smaller than the uncertainty (−18.9 to +11.7%) due to the supply of CDM quota available from prior to 2008.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

99

100

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

We find that the penalty for non-compliance should be substantially higher than the estimated market price for quota during the compliance period. Since the penalty must be set before the compliance period begins, the market price can only be an estimate. To get an estimate of an appropriate penalty, all parameters are set to the values that yield the highest price. That approach suggests that the present value of the non-compliance sanctions must be at least 2.3 times the expected price to be effective for all countries. If the Annex B seller is able to exercise market power, the market price will be higher and the penalty needs to be correspondingly higher. Thus, the sanctions are sensitive to market power by the seller, but this can be addressed by setting the penalty sufficiently high. The possibility that the sanctions for non-compliance faced by Annex B Parties under the Kyoto protocol may be lower than required, or may not be effectively enforced, is the motivation for the other liability proposals. If the Annex B seller is eligible to engage in emissions trading during 2008 and 2009 it can sell AAUs that it will ultimately need for compliance, so the first criterion is not met. The excess sales during 2008 and 2009 reduce the market price during those years, but the loss of eligibility prevents sales of AAUs from 2010 to 2012 and increases the market price during this period. The result is that the compliance cost to the Annex B buyer is approximately the same as for the least-cost, full-compliance reference case. If the Annex B seller is not eligible to engage in emissions trading until 2012, the demand during that year will limit sales of AAUs to less than the seller’s surplus. But the fact that AAUs are not available during the first 4 years of the commitment period raises the market prices during these years. This increases the compliance cost to the Annex B buyer substantially. 4.2. Limits on sales of AAUs Annual retirement is not effective in limiting sales of AAUs to the amount surplus to the seller’s compliance needs due to a combination of two factors. First, the quantity of AAUs available for sale (the seller’s assigned amount) is more than five times the seller’s annual emissions. This means that several years of emissions and quota sales must be deducted from the total before sales are limited. Second, actual emissions are only known with a lag of 2 years, which allows more sales to occur before the available supply is restricted. The result is that annual retirement does not limit the quantity that can be sold until 2012, by which time the seller has already sold a substantial quantity of the AAUs it will need for compliance. A limit on annual sales of AAUs equal to 20% of the seller’s total assigned amount divided by 5 is effective in limiting sales to amounts surplus to the seller’s compliance needs. The restriction affects the quota prices during the period enough to increase compliance costs for the Annex B buyer by over 12%. The EU proposal increases compliance costs to the Annex B buyer by almost 45% and forces the Annex B seller to bank nearly 75% of its surplus AAUs. The main difficulty with a limit on sales specified in this way is that the appropriate percentage is different for each country due to differences in the emissions limitation commitment, business as usual emissions, and marginal abatement costs. Furthermore, the limit that keeps the Annex B seller in compliance differs significantly from the one that maximizes the seller’s revenue. So each Annex B seller has an incentive to argue for a limit higher than the one that keeps it in compliance. This makes negotiation of a set of country-specific limits to maintain compliance by all Annex B Parties unlikely.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

