Why Carbon Pricing? Comparing design rationales for carbon taxes.

Transcription

1 Why Carbon Pricing? Comparing design rationales for carbon taxes. Master Thesis Hertie School of Governance Master of Public Policy Student: Johanna Arlinghaus Supervisor: Professor Mark Hallerberg i

2 Table of Contents Executive Summary... iv Introduction Optimal Tax Design and Experience to Date Static efficiency Dynamic efficiency Externalities and the administrative level of carbon taxes Impacts of Environmental Taxation Revenue Recycling The Advocacy Coalition Framework Comparing design rationales for carbon taxes Carbon Pricing in Australia Tax Design and Economic Efficiency Considerations Explaining policy change in Australia Carbon Pricing in Ireland Tax Design and Economic Efficiency Considerations Explaining Policy Change in Ireland Discussion Conclusion References List of Figures and Tables Figure 1: Matching marginal abatement cost and marginal external damages cost Figure 2: Imposing a tax on heterogeneous firms Figure 3: Components of dynamic efficiency of a tax Figure 4: Revenues from environmentally related taxation as percentage of GDP Table 1: Basic economic indicators Australia and Ireland, Table 2: Australian Climate Policy Domain in the early 2000s Table 3: Irish Climate Policy Domain in the early 2000s Table 4: Economic growth and government debt Ireland and Australia Table 5: Gross Value Added and Employment by Economic Activity 2011 in % Table 6: GHG profile Australia and Ireland, Breakdown by subsectors, % ii

3 List of Abbreviations ACF Advocacy Coalition Framework CPRS Carbon Pollution Reduction Scheme CO2 Carbon Dioxide CEP Clean Energy Package EC European Commission EEA European Energy Agency ECOTEC EITE Emissions-Intensive Trade-Exposed EPA Environmental Protection Agency EU European Union ETD Energy Taxation Directive ETR Environmental Tax Reform ETS Emissions Trading Scheme GHG Greenhouse Gases GDP Gross Domestic Product IBEC Irish Business and Employers Federation IPCC International Panel on Climate Change MBI Market-based Instrument MC Marginal Cost MAC Marginal Abatement Costs MEC Marginal External Cost MCCC Multi-Party Climate Change Committee NCCS National Climate Change Strategy OECD Organization for Economic Cooperation and Development PPP Polluter Pays Principle SEAI Sustainable Energy Authority Ireland VAT Value-Added Tax UNFCCC United Nations Forum on Climate Change iii

4 Executive Summary This thesis asks what effective and efficient carbon taxes are and examines why countries do not introduce such taxes. Carbon taxes place a value on carbon emissions, internalizing some portion of the costs associated with their environmental impacts. The tax rate should be set such that it equalizes marginal abatement costs and marginal external cost and coverage should be uniform across sectors. Carbon taxes have the additional advantage of raising revenue, associated with the possibility of achieving a double dividend. Due to distributional implications and concerns about impacts on competitiveness, few if any real-world taxes are designed and implemented in an optimal way. Both Ireland and Australia have not long ago introduced a carbon tax. These cases are interesting as both countries are commonly perceived as relative laggards in climate policy. This thesis thus follows the logic of the most similar systems design, attributing the distinct design features of these carbon taxes to the relevant differences found between the two countries. The research highlights that the Irish tax design does a good job in applying fairly uniform rates with regards to economic efficiency, while the Australian tax is characterized by significant assistance to industry. Differences in tax design can be attributed to the distinct impact of the recent financial crisis on the two countries, their economic structure and the different extents of lobbying. While Ireland experienced a fiscal crisis implemented farreaching measures for fiscal consolidation, Australia has largely profited from the strong growth of its neighbouring countries and the associated demand for resources. High carbon emissions associated with this development caused the Australian tax to be more prominent in the national debate than in Ireland. Additionally, in Ireland some of the most emissions-intensive industries were already covered by the EU ETS, excluding them from the carbon tax scheme. In contrast, Australia s carbon price includes all sectors of the economy, making strong lobbies for industry assistance more likely. The results reiterate the importance of international agreements and institutions that could facilitate the uniform introduction of carbon taxes to achieve efficient and effective pollution abatement. iv

