Carbon Pricing and Transfers
Feasible and Effective Multilevel Climate Policy in Federations
vorgelegt von Christina Roolfs Master of Sustainability Economics and Management ORCID: 0000-0001-7731-193X
an der Fakultät VI - Planen Bauen Umwelt der Technischen Universität Berlin zur Erlangung des akademischen Grades
Doktorin der Wirtschaftswissenschaften - Dr. rer. oec. -
genehmigte Dissertation
Promotionsausschuss: Vorsitzender: Prof. Dr. Stefan Heiland Gutachter: Prof. Dr. Ottmar Edenhofer Gutachter: Prof. Dr. Marco Runkel Gutachter: Prof. Dr. Matthias Kalkuhl
Tag der wissenschaftlichen Aussprache: 16. November 2020
Berlin 2021
Abstract
Regional and national governments face the challenge of enacting policies that rapidly decarbonize their economies so as to deliver on their climate targets. They need to fnd policy instruments that are both efective and feasible. Simultaneously reducing greenhouse gas emissions at multiple levels is a promising step forward. By analyzing the efectiveness and feasibility of multilevel climate policies in federal systems, this thesis translates Elinor Ostrom’s polycentric governance approach to formal public economics. General equilibrium models of low to medium complexity are developed and solved analytically and numerically for this purpose. The European Union (EU) serves as the primary example. In a nutshell, the following three results are generated: First, socially optimal transfers imply rich states being donors and poor states being recipients of carbon price revenues. From the perspective of rich states, these transfers may be too large and thus become politically infeasible. Second, choosing a uniform federal carbon price that maximizes the utility of the richest state leads to a minimum federal carbon price to which all member states would agree. The existence of such a minimum federal price depends on the degree of wealth heterogeneity among member states; the federal transfer rules; and whether or not states anticipate federal transfers. While transfers based on states’ historical emission levels (sovereignty) are always feasible and efective, excessive diferences in wealth between states limit the feasibility of equal per capita (egalitarian) transfers. Transfers based on states’ actual emission payments (juste retour) render federal policy inefective if states are able to anticipate them. Third, the potential for progressive uniform federal carbon prices depends on the relationship of member states’ diferences in wealth (vertical inequality) and CO2 emission intensity (horizontal inequality). If CO2 emission intensity and wealth are inversely correlated, then a uniform federal carbon price has a regressive efect. This is the case for most countries in the EU. The EU, however, transfers a major part of its ETS revenues based on the sovereignty rule. Such transfers can counteract the initial regressive efect leading to a progressive federal carbon price. Overall, this thesis provides a systematic analysis of price-based multilevel climate policy and revenue transfers in federal systems without solving the general distributional conficts in a federation. It identifes federal carbon prices that create a win-win situation for all member states and demonstrates that at appropriate minimum prices, the largest donor state voluntarily gives transfers to other member states. The results may be of interest to policymakers seeking feasible federal minimum carbon prices, appropriate transfer designs and fair burden-sharing in federal systems.
Zusammenfassung
Regionale und nationale Regierungen stehen vor der Herausforderung, politische Maßnahmen zur raschen Dekarbonisierung ihrer Volkswirtschaften zu ergreifen, um ihre Klimaziele zu erreichen. Sie müssen politische Instrumente fnden, die sowohl wirksam als auch durchführbar sind. Die gleichzeitige Reduzierung der Treibhausgasemissionen auf mehreren Ebenen ist ein vielversprechender Schritt nach vorn. Durch die Analyse der Wirksamkeit und Durchführbarkeit von Klimapolitiken auf mehreren Ebenen in föderalen Systemen übersetzt diese Arbeit Elinor Ostroms polyzentrischen Governance- Ansatz in die formale, öfentliche Finanzwirtschaft. Zu diesem Zweck werden allgemeine Gleichgewichtsmodelle von geringer bis mittlerer Komplexität entwickelt und analytisch und numerisch gelöst. Als primäres Beispiel dient die Europäische Union (EU). Zusammengefasst werden die folgenden drei Ergebnisse generiert: Erstens kann eine Preisuntergrenze im Emissionshandelssystem (ETS) dessen Mängel beheben. Sozial optimale Transfers implizieren, dass reiche Länder zu Gebern und arme Länder zu
Empfängern von CO2-Preiseinnahmen werden. Aus der Perspektive reicher Länder können diese Transfers zu groß und damit politisch undurchführbar sein.