101

4.3. Trading limited to surplus quota Prohibiting sales of AAUs until the Annex B seller establishes compliance in 2014 means that emissions trading under Article 17 cannot be used for compliance purposes during the first commitment period. This forces the Annex B seller to bank all of its surplus AAUs (1215 MtC). The Annex B buyer must rely more heavily on domestic action and purchases of CERs and ERUs increasing its compliance cost by over 50%.The permanent reserve proposal is effective in limiting sales of AAUs to amounts surplus to the seller’s compliance needs at a small increase in compliance costs or a small risk to the environment, depending upon the formula chosen to calculate the size of the reserve. The difference in performance for the two options as methods of calculating the permanent reserve is marginal. The formula for calculating the size of the permanent reserve is not sensitive to national circumstances. In addition, the results are very similar regardless of the market power of the Annex B seller. The parameters specified for both versions of the Swiss proposal lead to excess emissions of <2% which reduces compliance costs by almost 3% from the least-cost, full-compliance reference case. However, the value of the parameter that keeps the Annex B seller in compliance is different for each country. An Annex B seller maximizes its revenue by defining a compliance plan that leads to banking of a substantial part of its AAUs, which raises compliance costs for the Annex B buyer. If the Annex B seller is able to exercise market power, it maximizes its net income by banking 400–700 MtC under either version of the proposal. The prompt start leads to more stable prices and slightly lower compliance costs than the regular start. 4.4. Restoration of default Both options of the compliance reserve proposal can keep the Annex B seller in compliance at negligible added cost to the Annex B buyer. However, to achieve this result requires a different reserve requirement for each Annex B seller. It is impossible to determine the optimal reserve requirement for each seller in advance. Each Annex B seller has an incentive to argue for a requirement that is higher than the level needed to ensure compliance because this raises its net income. If the requirement is too high, sales of AAUs that are surplus to the seller’s compliance needs are restricted and the buyer’s compliance costs are increased. If the Annex B seller is able to exercise market power, it has an incentive to argue for an even higher reserve requirement. The compulsory insurance proposals do not perform well, possibly because the model is not well designed to evaluate these proposals. A requirement to pay the insurer a premium in the form of an AAU for each AAU sold does not create an incentive for sellers to limit their AAUs sales to quantities surplus to their compliance needs because giving AAUs to the insurance company does not impose a cost on the seller. The insurance company purchases lead to higher prices which leads to large increases in compliance costs for the Annex B buyer. The model is not well suited to an analysis of the performance of the escrow account proposal. The quantity adjusting behavior of the Annex B seller as a price taker cannot be modeled. The escrow account proposal can be effective in achieving compliance only if the sellers are able to exercise market power and estimate the appropriate minimum price. However, if the sellers have sufficient market power to affect the market price, it would be in their interest to demand a higher price for the AAUs.

102

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

4.5. User liability User liability leads to full compliance, but at a cost that is about 20% higher than for the full-compliance reference case. This is due to the assumption that the AAUs can only be used in 2012 after compliance by the seller has been established. This means that the buyer must rely on domestic reductions and ERU and CER purchases for compliance in 2008–2011, which raises the compliance costs for those years. Analysis of the shared liability proposal indicates that full user liability always performs at least as well as any shared arrangement. As in the case of user liability, the assumption that the AAUs can only be used in 2012 after compliance by the seller has been established raises compliance costs for the Annex B buyer. In general, the larger the issuer share of the liability, the greater the extent of the non-compliance. The analysis of the double liability proposal suggests that the penalty on the seller yields no additional benefit when compared with pure user liability. Arguably this result reflects the model assumptions. The buyer is assumed to be risk averse due the high domestic penalties it faces. In addition, the government is assumed to accept AAUs only if there is no risk of the transaction being invalidated. Thus, in the model, transactions are virtually never invalidated under user liability proposals. This means that the penalty faced by the issuer does not affect the results. 4.6. Progressive response to probability of default The traffic light proposal leads to over-compliance of about 250 MtC, which increases compliance costs for the buyer by over 25% from the full-compliance reference case. The limit that triggers a change in status varies with each country’s national circumstances. It would be difficult to find the most appropriate limit for each country in advance. The results are also sensitive to the ability of the Annex B seller to exercise market power. 5. Criteria-based assessment of liability proposals This section compares the performance of different liability proposals. Then, it discusses the sensitivity of the results to the assumptions, the compliance cost and distribution of income among sellers, and whether ERUs should be subject to the liability provision. 5.1. Environmental and economic performance The performance of the liability proposals in terms of excess emissions and compliance costs is shown in Fig. 2. The reference scenario — least-cost full compliance — is defined as the origin. The lower left quadrant is preferred since this means both lower compliance costs and lower total emissions compared with the reference case. Two of the proposals fall into this quadrant very close to the origin. This is surprising since it means that a liability proposal, where sales are restricted, achieves compliance at a lower cost than in the reference case, where there is no restriction on trade. Closer examination reveals that these results are due to smaller rounding errors than the 38 MtC for the reference case. Proposals in the upper right quadrant are undesirable because they involve both higher emissions and compliance costs. Variants of the escrow account, compulsory insurance and the Swiss proposal fall into this quadrant.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

103

Fig. 2. Comparison of liability proposals.