5 Introduction A large body of scientific evidence proves that human activity causes global warming, with its main source being greenhouse gases: Since the pre-industrial area, global concentrations of carbon dioxide, methane and nitrous oxide have increased markedly as a result of human activity due to increased use of fossil fuels, land use change and agriculture (IPCC 2007). Evidence shows that many natural systems are being affected by climate change, particularly by temperature increase (ibid). Climate change thus involves an externality: greenhouse gas emissions damage others at no cost to the agent responsible (Stern 2006) a market failure which justifies public intervention (OECD 1972). Carbon taxes place a value on carbon emissions, internalizing some portion of the costs associated with their environmental impacts. While a group of countries including Finland, Sweden, Norway and Denmark introduced carbon taxes in the 1990s, the pace of countries introducing carbon taxes has recently significantly slowed down. Economic theory has precise ideas on how such taxes should be designed in order to be both economically efficient and provide the right behavioural incentives for emissions reductions. The tax rate should be set such that marginal abatement cost match the marginal external cost of pollution. In addition, tax coverage should be uniform across sectors in order to induce pollution abatement where it most cost-efficient. Carbon taxes make energy for production processes more expensive and therefore lead to an increase in production costs. These extra-costs may impair the international competitiveness of affected firms and industrial sectors. Further, energy prices for private households will rise, giving rise to distributional concerns. Importantly, the choice of the revenue distribution may impact the tax political suitability. Through smart recycling of tax revenues, some of these impacts can be alleviated. Additionally, sectoral groups may lobby for special tax provisions and exemptions. However, assistance payments to industry do not rest on a strong economic case; they may lead to non-uniform treatment of sectors and an inefficient design outcome of the tax. Beyond this background, this thesis asks what an effective and efficient carbon tax is and examines why countries do not introduce such taxes. The Australian and Irish carbon tax and the policy process prior to tax introduction are used as case studies to analyse this research question. Ireland and Australia have both introduced a carbon tax in recent years. Ireland and Australia share the Anglo-Saxon tradition including many economic, cultural and political characteristics. This implies that the countries are suitable for a case study approach to carry out a comparative inquiry, following the logic of the Most Similar Systems Design (Przeworski and Teune 1970). Within this logic, differences in tax design can be attributed to a limited number of factors. From a viewpoint of economic efficiency, the Irish tax does a good job in applying fairly uniform 1

6 rates while the Australian tax is characterized by manifold exemptions to industry. The different policy outcomes can be explained with the countries respective geographic location, the most recent financial crisis as well as their economic structure and the distinct influence of lobbying activities on the policy process. This thesis contributes to the literature of environmental ad carbon taxes in that it provides a thorough analysis and comparison of these two recent cases. Their common stance of being relative climate policy laggards makes the joint investigation of their carbon taxes an interesting case. Additionally, these two countries fall out of the group of countries having introduced carbon taxes in the 1990s, making it worthwhile to ask why they have introduced a carbon tax now. In addition, this thesis puts environmental taxes in the context of the financial crisis, the increased popularity of green parties and the post-kyoto period, which have not received much attention in previous literature. All of these factors could have an influence on the future deployment of environmental and carbon taxes. Within the case studies, the tax design of the respective taxes is analysed under criteria of economic efficiency on the basis of the optimal tax theory introduced in the first section. More specifically, tax rate and coverage as well as revenue recycling are examined. In a second step, the different policy outcomes and the road to policy change are explained for the two countries in the period from the (non-) ratification of the Kyoto Protocol in 1997 up to the introduction of the carbon tax, using the Advocacy Coalition Framework (ACF) as an explanation for policy change. The ACF (Sabatier and Jenkins 1988, 1993, Sabatier 1998) was developed to study complex policy processes involving multiple actors. The ACF raised considerable interest for the analysis of policy change in OECD countries, especially for environmental policies (ibid.). Therefore, the ACF is chosen to explain differences in the policy outcomes with respect to carbon tax introduction in Ireland and Australia. The outline of this thesis is as follows: The first chapter reviews the theoretical rationale for introducing environmental taxes. It elaborates on the relevant definitions and discusses the common efficiency arguments underpinning environmental taxes. The second chapter introduces the ACF, which is used to analyse the policy process in the lead-up to the introduction of carbon taxes in Ireland and Australia. This is followed by the case studies of carbon tax design in Australia and Ireland in the third chapter. Within the case studies, carbon tax design is first analysed under economic efficiency aspects, secondly followed by an explanation of policy change in the respective countries. Thirdly, the discussion provides reasons for the different policy outcomes in the two countries. These are their geographic location, the most recent financial crisis as well as their economic structure and the influence of lobbying activities on the policy process. The conclusion argues that while major drivers for governments to introduce carbon taxes are compliance with international emissions reduction targets, competitiveness concerns play a 2

7 major role. This complicates the introduction of environmentally effective and economically efficient taxes. While green parties can play a pivotal role in implementing carbon taxes, financial crises can help to implement and sustain carbon taxes. At the same time, neighbouring countries, as well as the status of being an energy importing or exporting country have substantial influence on tax design. 1. Optimal Tax Design and Experience to Date This chapter explains the background and design rationales for environmental, including carbon taxes. The theory will be used to analyse the design features of the Australian and Irish carbon taxes in the third chapter. In 1972, the OECD first defined and recognized the polluter-pays principle (PPP) as an approach whereby the polluter should bear the expenses of carrying out pollution prevention and control measures decided by public authorities to ensure that the environment is in an acceptable state. These measures should not be accompanied by subsidies that could cause significant distortions in international trade and investment (OECD 1972). This spurred new interest in the use of market-based instruments (MBIs) as a means of joining economic and environmental sustainability (Barde and Godard 2012). Quantitative limits on pollution and price-based approaches are the most popular and employed MBIs to put a price on pollution (Nordhaus 2007). In the global-warming context, quantitative limits set targets on the time path of greenhouse gas emissions (GHG) of different countries. Countries can administer the amounts assigned to them in their preferred way, allowing the purchase and sale of emissions certificates across countries. Internationally, this approach has been implemented via the Kyoto Protocol, supra-nationally in the European Union (EU) via the EU Emissions Trading Scheme (ETS) and on national level among others via the SO2 allowance-trading programme in the US (ibid.). This thesis focuses on the second approach using harmonized prices, fees or taxes as a pricing method for pollution. This approach has not been used on the international level, but considerable experience with this approach exists on the national level. As the leading institution in global assessments of MBIs, the OECD, as well as Eurostat, the European Commission s (EC) Tax Directorate and the International Energy Agency (IEA), define an environmental tax as a tax whose tax base is a physical unit (or a proxy of it) that has proven specific negative impact on the environment. (OECD 1997). The affected taxes are divided into energy taxes including carbon taxes, transport, pollution and natural resource taxes (ibid). A carbon tax is a charge to be paid on each fossil fuel, proportional to the quantity of carbon emitted when it is burned (Baranzini et al 2000). A CO2 tax is specified per tonne of CO2 emitted and can easily be translated into a carbon tax by knowing that a ton of carbon corresponds to 3.67 tons of CO2. A true emissions tax would impose a charge on each 3