Zweitens führt die Wahl eines einheitlichen föderalen CO2-Preises, der den Nutzen des reichsten Landes maximiert, zu einem föderalen CO2-Mindestpreis, dem alle Mitgliedsländer zustimmen würden. Die Existenz eines solchen föderalen Mindestpreises hängt ab vom Grad der Wohlstandsheterogenität zwischen den Ländern, von den föderalen Transferregeln und ob die Länder föderale Transfers antizipieren oder nicht. Während Transfers auf der Grundlage des historischen Emissionsniveaus der Länder (Souveränität) immer durchführbar und wirksam sind, schränken übermäßige Wohlstandsunterschiede zwischen den Ländern die Durchführbarkeit gleicher (egalitärer) Pro-Kopf-Transfers ein. Transfers, die auf den tatsächlichen Emissionszahlungen der Länder beruhen (juste retour), machen die föderale Politik unwirksam, wenn die Länder in der Lage sind, diese zu antizipieren.
Drittens hängt das Potenzial für progressive einheitliche CO2-Preises auf föderaler Ebene vom Verhältnis zwischen den Wohlstandsunterschieden der Ländern (vertikale
Ungleichheit) und der CO2-Emissionsintensität (horizontale Ungleichheit) ab. Wenn reiche Länder eine niedrigere CO2-Emissionsintensität haben als arme Länder, dann hat ein föderaler CO2-Preis einen regressiven Efekt. Dies ist für die meisten Länder in der EU der Fall. Die EU transferiert den größten Teil der Einnahmen aus ihrem ETS auf der Grundlage der Souveränitätsregel. Solche Transfers können dem anfänglich regressiven
Efekt entgegenwirken und zu einem progressiven föderalen CO2-Preis führen. Insgesamt bietet diese Dissertation eine systematische Analyse der preisbasierten Mehrebenen-Klimapolitik und Transfers in föderalen Systemen ohne die allgemeinen
Verteilungskonfikte in einer Föderation zu lösen. Sie identifziert föderale CO2-Preise, die eine Win-Win-Situation für alle Mitgliedsländer schafen und zeigt, dass bestimmte
CO2-Mindestpreise gewährleisten, dass das größte Geberland freiwillig Transfers an andere Mitgliedsländer zahlt. Die Ergebnisse können für politische Entscheidungsträger von Interesse sein, die nach machbaren föderalen CO2-Mindestpreisen, geeigneten Transferdesigns und fairen Lastenverteilungen in föderalen Systemen suchen.
vi Table of Contents
Abstract iii
Title Page ii
Zusammenfassungv
List of Figures xi
List of Tables xiii
1 Introduction1 1.1 Economics of climate change and public economics ...... 5 1.1.1 Multilateral climate policy: Limits and solutions ...... 5 1.1.2 Voluntary provision of public goods ...... 6 1.1.3 Distributional concerns and fair burden-sharing ...... 8 1.1.4 Fiscal federalism and its application to multilevel climate policy9 1.2 The European Union as a laboratory for multilevel climate policy . . . 11 1.2.1 Multilateral climate policy ...... 11 1.2.2 Voluntary participation of member states in contributing to EU climate targets ...... 12 1.2.3 Burden-sharing and transfers ...... 13 1.2.4 The federal European Union ...... 14 1.3 Objectives and outline ...... 18 References ...... 23
2 Agreeing on an EU ETS Price Floor 29 Key Points for Policymakers ...... 30 2.1 Introduction ...... 30 2.2 The Main Shortcomings of the Current EU ETS Price Signal ...... 33 2.3 Guiding Principles for EU ETS Design with Heterogeneous Member States 36 2.3.1 Efciency Transfers and Solidarity ...... 36 2.3.2 National Preferences and Subsidiarity ...... 39 2.3.3 Institutional Design in a Non-Optimal World ...... 41
vii TABLE OF CONTENTS
2.4 Illustration of the Efects of an EU ETS Price Floor ...... 44 2.4.1 Cushioning Intra-ETS Leakage with a Carbon Price Floor . . . 44 2.4.2 Implications for the European Power Sector: Numerical Simulation 46 2.4.3 Some Implementation Issues ...... 49 2.5 Concluding Remarks ...... 52 References ...... 53
3 Make or Brake 61 3.1 Introduction ...... 63 3.2 Literature review ...... 66 3.3 The model ...... 69 3.3.1 Private sector agents ...... 71 3.3.1.1 Firms ...... 71 3.3.1.2 Consumers ...... 71 3.3.1.3 Market clearing and reaction of private sector . . . . . 72 3.3.2 Multilevel emission policy ...... 72 3.3.3 State policy ...... 73 3.3.4 Decentralized policy equilibrium ...... 76 3.3.5 Federal policy ...... 76 3.4 Impact of transfer rules ...... 78 3.4.1 Analytical results ...... 78 3.4.1.1 Main fndings ...... 79 3.4.1.2 Capital-homogeneity-restriction of egalitarian transfers 83 3.4.1.3 Transfer anticipation and state policy in multilevel equilibrium ...... 84 3.4.1.4 Aggregate emission reduction at the federal minimum price ...... 86 3.4.2 Numerical analysis ...... 86 3.4.2.1 Assumptions and specifc functional forms ...... 87 3.4.2.2 Results ...... 88 3.5 Conclusion ...... 94 References ...... 97 Appendix 3.A First-order conditions of states ...... 102 Appendix 3.B Stackelberg-Leader’s frst order conditions ...... 103 Appendix 3.C Cobb-Douglas technology ...... 103 3.C.1 Firm’s problem ...... 103 3.C.2 Market clearing and reaction functions of frms and consumers 103 Appendix 3.D Proof juste retour anticipated* ...... 104 Appendix 3.E Proofs Proposition2...... 105 3.E.1 Egalitarian unanticipated ...... 105