Proposals in the upper-left and lower-right quadrants involve a trade-off between compliance costs and excess emissions. At the far left-hand side of the chart are proposals that reflect a compliance effort of US$ 1/tC. They have the maximum excess emissions (4029.28 MtC = 100%) and reduce the Annex B buyer’s compliance costs to about 3% of the reference case. These proposals include issuer liability with a penalty of zero and a non-binding limit on sales. At the lower right of the chart are proposals that restrict AAU sales to levels below the amount surplus to the seller’s compliance needs. These proposals force the Annex B seller to bank some or all of its surplus AAUs, reducing total emissions below those of the reference case and increasing buyer’s compliance costs. A prohibition on sales until after compliance is established in 2014 leads to banking of about 30%

104

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

and increases compliance costs by over 50%. The European Union’s proposal to limit sales leads to banking of over 20% and increases the buyer’s compliance costs by almost 45%. The important result from Fig. 2 is that several of the liability proposals lie very close to the origin, meaning they lead to outcomes very close to the least-cost, full-compliance reference case. In other words, under the worst case assumptions several of the liability proposals protect against deliberate abuse of emissions trading for a small “premium”, which can take the form of increased compliance costs or a risk of non-compliance. The proposals that lie close to the origin in Fig. 2 are those that have check marks in the first two columns of Table 2, namely • Penalties of at least US$ 40/tC for non-compliance. • Permanent reserve, with the size of the reserve defined by option 1. • Permanent reserve, with the size of the reserve defined by option 2. • Swiss proposal with a regular start and annual emissions up to 33% above the average assigned amount. • Swiss proposal with a prompt start and annual emissions equal to the average assigned amount. • Compliance reserve defined by option 1 with a limit of 300%. • Compliance reserve defined by option 2 with a limit of 1600%. • Escrow account with a minimum price of US$ 20/tC. 5.2. Sensitivity to national circumstances and market power While all of these proposals can achieve results close to the reference case, not all of them ensure this result for different sellers. As shown in Table 2, the operational specification of the Swiss proposal and of the compliance reserve that achieves the desired environmental and economic results is sensitive to national circumstances. This means that a different value for the parameters is needed for each Annex B seller to achieve the desired results. Table 2 also shows that the performance of the Swiss proposal and compliance reserve are sensitive to market power on the part of the seller. If the seller is able to exercise market power, it has an incentive to withhold AAUs from the market to increase prices and its net revenue. Even if the seller is not able to exercise market power, it has an incentive to argue for a different specification than the one which just achieves compliance. In the case of the escrow account as well, the minimum price the Annex B seller needs to charge varies with national circumstances. If the seller is able to sustain a minimum price it has an incentive to establish a price higher than the one that just achieves compliance. Thus, for the Swiss proposal, compliance reserve and escrow account it would be difficult to ensure that the proposal is implemented for each Annex B seller in a way that achieves good environmental and economic performance. Penalties are sensitive to market power by the Annex B seller(s). If the seller(s) has limited market power, non-compliance may still be a concern. In that case a higher penalty can address the concern. If an Annex B seller has significant market power, it is likely to restrict sales beyond the quantity needed for compliance to maximize its revenue. 35 Then, the concern becomes the impact on the compliance cost of the buyer, rather than non-compliance by the Annex B seller. Table 2 indicates that the permanent reserve options are not sensitive to national circumstances and market power. That is not strictly true. The sensitivity of the results to national circumstances falls within 35

See Ellerman and Wing (2000).