8 unit of GHG released - this is difficult to implement in practice since emissions are difficult to measure directly (ibid.). Despite being based on the idea of Pigouvian taxation, few if any environmental tax implemented can be termed a Pigouvian tax under the textbook definition (Milne 2012). However, under the least-cost abatement theory and the PPP, second-best taxes can have significant environmental attributes and warrant recognition as environmental economic instruments (ibid.). 1.1 Static efficiency Static efficiency is the main quality attributed to well-designed environmental taxes in case they apply to a responsive, flexible market economy that does not suffer from significant distortions except for the environmental issue that is targeted (Barde and Godard 2012). Pigou (1920) posited that the optimal tax to address a negative environmental externality is equal to the marginal external cost (MEC) from the polluting activity to society. In a perfect market, this way of taxing results in the most efficient use of resources. According to theory, a polluter will react to the environmental tax by reducing emissions to the level where the unit rate of the tax and the marginal pollution abatement cost (MAC) the cost of removing one additional unit of pollutant are equal. Figure 1 illustrates that if a tax at a rate t* is imposed, the polluter will reduce pollution from B to P*. Beyond P* it is cheaper to pay the tax than to abate more (Barde and Godard 2012). Static efficiency gains can be realized at the level of abatement measures by industry, the impact on consumer decisions and industry structure (OECD 2001). Figure 1: Matching marginal abatement cost and marginal external damages cost. Source: Author, adapted from Barde (2012). 4

9 However, full internalisation of external costs is difficult to achieve in practice, because the exact value of the environmental damage is generally not known (EEA 2005). As a result, the Baumol-Oates Theorem or the theory of least-cost abatement is often used for determining the optimal tax rate in the absence of the exact knowledge on MEC. Relying on this theory, a government tries to determine how a specific environmental standard can be achieved at the least-cost for the private sector (Milne 2012). The starting point is thus the question as to the desired level of abatement. Setting the tax rate equal to that level will yield the desired amount of change in behaviour, taking into account the cost of prevention and control which an entity is likely to occur when adapting its behaviour. In general, this will not involve equal abatement efforts by all polluters (ibid.). A key feature of taxes is that they minimize total abatement costs by equalizing MAC across polluters. Figure 2 shows three firms with different MAC. In the case public authorities introduce a tax t, firm 1 would increase its abatement effort so as to lower emissions from E11 to E12. Its MAC increase until it is cheaper to pay the tax than to further reduce emissions. Firm 3, initially abating a lot of emissions, has an incentive to reduce abatement, so its emissions increase from E32 to E32 until MAC equals t. Given that MAC of firm 2 already equal t, its situation does not change. Figure 2: Imposing a tax on heterogeneous firms. Source: Own graphic based on Barde (2012: 39) 1.2 Dynamic efficiency Compared to static efficiency gains, dynamic efficiency gains need a longer time to be realized. They can be achieved because taxes provide continuing incentives for firms to further reduce emissions through cost-effective abatement, innovation of cleaner production techniques and better abatement technologies (OECD 2001). Thus, taxes give a double stimulus (Barde and Godard 2012). First, pollution reduction is stimulated when abatement costs decrease. In Figure 3, the marginal abatement 5

10 cost is reduced from to due to technological progress. In the case of emissions at level, the producer incurs savings depicted by area B. With tax t, emissions will be reduced to. Second, a tax is a stimulus to technological change, i.e. to develop more efficient pollution-abatement techniques. Technological progress reducing marginal abatement costs from to, results in cost savings from reduced abatement costs (surface D+B), as well as reduced tax payments due to lower final levels of pollution (surface C, when emissions are reduced from to The notion that well-designed environmental regulation can spur innovation and improve competitiveness became known as the Porter Hypothesis (Porter 1995). Unfortunately its validity is limited in a situation where the polluter is not the producer of new technologies (Barde and Godard 2012). In that case, the polluter buys the technologies from the supplier, which may not take the tax into account in production. In addition in order for dynamic efficiency gains to be realized, the future evolution of the tax rate should be predictable. If innovation involves a remodelling of the whole production chain, significant investments may be required. This increases the importance of the expected future tax rate compared to the current one. However, if the future tax rate is uncertain, necessary investment may not be triggered and dynamic efficiency gains do not occur (ibid.). Figure 3: Components of dynamic efficiency of a tax. Source: Author based on Barde and Godard (2012) 6