viii TABLE OF CONTENTS
3.E.2 Egalitarian anticipated* ...... 108 3.E.3 Sovereignty unanticipated ...... 109 3.E.4 Juste retour unanticipated ...... 110 Appendix 3.F Sovereignty anticipated* ...... 111 Appendix 3.G Proof of Proposition4...... 112 Appendix 3.H Capital-homogeneity-share consideration ...... 112 Appendix 3.I General CES-function ...... 112 3.I.1 Firm’s problem ...... 112 3.I.2 Market clearing and reaction functions of frms and consumers . 113
Appendix 3.J Estimation of αE, γ and σ ...... 113 Appendix 3.K Sensitivity analysis ...... 114 3.K.1 Linear vs log utility and capital stock diferences ...... 114 3.K.2 Robustness test for sigma-variation ...... 115 3.K.3 Consumption changes at minimum prices ...... 115 3.K.4 Aggregate emission levels at minimum prices ...... 117
4 Technology Beats Capital 119 4.1 Introduction ...... 121 4.2 Literature ...... 123 4.3 The model ...... 125 4.3.1 Economic agents ...... 126 4.3.1.1 Consumers and revenue recycling ...... 127 4.3.1.2 Firms and emission efciency ...... 127 4.3.1.3 Market clearing ...... 128 4.3.2 Technology and capital impact ...... 129 4.3.3 Multilevel emission tax choices ...... 132 4.3.3.1 State governments ...... 132 4.3.3.2 Federal authority ...... 133 4.4 Numerical application ...... 135 4.4.1 Data and calibration ...... 135 4.4.2 Results immobile capital ...... 137 4.4.3 Results contrasting mobile to immobile capital ...... 141 4.5 Conclusion ...... 145 References ...... 149 Appendix 4.A Conditional input demand ...... 153 Appendix 4.B Market clearing with immobile capital ...... 153 Appendix 4.C Market clearing with mobile capital ...... 154 Appendix 4.D Few comparative statics ...... 154 Appendix 4.E Proof of Lemma 1 ...... 155 Appendix 4.F Proof of Lemma 2 ...... 156
ix TABLE OF CONTENTS
Appendix 4.G Alternative interpretation of Lemma 1 and Lemma 2 . . . . 157 Appendix 4.H State i’s frst-order conditions ...... 158 Appendix 4.I Technical description of federal solutions ...... 158 Appendix 4.J A proxy ...... 158 Appendix 4.K Replication of utility structure ...... 158 Appendix 4.L Tax ranking with mobile capital ...... 158
5 Synthesis and Outlook 161 5.1 Federal carbon pricing and transfers ...... 164 5.2 Models for fnding feasible and efective multilateral and federal policies 165 5.3 Why study multilevel policies ...... 167 5.4 Next steps ...... 168 5.4.1 Climate policy ...... 168 5.4.2 Beyond climate policy ...... 170 5.5 Outlook ...... 171 References ...... 173
Tools and Resources 175
x List of Figures
1.1 Diferences of EU member states ...... 15 1.2 Efective carbon price stringency across EU member states ...... 17
2.1 Evolution of EUA price and EUA future contracts for the year 2020 . . 31 2.2 Estimated efective carbon prices by country derived from the electricity, transport, pulp and paper, and cement sectors ...... 33 2.3 Optimal transfers per person according to private consumption in the year 2010 in thousands of dollars of purchasing power parity ...... 42 2.4 Optimal transfers per country according to private consumption in the year 2010 in billions of dollars of purchasing power parity ...... 43 2.5 Illustration of the cushioning efect of intra-ETS leakage with a carbon price foor ...... 45 2.6 Change in (a) annual CO2 emissions and (b) electricity production due to a higher CO2 price in Germany ...... 48 2.7 Change in annual electricity production (a) in European countries except Germany and (b) in Germany in the year 2020 due to a higher CO2 price in Germany ...... 49
3.1 Stylized illustration of the proof structure ...... 82 3.2 Comparison of capital-homogeneity-restriction ...... 85 3.3 Feasible federal price range ...... 88 3.4 Robustness test of the feasible federal price range ...... 91 3.5 Efective (consolidated) emission price per state ...... 92 3.6 Impact of population size diferences ...... 94 3.7 Sensitivity analysis of capital stock in rich states ...... 114 3.8 Robustness test of the feasible federal price range ...... 115 3.9 Sensitivity analysis of consumption changes for poor states ...... 116 3.10 Sensitivity analysis of consumption changes for rich states ...... 116 3.11 Ratio of aggregate federal emissions subject to transfer rule at each minimum price ...... 117
xi LIST OF FIGURES