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

105

2% of the maximum possible non-compliance and within 5% of the buyer’s reference case compliance cost. The sensitivity of the results to market power on the part of the Annex B seller(s) likewise falls within these ranges. In summary, sanctions involving significant penalties that are effectively enforced are sufficient to deter deliberate non-compliance through emissions trading. If the non-compliance penalties are weak or poorly enforced, only a permanent reserve can achieve results close to the least-cost, full-compliance reference case under all circumstances. 5.3. Sensitivity tests To test the stability of the results to various assumptions used in the model several sensitivity analyses were conducted. Variations in the transaction costs for JI and the CDM can increase the compliance cost for the reference case by 8.3% or reduce it by 5.4%. Depending on the size of the CDM supply from prior to 2008 the compliance cost for the reference case can increase by 11.7% or reduce it by 18.9%. Both tests suggest that the costs associated with the project-based mechanisms affect the total costs significantly. However, they do not affect the ranking of the liability options as such. The results for proposals with a user liability component are not sensitive to risk aversion by the buyer, at least to the manner in which this was modeled. Finally, the discount rate does not have a significant impact on the relative performance of the different proposals. 5.4. Compliance cost and distribution of income among sellers The Annex B buyer needs to make domestic reductions and quota purchases totaling 5680 MtC over the commitment period to achieve compliance. For liability proposals with excess emissions and compliance costs close to those of the reference case, this is achieved through domestic reductions by the buyer (25–30%), purchases of CERs (about 40%), purchases of ERUs (1–5%), and purchases of AAUs (25–30%). All of the liability proposals that achieve compliance by the Annex B seller with compliance costs close to those of the reference case result in approximately the same distribution of net income among the sellers. The compliance cost for the Annex B buyer is about US$ 140 billion, while the Annex B seller and the non-Annex B seller each receive net income of about US$ 50 billion, leaving a net cost of about US$ 40 billion. In other words, a choice among the liability proposals that achieve the desired environmental and economic results does not involve equity considerations. 5.5. Application of the liability provision to ERUs The sale of ERUs by Annex B countries can contribute to non-compliance just as the sale of AAUs can. The sale of ERUs reduces the seller’s remaining assigned amount. If the remaining assigned amount is insufficient to cover the country’s actual emissions, the sale can be argued to have led to non-compliance. Should ERUs, then, be subject to the same liability provisions as AAUs? We have assumed that decisions on the quantity of ERUs to be awarded for each JI project are subject to an international review process similar to that for the CDM. Then the ERUs reflect real reductions and so do not contribute to non-compliance by the host country. This suggests that if the quantity of ERUs

106

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

awarded is decided through an independent international process, they should be valid once issued in the same way as CERs. On the other hand, if the decision on the quantity of ERUs awarded is left entirely to the host government, it could award ERUs beyond the emissions reduction achieved to help attract JI projects that yield employment or other benefits. Then, it can be argued that ERUs and AAUs contribute equally to non-compliance by the host government. The analysis indicates that excess emissions associated with ERUs are in the range of 1–6% under several of the liability proposals. This suggests that if the decision on the quantity of ERUs awarded is left entirely to the host government, ERUs should be subject to the same liability provision as AAUs. 6. Policy implications The analysis confirms that sanctions lead to compliance if the penalty faced by the Annex B seller is substantially higher than the market price for quotas and is effectively enforced. The model suggests that the present value of the non-compliance sanctions must be at least 2.3 times the expected price. A penalty of 1.3 AAUs from the second commitment period for each ton of excess emissions, for example, does not provide a financial incentive to comply. 36 The possibility that the sanctions for non-compliance faced by Parties under the Kyoto protocol may be lower than required, or may not be effectively enforced, is the motivation for the other liability proposals. They seek to prevent an Annex B seller from benefiting financially through emissions trading and deliberate non-compliance. Our analysis indicates that several of the liability proposals can achieve results essentially equivalent to the least-cost full-compliance equilibrium even assuming the worst behavior by the Annex B seller. In other words, these proposals can prevent abuse of emissions trading at negligible cost in terms of excess emissions or extra compliance costs. The permanent reserve proposal best meets the criteria adopted. It requires Annex B sellers to establish a reserve equivalent to an estimate of total emissions during the commitment period. Trading is limited to AAUs surplus to that reserve. A permanent reserve rule would mean that Annex B countries could not use AAUs as the allowances for domestic emissions trading programs. Annex B countries could, however, establish domestic emissions trading programs that use domestic allowances which can be exchanged for surplus AAUs when needed for international sales. In practice, the size of the reserve should be capped at some percentage, say 95%, of each Annex B country’s assigned amount. Such a cap would enable a source in a country that is a net buyer to sell surplus allowances on the international market, or transfer surplus allowances to an affiliate in a different Annex B country. 37 The quantity of AAUs needed to accommodate such transactions is likely to be small because 36