11 1.3 Externalities and the administrative level of carbon taxes As previously stated, climate change is a global problem, to which nearly all countries contribute and that affects nearly all countries. Thus, the question at which level environmental taxes should be imposed at local, national, supra- or international level is an important one. The literature on fiscal and environmental federalism posits that the control of externalities and the provision of public goods work best if the governmental authority responsible for addressing the problem matches as closely as possible the geographic extent of the environmental problem itself (Burtraw 2012). For example, environmental benefits from a decrease in fossil fuel consumption may enhance local air quality by reducing the emission of air pollutants (Baranzini et al. 2000). However, climate change depends on total global GHG emissions and their trajectory, not on their geographic location (Nordhaus 2007). As a result, the key environmental issue is to balance costs and benefits of global emissions reductions. Even in countries with a federal structure of government, policies to address climate change belong at the national level and the international activities required to address climate change should arise from negotiations among sovereign nations (Burtraw 2012). Thus, the theoretically superior solution would be an international carbon tax negotiated by sovereign nation states or a harmonized tax, none of which has materialized so far 1. Emissions reductions will be most efficient if the marginal cost of emissions reductions is equalized across space and with appropriate discounting across time. The spatial component of efficiency is that the marginal costs of pollution abatement are equalized across all countries, industries and sources (see Figure 2). Inter-temporal efficiency requires the profile of emissions be timed to attain the ultimate goal. This goal may be to stabilize temperature or a cost-benefit criterion. In this case, dynamic models require that the market price of carbon should rise over time as the tax base will in the ideal situation erode due to abatement. An internationally harmonized carbon tax would be spatially efficient among those countries that harmonize taxes. If the carbon tax increases over time, it would also be efficient in an inter-temporal way (ibid.). The EU s Energy Taxation Directive (ETD) adopted in 2003 is an attempt to harmonize member state s energy taxes supra-nationally to avoid competitive distortions in the energy sector within the Internal Market (EC 2011). It sets out common rules for what should be taxed and outlines which exemptions exist. Minimum rates, mainly based on the volume of energy consumed are laid down for products used in heating, electricity and motor fuels those sectors not covered by the EU ETS. Member states are allowed to set higher national tax rates as they wish. The directive has been criticised for creating unfair competition between fuel sources and benefits for certain types of fuel compared to others by defining 1 For an overview on the discussion on global environmental taxes, see Thalmann (2012). 7

12 the tax rate on the basis of volume. For example, renewable energies face discrimination because they are taxed at the same rates as energy from fossil fuels they are meant to replace (ibid.). A proposal to revise the 2003 ETD has been published in In the proposed Directive, energy taxes would be linked to energy and CO2 content by introducing a single minimum rate for CO2 emissions (20EUR/t CO2) and minimum tax rates for energy content (i.e. gigajoule) of a fuel rather than volume. Both CO2 and energy content elements would be combined to produce the overall rate at which a product is taxed. The energy element would apply to all fuel used for transport and heating, while the CO2 element of the tax will complement the emission trading system, covering sectors that remain outside transport, households, agriculture and small industries (EC 2011). However, a decision in favour of the revised ETD is currently put on hold - a period of austerity and high fuel costs is not seen as the right time to raise minimum tax rates on energy (euractiv 2012). In addition, any decision made requires unanimity in the Council of Ministers, where Poland has already blocked moves for higher emissions reduction targets (ibid.). Interest in energy and carbon taxes seems to have lost momentum at the EU level since adoption of the EU ETS (Helm 2009, Speck 2009). Consequently, the decision for increasing carbon and energy taxes is again left to the respective national levels, which is the approach taken by Australia and Ireland. 1.4 Impacts of Environmental Taxation When deciding whether to implement environmental including carbon taxes it is important to investigate impacts. These are their environmental effects, distributional implications and consequences for national competitiveness. Overall, the literature agrees that while environmental and carbon taxes can achieve emissions reductions, it is possible to mitigate distributional and competitiveness effects of environmental taxes by making smart use of the revenues. While ex ante modelling studies generally play a large role in decision-making, ex post assessments are subject to methodological difficulties (Bosquet 2000). This is because environmental taxes are generally introduced in conjunction with other tax and compensation measures and due to the absence of a counterfactual scenario (OECD 1997, Baranzini et al. 2000). Despite this difficulty, evidence from northern EU states found that environmental taxes have significant positive impacts on the environment (ECOTEC 2001, EEA 2000, Baranzini et al. 2000, Fujiwara et al. 2006). An important concern in the discussion is that environmental taxes are likely to have a regressive impact on income distribution (Kosonen 2012: 161, EEA 2011). Since low-income households generally tend to spend larger parts of their income on energy, the tax will have a larger stronger impact on the 8