4.1 Capital per capita and CO2-emission intensities in EU member states for the year 2014...... 122 4.2 Schematic sketch of production function indicating interplay of output and substitution elasticities...... 131 4.3 Stylized representation of the minimum and maximum uniform federal tax for two states ...... 134 4.4 Countries cluster with respect to capital per capital and population size into four groups ...... 138 4.5 Optimal uniform federal tax rates T i subject to equal per capita (EQ) and sovereignty (SO) transfers...... 139 4.6 Federal payment, transfer, and net payment at the federal minimum tax given equal per capita and sovereignty transfers per capita ...... 140 4.7 Federal payment, transfer, and net payment at the federal minimum tax given equal per capita and sovereignty transfers per country ...... 142 4.8 Net federal payment under equal per capita transfers ranked by capital per capita...... 143 4.9 Net federal payment under sovereignty transfers ranked by capital per capita...... 144 4.10 Incidence of the federal minimum emission tax...... 146 4.11 Relative change of utility levels between the multilevel and decentralized equilibrium under equal per capita and sovereignty transfers...... 159 4.12 Normalized uniform federal tax rates that maximize the utility of the respective country given mobile capital...... 160
xii List of Tables
1.1 Overview of the main objectives, challenges covered, and methodological characteristics for the main chapters of the thesis...... 21
2.1 Policy Scenarios ...... 47
3.1 Transfer criteria and operationalized federal transfer rules...... 70 3.2 Population-capital endowment matrix...... 87 3.3 Sensitivity analysis on parameters and the assumption of linear or log utility in consumption...... 88
4.1 Transfer rules...... 131 4.2 Correlation coefcients between ranking of T i and countries’ heterogeneities.139 4.3 Correlation coefcients given mobile capital...... 142 4.4 Correlation coefcients of net payments and heterogeneous country data. 144
xiii
1 Introduction
1 1. Introduction
For almost two centuries, the burning of fossil fuels has brought wealth and prosperity to many nations, while the accumulation of greenhouse gases in the atmosphere has increased rapidly. This causes global mean temperature to rise resulting in human- induced climate change. The global mean temperature exceeding pre-industrial levels by more than 1.5◦C to 2◦C would have irreversible and pervasive consequences for humanity and ecological systems. Some regions would become uninhabitable, and water and food supplies may become scarce or lost due to sea level rises in some places and desertifcation in others. In turn, this may threaten social cohesion due to conficts over resources such as water and food exacerbating inequality in already polarized societies (IPCC, 2014b; IPCC, 2018). With the adoption of the Paris Agreement, almost all countries in the world have made individual voluntary commitments to meet specifc emission levels in an efort to keep
“... the increase in the global average temperature to well below 2◦C above pre-industrial levels and pursuing eforts to limit the temperature increase to 1.5◦C above pre-industrial levels, recognizing that this would signifcantly reduce the risks and impacts of climate change”. Paris Agreement (UNFCCC, 2016, p. 3, Article 2a) The voluntary nationally determined contributions committed under the Paris Agreement, however, will miss the emission reduction that would be needed to achieve cost-optimal climate protection at 2◦C by 2030 (cf. Luderer et al., 2018; Roelfsema et al., 2020; Vrontisi et al., 2018). Following the trajectory of nationally determined contributions until 2030 results in a temporary exceeding of the 1.5◦C limit (IPCC, 2018). Adherence to this limit would be possible with more stringent mitigation strategies than those of nationally determined contributions, but only if implemented in time. Even though the social cost of storing greenhouse gases in the atmosphere is considered to be tremendous (cf. Carbon Pricing Leadership Coalition, 2017; IPCC, 2014a; Mattauch et al., 2020; Nordhaus, 2019), currently almost all individual emitters across the world can use the atmosphere as a storage space free of charge or at too low cost. From an economic point of view, this poses two elementary problems: First, a low price does not refect the scarcity of the storage space provided by the atmosphere, which leads to it being overused. Second, many ethical positions justify the assumption that the atmosphere belongs to all humans, i.e. it is common property. Charging a sufciently high carbon price could address these two points: It would make polluters pay for emissions1 and at the same time, a higher carbon price generates more governmental revenue which can be used to beneft humans, for instance, by providing public services, addressing inequality concerns, or reducing other distortionary taxes (e.g Franks et al., 2018).