Assuming that the real price of an AAU remains constant, the present value (using a 5% discount rate) on 1 January 2008 of 1.3 AAUs on 31 December 2019 when they might be needed for compliance for the second period commitment is 0.72. If the 1.3 AAUs are needed on 31 December 2017, the present value is 0.80. If the seller had to use first commitment period AAUs to establish compliance in 2014, the present value is 0.92. In every case the present value of the (loss of or) requirement to purchase 1.3 AAUs in the future is less than the revenue (1.0) from the sale of AAU on 1 January 2008. Thus, this penalty does not create a financial incentive to comply. 37 Since AAUs cannot be used in the domestic emissions trading program, the source would have surplus domestic allowances. It would exchange the surplus allowances for AAUs and then sell or transfer the AAUs.

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

107

• The country is a net buyer, so the domestic price will be at least as high as the international price and the export demand is likely to be small. • As a net buyer the Annex B country will accumulate AAUs, ERUs and CERs which can be resold to meet the export demand. 38 • Transferring funds is a good substitute for transferring quotas. 39 • The lower the percentage at which the reserve is capped, e.g. 80 versus 95%, the less effective the liability rule and the greater the environmental risk. As long as sufficient quotas are available to accommodate the desired transactions, compliance costs are not affected by keeping the cap relatively high. Some of the more recent liability proposals are based on the permanent reserve concept. 40 They differ in terms of the way the size of the reserve is calculated, whether the reserve can be adjusted during the commitment period, and other details. Our analysis suggests that several formulae for estimating the size of the reserve can achieve the desired results, although with slight differences in the level of environmental protection and compliance cost. Some of the more recent proposals also allow sales from the reserve on a user liability basis. The same result can be achieved through commercial contracts for forward delivery, so the need for a user liability component of the rule can be questioned. 41 While not sufficient to ensure compliance on their own, eligibility requirements and annual retirement of AAUs equal to actual emissions, contribute to compliance at little or no cost. Hence these provisions could complement other liability proposals. The compulsory insurance, escrow account, user liability, shared liability, double liability and traffic light proposals are difficult to analyze given the model structure, so those results should be interpreted with care. The analysis also suggests that if the decision on the quantity of ERUs awarded for JI projects is left entirely to the host government, ERUs and AAUs are equivalent and hence ERUs should be subject to the same liability provisions as AAUs.

Acknowledgements The authors gratefully acknowledge the generous cooperation and assistance provided by Prof. Denny Ellerman of MIT. In addition, Christian Albrecht, Kyle Danish, Denny Ellerman, Robert Nordhaus, Andrea Pinna, Richard Rosenzweig, Tom Wilson, ZhongXiang Zhang and an anonymous referee provided 38

If the rules for the mechanisms prohibit resale of CERs, ERUs or AAUs, a country could deposit acquired quotas in its permanent reserve account and withdraw an equal quantity of its own AAUs (maintaining the size of the reserve) and then export the AAUs. 39 If firm A in country X wanted to transfer AAUs to an affiliated firm B in country Y but country X had no surplus AAUs available, firm A could sell its surplus allowances domestically and send the money to firm B, which could then buy the quota on the international market. If the firms are unrelated, country X’s lack of surplus AAUs simply means that A must sell its surplus allowances domestically (at a price that should be at least as high as the international price) and that B must buy from sellers in other countries. This helps keep X in compliance. 40 See CIEL, 2000, Environmental Defense (2000) and Nordhaus et al. (2000b). 41 A rule that allows AAUs from the permanent reserve account to be sold on a user (buyer) liability basis means that those AAUs could not be used by the buyer until the Annex B seller has established compliance with its emissions limitation commitment. Under a forward contract the seller agrees to sell AAUs to the buyer once the seller has established compliance with its emissions limitation commitment. The price could be fixed at the time the contract is negotiated or be established at the time the AAUs are transferred.