13 expenditure of low-income household. In general, this depends on the price elasticity of demand relative to the price elasticity of supply: the lower the demand elasticity and the higher the supply elasticity, the more the tax will be reflected in consumer prices and thus the higher its potential regressive impact (Kosonen 2012). Specifically for energy taxes, the short-term demand elasticity is low, as switching costs and the behavioural changes as a response to the tax usually take time. In addition, alternatives may not be readily available and consumer s may be locked in to their current energy consumption patterns, e.g. through previous investment into household infrastructure. However, in the middle- to long run, consumers will be more responsive to price changes (Fujiwara et al 2006). A recent study by the EEA (2011) confirmed mild regressive impacts of carbon taxes, but a number of studies conclude that those impacts can be alleviated by smart recycling of tax revenues (EEA 2011, Baranzini et al 2000, Ekins and Speck 1999). Competitiveness effects due to increases in production costs are a major political obstacle for the introduction of environmental taxes and usually induce sectoral groups to lobby against the tax or for exemptions and tax rate reductions in many countries (OECD 2001, Fujiwara et al 2006: 15, Jotzo 2011, Pearce 2006, Pezzey 2010). However, the idea that environmental taxes undermine competitiveness is often based on a static view of competitiveness in which technology and innovation are exogenous (Fujiwara et al. 2006). Placing environmental taxes in a dynamic framework (Figure 3) shows that environmental and carbon taxes can actually enhance competitiveness through increasing innovation. Empirically there is little or no evidence that environmental taxes have negative impacts on competitiveness (Fujiwara et al. 2006, Ekins and Speck 1999, ECOTEC 2011). Newer studies confirm that environmental taxes provide incentives for eco-innovation in the short-run, which can boost efficiency and competitiveness and ensure sustained prosperity in the longer term (OECD 2010, EEA 2011). 1.5 Revenue Recycling A positive effect of environmental taxes from a policy perspective is their revenue-raising potential and the idea that environmental taxes could be used to finance reductions in other taxes ( double dividend hypothesis ) (Baranzini et al 2000, Jaeger 2012). In this context, an Environmental Tax Reform (ETR) is a reform of the national tax system shifting the burden of taxes away from taxes capital and labour and social security contributions, to sources of environmental externalities, in particular energy or environmental pollution (EEA 2005). ETR is generally associated with revenue-neutrality, meaning that the overall tax burden is left unchanged (ibid). 9

14 The theory of optimal taxation recognizes that taxes distort behaviour, introducing inefficiencies in resource allocation compared to the undistorted optimum (Jaeger 2012). Environmental taxes however are meant to correct behaviour in a way that eliminates the inefficiencies arising from above-optimal consumption. Integrating these two ideas, the double-dividend hypothesis posits that environmental taxes reducing pollution and raising revenue serves both environmental and public expenditure goals. Thus, recycling carbon tax revenues by reducing distortionary taxes may have positive impacts on economic growth, employment and technological development (Hourcade 1996, Baranzini et al. 2000, Jaeger 2012). Given this revenue-recycling benefit, the environmental aspects of environmental taxes might even turn out to be less important (Bovenberg 1999). In contrast, the Mirrlees Review by the British Institute for Fiscal Studies (IFS) questions the double dividend hypothesis 2. The authors argue that environmental taxes raise the production costs, increasing output prices and reducing the net return from each hour worked just like income taxes and social security contributions do (Fullerton, Leicester and Smith 2010). By implication, environmental taxes should primarily be justified on environmental grounds (ibid.). In practice, there are manifold ways to deal with the revenues collected from environmental taxation. Revenues may be absorbed into the general government budget, redistributed to the population or the economy or used to decrease other taxes to achieve revenue neutrality. Revenues may be earmarked in that they are in advance allocated to finance specific environmental programmes or used to directly compensate those members of the population which are most affected by the tax (Baranzini et al. 2000: 399). Each of these has other fiscal, distributional and efficiency implications. The recent global and financial crisis and the associated problems with fiscal deficits led to a focus on the fiscal and revenue implications of environmental taxes. In OECD countries, environmental taxes contribute around 2% to GDP in 2010 with a growing tendency (OECD 2010). In 2010, 70% of the revenues from environmental taxes were raised on various energy products, particularly on motor vehicle fuels. On top of that, 27% of the total revenues were raised via taxes on motor vehicles and various transport activities. Taxes on all other environmentally related tax-bases, for which the price elasticity is larger than for energy and vehicles, only raise about three per cent of the total revenue (ibid.). Figure 4 shows that Denmark with 4.2% and the Netherlands with 4.5% of GDP are those OECD countries generating most revenue from environmental taxes. Australia and Ireland are located in the middle- to lower range, with Australia generating 1.8% and Ireland generating 2.4% of GDP as revenue from those taxes in There is an ongoing debate on the relevance and magnitude of these additional benefits. For a full account, see Jaeger (2012). 10