1In addition, the higher the price the more advantageous it is for polluters to invest in carbon-neutral technologies.
2 Emission mitigation by pricing carbon emissions is an example of Pigouvian taxation, as it corrects for overuse of the atmosphere by charging a price for its usage. At the same time, carbon pricing raises distributional concerns about unfair burden-sharing, which in turn may hamper its feasibility (e.g Goulder, 2020). There is no consensus about the fairest rule to share the burden of climate policy (e.g Kverndokk, 2018). It seems indisputable, however, that countries that already achieved great prosperity through high emission levels in the past should now take on more of the burden of the post-carbon transition than countries that have emitted less rendering them unable to accumulate as much wealth. This is also laid down in the United Nations Framework Convention on Climate Change (UNFCCC):
“The Parties should protect the climate system for the beneft of present and future generations of humankind, on the basis of equity and in accordance with their common but diferentiated responsibilities and respective capabilities. Accordingly, the developed country Parties should take the lead in combating climate change and the adverse efects thereof.” (UNFCCC, 1992, p. 4, Article 3)
However, not only a large body of literature in political philosophy and climate ethics (e.g Gardiner, 2010) but also political debates are inconclusive as to how exactly these burdens should be distributed. The reality of global warming — that efective measures need to be implemented to limit temperature increase appropriately in time — calls for pragmatic perspectives on equitable distributions. For a long time, political and scientifc discourse was expected to deliver a global solution to combat global warming. But current political developments, such as the Paris Agreement, are far from such a solution. Instead, individual countries and regions are attempting to tackle the problem alone and are in the process of deciding on future climate policy implementation. The editorial board of Nature recently stated that the primary goal of science in this decade should be to help political actors to fnd
feasible policy measures that drastically reduce CO2 emissions to achieve the necessary temperature targets (Nature, 2019). In the spirit of more decentralized action, Elinor Ostrom argues that
“Given the slowness and confict involved in achieving a global solution to climate change, recognizing the potential for building a more efective way of reducing greenhouse gas emissions at multiple levels is an important step forward”. (Ostrom, 2009, p. 38)
This thesis serves Ostrom’s case by translating her notion of polycentric governance formalized to public economics. It analyzes how to build efective multilevel climate policy in federations by focusing on the benefts of such governance systems. Well-known examples of large federations are the United States, Canada, and the European Union
3 1. Introduction
(EU). Examples of smaller federations include Switzerland and the Federal Republic of Germany. This thesis concentrates on climate policies in the European Union, although many results are also applicable to other federations or multilateral systems. Common but also diferent interests of member states generally characterize federal systems. Common interests make the federation possible in the frst place. It is the diferent interests of the member states, however, that make it challenging to fnd common federal policies. Importantly, the size and design of redistribution between member states make distributional conficts in a federation inevitable. But how can demand for environmental action that difers across member states be leveraged on to improve the overall policy outcome? Specifcally, the thesis examines how carbon pricing design can succeed without solving the general distributional conficts in a federation. It identifes federal minimum carbon prices that create a win-win situation for all member states. It demonstrates that at such a minimum price, the largest donor state voluntarily gives transfers to other member states. In Chapter 2, I examine the shortcomings of the current EU emissions trading system (ETS) and relate these to the principles of solidarity, subsidiarity, and efciency. I show how a price foor (minimum price) can remedy shortcomings of the EU ETS, and discuss what socially optimal transfers in the EU could look like in theory. Furthermore, I present implications of a price foor in combination with additional carbon prices at member state level for the European electricity sector. With a uniform EU carbon price, I fnd that socially optimal transfers involve rich states becoming donors, and poor states becoming recipients of carbon price revenues. From the perspective of rich states, these transfers may be too high, making them politically unfeasible. Chapter 3 focuses on the feasibility uniform federal carbon price and transfer policy with diferently wealthy member states. I develop a general equilibrium model and focus on commonly used federal revenue distribution rules of equal per capita (egalitarian) transfers, and transfers in proportion to historical emission share levels (sovereignty) or actual emission payments of states (juste retour) — a rule that is often requested by EU member states. The heterogeneity of member states plays a decisive role in fnding feasible federal policy. If the diferences in wealth among states are too high, then a uniform federal price, together with federal egalitarian transfers, becomes infeasible. Anticipated and unanticipated federal sovereignty transfers as well as unanticipated juste retour transfers, however, make federal policy feasible irrespective of wealth diferences. In Chapter 4, I investigate the scope of progressive federal carbon price design if there are diferences in wealth (vertical inequality) and diferences in CO2 emission intensity (horizontal inequality) among member states. Therefore, I develop a general equilibrium model and apply it to EU data. In the EU, wealth (vertical) and CO2 emission intensity (horizontal inequality) are inversely correlated. In other words, rich states often have lower emission intensity, and vice versa. I fnd that in such a case, the uniform federal