108

E. Haites, F. Missfeldt / Climate Policy 1 (2001) 85–108

helpful comments. The authors are solely responsible for the content. The authors also wish to acknowledge financial support from EPRI. References Babiker, M.H., Reilly, J.M., Eckaus, R.S., Wing, I.S., 2000. EPPA 3.0 Documentation. Massachusetts Institute of Technology, Cambridge, MA, mimeograph, in preparation. Baron, R., 1999. An Assessment of Liability Rules for International GHG Emissions Trading. International Energy Agency Information Paper, Paris, October. Center for Clean Air Policy (CCAP), 2000. Recommendations on the Implementation of International Emissions Trading and Related Inventory, Reporting, Review and Compliance Procedures and Mechanisms. UNFCCC Secretariat, Center for Clean Air Policy, Washington, DC, January. Center for International Environmental Law (CIEL), 1999. The Compliance Fund: A New Tool for Achieving Compliance Under the Kyoto Protocol. Washington, DC, June. Center for International Environmental Law (CIEL), 2000. Hybrid Liability Revisited: Bridging the Divide between Buyer and Seller Liability. Draft, Washington, DC, September. Climate Action Network, 2000. Submission by US Climate Action Network in Response to the Note by the Chairman of the Contact Group on Mechanisms. Washington, DC, January. Chayes, A., Handler Chayes, A., 1998. The New Sovereignty: Compliance with International Regulatory Agreements. Harvard University Press, Cambridge, MA. Ellerman, A.D., Wing, I.S., 2000. Supplementarity: An Invitation to Monopsony? Report 59. MIT Joint Program on the Science and Policy of Climate Change, Cambridge, MA, April. Environmental Defense, 2000. The Kyoto Protocol: Maintaining the Dynamic Balance. Working Paper Draft 3.2, Washington, DC, August. Haites, E., 1998. International Emissions Trading and Compliance with GHG Limitation Commitments, Working Paper W70. International Academy of the Environment, Geneva. Haites, E., Missfeldt, F., 2000. Costs and Environmental Impacts of Liability Proposals. EPRI, Palo Alto, CA, unpublished. Hargrave, T., et al., 1999. Defining Kyoto Protocol Non-compliance Procedures and Mechanisms, Center for Clean Air Policy. Leiden International Emissions Trading Papers, Washington, DC, October. Nordhaus, R.R., Danish, K.W., Rosenzweig, R.H., Fleming, B.S., 2000a. International emissions trading rules as a compliance tool: what is necessary, effective and workable? Environmental Law Reporter 30, in press. Nordhaus, R.R., Danish, K.W., Rosenzweig, R.H., Fleming, B.S., 2000b. A framework for achieving environmental integrity and the economic benefits of emissions trading under the Kyoto protocol. Environmental Law Reporter 30, in press. Reilly, J., Prinn, R., Harnisch, J., Fitzmaurice, J., Jacoby, H., Klicklighter, D., Melillo, J., Stone, P., Sokolov, A., Wang, C., 1999. Multi-gas assessment of the Kyoto Protocol. Nature 401, 549–555. Shavell, S., 1987. Economic Analysis of Accident Law. Harvard University Press, Cambridge, MA. United Nations Framework Convention on Climate Change (UNFCCC), 1997. The Kyoto Protocol. Climate Change Secretariat, Bonn. United Nations Framework Convention on Climate Change (UNFCCC), 1999. Principles, Modalities, Rules and Guidelines for the Mechanisms under Articles 6, 12 and 17 of the Kyoto Protocol. Submissions from Parties. Note by the Secretariat. Addendum, Corrigendum (FCCC/SB/1999/MISC.3/Add.2/Corr.1), Bonn, June. United Nations Framework Convention on Climate Change (UNFCCC), 2000a. Procedures and Mechanisms Relating to Compliance Under the Kyoto Protocol: Report of the Joint Working Group on Compliance on Its Work During the Twelfth Sessions of the Subsidiary Bodies. Proposals from the Co-Chairmen of the Joint Working Group on Compliance, FCCC/SB/2000/CRP.3/Rev.1, Bonn, June. United Nations Framework Convention on Climate Change (UNFCCC), 2000b. Mechanisms Pursuant to Articles 6, 12 and 17 of the Kyoto Protocol: Consolidated Text on Principles, Modalities, Rules and Guidelines. Note by the Chairmen, (FCCC/SB/2000/4), Bonn, June. Weyant, J.P., Hill, J.N., 1999. Introduction and overview, the costs of the Kyoto protocol: a multi-model evaluation. Energy J., Special Issue, May, pp. vii–xliv.