15 Figure 4: Revenues from environmentally related taxation as percentage of GDP. Source: OECD/EEA database on instruments for environmental policy Exploring the potential of carbon pricing as a way to raise tax revenues and contribute to fiscal rebalancing in Europe, the CETRiE Project examines reforms of the energy taxation systems of Hungary, Poland and Spain (Vivid Economics 2012). It is demonstrated that direct taxes could have twice as much negative impact on GDP than energy taxes while raising the same amount of revenue. Further, energy taxes reduce energy imports and increase energy security (ibid.). In contrast, while the Mirrlees Review sees a greater role for environmental taxation in the UK, it underlines that environmental taxes cannot transform the composition of the tax system. In sum, environmental and carbon taxes currently raise revenue of up to 4.5% of GDP in OECD countries, with the largest part of it being fuel and energy taxes. Revenue recycling and ETR can make use of the revenues, such that the overall efficiency of the tax system may be enhanced. While some studies see a larger role for environmental and carbon taxes in the future, the extent of this double dividend is questioned by others. 2. The Advocacy Coalition Framework Having surveyed the literature on environmental and carbon taxation, this chapter introduces the Advocacy Coalition Framework, which will be used for explaining policy change in Australia and Ireland in the lead-up to carbon tax introduction. The ACF rests on five premises (Sabatier 1998). First, theories of the policy process or policy change should address the role of technical information concerning the magnitude and facets of the problem, its 11

16 causes and the probable impacts of various solutions. Second, a time perspective of a decade or more is needed to understand policy change. Third, the most useful unit of analysis for understanding policy change in modern industrial societies is a policy subsystem or domain, rather than a governmental organization or programme. Subsystems are made up of a variety of actors from public and private organizations who are actively concerned with a particular issue, such as air pollution or energy policy and who regularly seek to influence public policy in that domain. Advocacy coalitions consist of actors in a subsystem who share basic policy beliefs and who collaborate over time. Fourth, in addition to administrative agencies, legislative committees and interest groups, journalists, researchers and policy analysts as well as actors at all levels of government should be taken into account in policy analysis. Fifth, public policies incorporate implicit theories about how to achieve objectives and can thus be conceptualized much the same way as belief systems. They involve value priorities, perceptions of important causal relationships, perceptions of world states including the magnitude of the problem and perceptions or assumptions concerning the efficacy of various policy instruments (Sabatier 1998). At the level of individual subsystems, advocacy coalitions seek to influence the policy process using various strategies and instruments such as lobbying elected officials, commissioning research or influencing public opinion. It is assumed that belief systems are hierarchical, meaning that core beliefs are more resistant to change than secondary ones. Also government programmes can be seen as being made up of policy cores and secondary aspects. Policy change is a function of processes within the subsystems in question, influenced by relatively stable parameters and external system events. The result of policy change is one or more changed or new governmental programmes, which produce outputs and impacts at the operational level. Within subsystems, policy-oriented learning occurs. Although policy-oriented learning can contribute to policy change, major shifts in the distribution of political resources, leading to modification of the core aspects of governmental policy or programme are usually the results of perturbations external to the subsystem. The ACF identifies four categories of external system events which can promote policy change: major socio-economic changes such as dislocations or the rise of social movements, changes in public opinion particularly regarding the relative seriousness of various problems, or changes in systemic governing coalitions as well as policy decisions and impacts from other subsystems. From this setting, the ACF develops the following hypotheses, which will be applied to the cases at hand: 12

17 Coalition Hypothesis 1: On major controversies within a mature policy subsystem when policy beliefs are in dispute, the line-up of allies and opponents tends to be rather stable over periods of a decade or so. Coalition Hypothesis 2: Actors within an advocacy coalition will show substantial consensus on issues pertaining to the policy core, although less so on secondary aspects. Coalition Hypothesis 3: An actor (or coalition) will give up secondary aspects of his/her (its) belief system before acknowledging weaknesses in policy core. Policy Change Hypothesis 1: The policy core of a governmental programme in a specific jurisdiction will not be significantly revised as long as the subsystem advocacy coalition that initiated the programme remains in power within that jurisdiction except when the change is imposed by a hierarchically superior jurisdiction. Policy Change Hypothesis 2: The policy core of a governmental action programme are unlikely to be changed in the absence of significant perturbations external to the subsystem, that is, changes in socio-economic conditions, system-wide governing coalitions, or policy outputs from other systems. Learning Hypothesis 1: Policy-oriented learning across belief systems is most likely when there is an intermediate level of informed conflict between the two. In such a situation, it is likely that: a) each coalition has the technical resources to engage in such a debate; b) the conflict is between secondary aspects of one belief system and core elements of the other or, alternatively, between important secondary aspects of the two belief systems. Learning Hypothesis 2: Problems for which accepted quantitative data and theory exist are more conducive to policy-oriented learning than those in which data and theory are generally qualitative, quite subjective, or altogether lacking. Learning Hypothesis 3: Problems involving natural systems are more conducive to policy-oriented learning than those involving purely social or political systems because in the former many of the critical variables are not themselves active strategists and controlled experimentation is more feasible. Learning Hypothesis 4: Policy-oriented learning across belief systems is most likely when there is a forum which a) is prestigious enough to force professionals from different coalitions to participate; and b) is dominated by professional norms. 3. Comparing design rationales for carbon taxes This chapter introduces the case studies, which will first be analysed under from an economic efficiency perspective making use of the theory introduced in the first section. In discussing the efficiency features of the respective carbon taxes, it is important to note that almost no tax implemented in practice conforms to theoretical standards. Second, the road to policy change will be analysed using the ACF introduced in the previous section. The case studies chosen follow the logic of the most similar systems design. This theory posits that if some important differences in tax design are found among these otherwise fairly similar countries, then the number of factors attributable to these differences will be sufficiently small to warrant explanation of those differences alone (Przeworski and Teune 1970). Thus, the distinct design features of the Irish and Australian carbon taxes will be explained with important differences found in surrounding conditions. The discussion as a third part of this section highlights, compares and analyses the distinct tax design features between of the two countries and the conditions which have led to these differences. 13