4 1.1 Economics of climate change and public economics
carbon price becomes regressive. I show that appropriate transfers can counteract the regressive efect: transfers based on historical emission share levels (sovereignty transfers), i.e., how the EU calculates the bulk of revenue from the ETS, produce a progressive federal policy for EU citizens; in contrast, equal per capita pertains to the regressive efect of a uniform federal carbon price. The remainder of the introduction is structured as follows. In Section 1.1, I discuss the key topics within the economics of climate change and public economics with direct bearing on this thesis. Section 1.2 explains why I consider the EU as a laboratory for multinational and multilevel policies. Section 1.3 concludes with a more detailed formulation of the objective of this thesis and an outline of the main chapters of this thesis.
1.1 Economics of climate change and public eco- nomics
In this section, I review the issues of the economics of climate change and public economics that have a direct bearing on the core problems addressed in this thesis. In addition, I briefy explain how my thesis is linked to the issues reviewed, and how it sheds a diferent light on them. The issues discussed here are limits and solutions for multinational climate policy (Section 1.1.1), the role of voluntary contributions for public good provision (Section 1.1.2), proposals of fair distribution and burden-sharing in climate change mitigation (Section 1.1.3), and policymaking in federations (Section 1.1.4).
1.1.1 Multilateral climate policy: Limits and solutions Mitigating climate change is a global public good. Mitigation thus creates free-rider incentives, as any mitigation efort made by a state, country or region also benefts those that do not undertake their own mitigation eforts. Moreover, mitigation is thwarted by a lack of incentives for multilateral cooperation. The literature on international environmental agreements investigates these incentives and proposes solutions in game- theoretic frameworks. An overview of the problem is provided, for instance, in Barrett (1994); solution proposals are summarized, for example, in Kornek (2015, Ch. 1). One way to improve multilateral cooperation is to provide an appropriate focal point in such a climate policy game by using a uniform multinational carbon price instead of quantity targets (Cramton et al., 2015; Weitzman, 2014). In simple terms, this is because a single price is more tangible and easier for states to negotiate than emission reduction quantities. Another strand of the literature departs from the game-theoretic perspective and instead examines the efciency conditions of multinational climate policy. These studies
5 1. Introduction take a top-down perspective by considering a social planner or centralized policy (Chichilnisky and Heal, 2000; Chichilnisky and Heal, 1994; Sandmo, 2007; Shiell, 2003), or a top-level government with delegation authority over all lower levels of government (e.g d’Autumne et al., 2016). Chichilnisky and Heal were the frst to show that equity and efciency cannot be considered separately for carbon emission mitigation due to its global public good nature. In doing so, they contradicted the common economic wisdom that efciency can be considered separately from equity concerns. Their fnding implies that rich countries must give transfers to poor countries based on efciency reasons. See Engström and Gars (2015) for a recent overview of this literature strand. These theoretically optimal transfers, however, may be very high, and may become infeasible if they encounter the resistance of states or countries that follow their self-interest (cf. Gruber, 2000; Sandmo, 2004; Sandmo, 2007; Stavins, 1997; Wiener, 2007). Shifting the focus from multinational efciency concerns towards efective and feasible multinational climate policy is consistent with the pessimistic insights of the game theory literature and the outcome of the Paris Agreement (see also Franks, 2016, Ch. 1). The research presented in this thesis comes into play at this juncture. In Chapter 2, I apply the model of Chichilnisky and Heal. The results show that, even in the EU, theoretically optimal transfers across member states may be too high to be politically feasible. In Chapters 3 and 4, I depart from the concept of optimal transfers. Instead, I consider federal transfers that are applied in political practice in an efort to relate my results to actual policymaking. Chapter 3 expands on the fndings of Chapter 2 by developing a novel theory of voluntary participation of diferently rich states in federal policy in a game-theoretic framework. I model a uniform federal carbon price and commonly used transfers in coexistence with member state policies. In Chapter 4, I extend the dimension of wealth heterogeneity (vertical inequality) by the dimension of technological diferences in CO2 emission intensity (horizontal inequality) and study the burden implications of a uniform federal carbon price and transfer policy.
1.1.2 Voluntary provision of public goods The voluntary decentralized provision of public goods has been investigated for decades in a non-climate change related context. Motivated by NATO expenditure in the arms race with the Warsaw Pact, it was pioneered by Olson (1965) and Olson and Zeckhauser (1966). Olson and Zeckhauser (1966) show that the USA, as the wealthiest NATO member, also contributed the most to NATO expenditure in the arms race2. Later on, Olson (1986) shows that benevolent yet hegemonic states tend to create multinational systems (or alliances) for the provision of public goods. While public good provision benefts all states belonging to the alliance, the hegemonic state voluntarily bears a
2Some of today’s threats are, for example, climate change and pandemics. In this spirit, today’s multilateral entities fghting a common threat can be considered multilateral climate cooperations or funds but also organizations such as the World Health Organization.
6 1.1 Economics of climate change and public economics disproportionately large share of the costs of its provision. Bergstrom et al. (1986) formalize this consideration. When a common threat is involved, this strand of literature predicts that rich entities voluntarily contribute more to mitigating this threat than poor ones (see Bergstrom et al., 1986; Olson and Zeckhauser, 1966). In addition, Bergstrom et al. (1986) fnd that income redistribution from rich to poor entities can have a negative impact on the level of voluntary provision of public goods. Concerning the mitigation of climate change, two critical components can be derived from this research strand and specifcally from the framework of Bergstrom et al. (1986). The frst critical component relates to their fnding that a reduction in income disparities can reduce contributions to climate change mitigation. This perspective ignores the fact that individual wealth levels often stem from past high emission levels, meaning that wealth may have been generated at the expense of disadvantaged entities. It is also expected that
"the people at greatest risk from climate hazards are the poor, the vulnerable and the marginalized". (UN Economic and Social Council, 2016, preface, p. iii)
Combating climate change, therefore, also requires policy measures that incorporate equity concerns, which will be elaborated in the next section. Second, Bergstrom et al. (1986) originally frame their paper by considering individuals. Their fnding that the rich voluntarily donate more than poor individuals implies that disproportionately wealthy private entities could make a more signifcant contribution to combating climate change by private donations (in the sense of charity).3 Another interpretation is that they would also be willing to pay higher carbon prices. As an implicit donation, these carbon price revenues would then be available to governments to fnance public goods and services, which include the mitigation of and adaptation to climate change. This thesis expands on the insights into voluntary contributions and the vital role of diferent states (or countries) by turning Olson’s approach upside down. While Olson shows that large wealthy states have an incentive to create multinational regimes, I show in Chapter 3 how a multinational regime (federation) can encourage the wealthiest member state to become the largest donor to federal revenue. Chapter 4 shows that, even voluntary federal carbon pricing can put poor states in donor positions, making federal carbon pricing regressive.
3Anecdotal evidence of this theory applied to climate change is provided by Jef Bezos, the richest person in the world. He recently announced that he would donate USD 10 billion to help combat climate change (Bezos, 2020). In contrast, the Green Climate Fund set up by the UNFCCC whose funds mainly stem from developed countries, gathered pledges for donations worth USD 10.3 billion (Green Climate Fund, 2020).