18 The similarities between Australia and Ireland are large. Besides the English language, both countries have a two-chamber, parliamentary democracy with proportional representation on national level. In addition, Australia and Ireland share the common law tradition based on the English model and the dominant religion is Christianity. Table 1 displays some of the main economic indicators from the year 2005, which was selected as data were gathered before the economic crisis. Both countries are characterized by approximately the same level of income, a low level of public expenditure, similar levels of taxation and a negative current account. Furthermore, the two countries take on a similar role with respect to environmental policies. As discussed below, both Australia and Ireland can historically be characterized as relative laggards in the international climate policy arena. Another similarity with respect to this move is the pivotal role of the Green Party in the policy process. Production and income Australia Ireland Australia Ireland Expenditure GDP per capita (USD current PPPs) Public expenditure on health (% GDP) Gross national income (GNI) per Private expenditure on capita (USD current PPPs) health (% GDP) Household disposable income (annual growth %) 4,5 7,8 Public social expenditure (% GDP) Economic growth Public pension expenditure (% GDP) Real GDP growth (annual growth %) 3,0 5,3 Taxes Net saving rate in household disposable income (%) 0,4 1,7 Total tax revenue (% of GDP) Gross fixed capital formation 9,3 14,7 Taxes on income and profits (% of GDP) Economic structure Taxes on goods and services Real value added: agriculture, 2,8-17,5 Taxes on the average forestry, fishing worker (annual growth %) Real value added: industry 0,6 2,2 Trade (annual growth %) Real value added: services (annual growth %) Government deficits and debt 4,3 8,3 Current account balance of payments (% GDP) Environment Government deficit (% GDP) 1,6 1,7 CO2 emissions from fuel combustion (Mln tonnes) General government debt 28,5 32,4 Population 5,6 5,8 2,8 1,8 16,5 16,0 3,3 3,4 29,8 30,3 17,6 11,7 8,3 11,4 28,5 23,5-5,7-3, General government revenues (% 35,5 35,4 Total population ( GDP) persons) General government expenditures (% GDP) 33,9 33,8 Population growth rates (%) 1,3 2,2 Table 1: Basic economic indicators Australia and Ireland, Source: OECD, Country Statistical Profiles

19 3.1. Carbon Pricing in Australia Tax Design and Economic Efficiency Considerations Tax Rate and Coverage On 1 July 2012 Australia introduced a carbon price of A$23 per tonne of CO2 (= at 2013 prices) for three years. This price rises at 2.5% each year in real terms to A$25.4/t (= 19.90) (Australian Government 2012). As an unlimited amount of permits is sold at this price and neither international trading nor banking is allowed, the scheme acts like a carbon tax during the first years (Jotzo 2012, Spash 2012). On 1 July 2015, the carbon price will transition to a fully flexible price with the price being determined by the market (Australian Government 2012). This scheme thus functions as an emissions trading system (Jotzo 2012, Spash 2012). The price will be kept within between a price floor of A$15 and a ceiling of A$20 for another three years until it is allowed to float freely. The minimum price is meant to support investment security and ensures that the price is high enough to achieve a behavioural effect, while the price ceiling ensures that the price will not rise too high (Jotzo 2012). A process to internationally link Australia s future ETS with other systems such as the EU ETS is foreseen (ibid.) and negotiations with the EU and China have already started (Australian Government 2013a). While the price rises as requested by economic theory, the carbon price alone is most likely not high enough to allow Australia to reach its emissions targets 3 (Garnaut 2011b). In addition, the price rise of 2.5% per is lower than the recommendation of 4% (ibid.). Due to the recent introduction, it is still too early to assess dynamic efficiency gains resulting from technology improvements. Previous sections have argued that an internationally uniform carbon tax would be the most efficient in economic terms due to an international equalisation of marginal abatement cost. In the face of the recent price collapse of the carbon price under the EU ETS with prices varying between near zero in 2007, 32 in 2006, crashing again at the beginning of 2013, Australia s price on emissions is much higher than the EU price (Helm 2009, Carrington 2013, Jotzo 2012). Should Australia want to pursue effective domestic abatement with high sufficiently high carbon prices, keeping a national minimum price may make sense in practice, even if this is not the efficient approach terms of economic theory. The price covers the stationary energy sector, rail transport, non-transport use of fuels, industrial processes, non-legacy waste and fugitive emissions. Also landfill facilities with direct emissions of more than tonnes CO2 emissions each year will be liable. The carbon price excludes household transport fuels, light vehicle business transport and off-road fuel use by agriculture, forestry and fishing 3 National GHG reductions targets are that Australia unconditionally reduces its GHG emissions by 5% compared to 2000 levels and by up to 15% by 2020 conditional on a global agreement. It commits to 25% reduction if by 2020 the world agreed on a deal under which global temperature increases could be limited to 2 C ( 450ppm CO2 equivalent or lower). 15

20 industries (ibid.). Overall, the carbon price covers around 60% of Australia s GHG emissions, including those from electricity generation, industry and households by upstream liability, as well as from mines (Jotzo 2012). Via separate regulation, an equivalent emissions price will be imposed on some uses of transport fuels through changes to fuels taxes and on synthetic greenhouse gases (ibid.). Coverage is relatively broad as compared to e.g. the EU ETS which covers around 50% of the EU s GHG and the scope of European carbon taxes, which do not cover more than 50% of a country s emissions due to their participation in the EU ETS and the avoidance of double taxation. Revenue recycling, equity and competitiveness The revenue from the carbon tax is estimated to exceed A$4billion and rise to A$8 to 12 billion, depending on the carbon price trajectory and coverage (Hatfield-Dodds 2011). Assistance for households is provided through income-tax cuts and increased welfare payments (Australian Government 2012). This results in A$5 billion of transfers/year, which means that the majority of lower income households will be overcompensated for the increased living costs, while higher-income households bear most of the net costs (Jotzo 2012). The case for household assistance as a means to alleviate the tax s regressive impacts has been made in previous sections. Using tax revenue to cut income and other direct taxes rests on a strong economic efficiency case due to the possibility of a double dividend. Industry assistance is provided through the $A8.6 billion Jobs and Competitiveness Programme for emissions-intensive and trade-exposed industries (EITE) such as steel making and aluminium smelting 4, as well as for local communities and workers (Australian Government 2013). An Energy Security Fund will allocate free carbon units and cash payments to strongly affected coal-fired electricity generators. The $A1.2 billion Clean Technology Programme in manufacturing industries is supposed to improve energy efficiency, the A$300 million Steel Transformation Plan supports the steel industry in the transition phase and the A$1.3 billion Coal Sector Jobs Package provide additional support to help the coal industry, especially targeted at those mine which pollute most. More than A$13 billion will be invested in clean energy projects. While emissions from farming are not covered by the carbon price, the Carbon Farming Initiative additionally supports farmers via an offset scheme, generating carbon credits in agriculture which can be sold to other businesses wanting to offset their own carbon pollution (ibid.). This results in the EITE industries receiving free permits with a value of more than A$3billion per year, as well as cash and free permits for the most emissions-intensive coal fired power stations of A$5.5 billion over five years and coal mines of A$1.3 billion over 6 years (Jotzo 2012). Through this 4 Other EITE industries are production of bulk flat glass, methanol, carbon black, silicon, smelting zink, manufacture of newsprint, production of magnesia, packaging and industrial paper manufacturing, aluminia refining, integrated iron and steel manufacturing and petroleum refining 16

21 programme, the most emissions-intensive activities receive assistance to cover 94.5% of industry average carbon costs in the first year of the carbon price (Australian Government 2012). Less emissions-intensive activities receive assistance to cover 66% of industry average carbon costs in the first year (ibid.). The official justification for the Jobs and Competitiveness Programme is that assistance is provided to entities that face high carbon costs and are constrained in their capacity to pass through costs in global markets, issuing free carbon permits to eligible applicants (Australian Government 2011a). Essentially, these companies are being compensated for losses in competitiveness while giving them incentives for improving efficiency (Jotzo 2012). Thus, resistance against the introduction of any carbon tax seems to have been overcome with assistance payments to industries. These payments do not rest on a strong economic case (Jotzo 2012). Excessive industry assistance gives rise to several inefficiencies; it prevents the marginal costs of abatement from being equalized across polluters, resulting in too little abatement from EITE industries and too much abatement in less carbon-intensive industries. Further, excessive assistance to industry eats up revenue for investment in low-carbon technology and energy savings for households, as well as compensation for low-income households and tax cuts (ibid.). In contrast, industry support may be valid on grounds of carbon leakage to non-carbon constrained countries 5 (Pezzey et al 2010). While the EITE sectors have the highest potential for carbon leakage, data limitations prior to tax introduction of the real extent of carbon leakage open up rhetoric space for the arguments of industry lobbyists. In addition, evidence for carbon leakage in Australia is rare (Pezzey 2010), suggesting that industry is largely over-protected through this assistance. As this outcome cannot be sufficiently justified on economic grounds, large assistance to industry is most likely the result of the influence of sectoral lobbies by EITE industries (Jotzo 2012). Consequently, the Australian carbon pricing scheme was heavily criticised, as it provides distortionary subsidies via free allocation of emission permits on the basis of grandfathering to the EITE companies (Spash 2012: 73). This contradicts standard theory of emission taxes and violates the PPP. Further, free permits were allocated on the basis of what polluters themselves requested to be needing, risking over-allocation and price collapse of the trading system (Spash 2012). Further, in the fixed-price period, holders of freely allocated permits will be allowed to sell unused permits to the government at fixed value and are even allowed to trade whereas purchased permits cannot be traded. As only less energyintensive industries have to buy their permits, this regulation offers preferential treatment for large polluters relative to smaller ones. In addition, free permit allocation gives firms incentives to pass on the 5 Carbon leakage is considered to be the movement of EITE industries from the country with higher carbon price to other countries that imposed less carbon constraint. This could result in an increase in global emissions in the case that the activity moves to a countries that uses a more emissions-intensive production process than the country of origin. 17