7 1. Introduction
1.1.3 Distributional concerns and fair burden-sharing Ensuring a fair climate policy increases the chances of its acceptance and the near-term feasibility of more efective climate policies (Goulder, 2020). A large strand of literature develops and discusses a broad spectrum of equity criteria for burden-sharing in the climate change context (e.g Burtraw and Toman, 1992; Cazorla and Toman, 2001; Grubb et al., 1992; Kverndokk, 2018; Kverndokk and Rose, 2008; Pottier et al., 2017; Ringius et al., 2002; Rose, 1992; Rose and Stevens, 1993; Rose, Stevens, et al., 1998). Voluntary participation in multinational policy is considered one such equity criterion (e.g. Cazorla and Toman, 2001; Kverndokk and Rose, 2008). Equity criteria can be ex-post or ex-ante criteria. Ex-post equity criteria consider the situation after policy has been introduced (as in the case of voluntary participation). Ex-ante criteria use independent parameters, such as historical emission levels, population sizes or given levels of prosperity. Recent surveys of the literature are provided in Kverndokk (2018), Kverndokk and Rose (2008), Paterson (2001), and Pottier et al. (2017). It may come as no surprise that both public debate and academic contributions are inconclusive about the “best” criterion to ensure fair burden-sharing (e.g Kverndokk, 2018). In fact, it may never be possible to identify the “one criterion that suits all” situations. Goulder (2020) argues that the complexity of the problems and interests related to climate change of individual states and stakeholders requires more economic models that provide detail on the distribution of climate change and climate policy burden. Economics research has yet to fully grasp this problem (Goulder, 2020). In this vein, Roberton Williams III tweeted on Goulder’s talk at the annual meeting of the Allied Social Science Associations that
“We need to pay more attention to distributional efects. Economists far too often call the most efcient policy "best", even when it has problematic distributional efects. And distribution is a key element to political feasibility”. (Williams, 2019)
The research of my thesis is at the junction of distribution and feasibility. I examine burden-sharing criteria for the distribution of revenue from a uniform federal carbon price in terms of their potential feasibility in the sense that all states would agree to this federal policy. In the process, I mainly focus on distributional and feasibility concerns among unequal states or countries. In Chapter 2, I highlight the potential difculty of the most efcient federal policy as it may encounter the dissent of some states. For this reason, I develop a theory of voluntary participation of diferently wealthy states in federal policy in Chapter 3, and study commonly used transfer rules to distribute federal revenue. In Chapter 4, I add another critical dimension for climate policy, namely horizontal inequality in terms of the quality of a state’s mitigation technologies.
8 1.1 Economics of climate change and public economics
1.1.4 Fiscal federalism and its application to multilevel climate policy Research on federal systems accumulates several of the issues outlined in the previous three literature strands: A federation represents a multilateral system; requires some scope of voluntary participation of states, especially if new policies are to be introduced (viability); and needs to deal with burden-sharing arrangements between diferent member states. In addition, a federation already has an existing institutional structure of coexisting state and federal policies such that multilevel policies can, of course, be observed. As such, federal analysis can shed light on feasible climate policies in federations, whilst providing insights into multinational or multilevel climate policies in general. The literature on fscal federalism contains extensive research on federal systems. It investigates the efcient regulatory competence and the viability of federations (cf. Oates, 2005). The frst generation of literature on fscal federalism is particularly interested in the allocation of regulatory competence by determining which regulatory tasks should be centralized, and which should be left to decentralized decision-makers, especially concerning the provision of (local) public goods. The frst generation, based primarily on efciency and equity arguments, concludes similar to the literature on multinational environmental policy that the primary responsibility for the provision of federal public goods should be left to the central federal government (e.g Musgrave, 1959; Oates, 1972; Oates, 2000). On the other hand, decentralized regulation by state-level governments can be more efcient if state levels know or can address the needs of the local population better than the federal level (e.g Oates, 1972; Oates, 2011). The subsidiarity principle refects the proportionality between decentralized (state) and centralized (federal) regulation. Applying the subsidiarity principle means that the exercise of federal regulation is only appropriate if federal policy can improve the outcome of decentralized state policies. Thereby, the subsidiarity principle also represents a limitation of federal regulatory competence. The second generation of fscal federalism explores situations in which a federal government has only limited control or competence over the political actions of its states; it is therefore concerned with the viability of federal institutions (Oates, 2005). Only a few studies address limited environmental control in federal systems. Williams (2012) examines limited control by considering the simultaneous existence of state and federal emissions policies. In terms of efciency, he fnds that a federal tax is superior over quantity controls since the additivity of taxes prevents a mutual overruling of state and federal policies. Depending on the federal policy instrument used, he fnds that there are sometimes no optimal transfers. This thesis contributes to federalism research by translating Ostrom’s polycentric governance approach to public economics by analyzing the viability of multilevel climate
9 1. Introduction policy in federations. In Chapter 3, I develop such a model structure by applying the principle of subsidiarity to cover coexisting state and federal policy with the requirement that federal policy needs to deliver Pareto improvements for all states. Thereby, federal policy ensures the unanimous consent of all states towards federal policymaking. This thesis expands on the second generation of fscal federalism by including unanimity requirement and environmental regulation. In Chapter 4, I extend the novel model structure developed in Chapter 3 to analyze the progressive potential of federal transfers. As I have outlined before, previous literature argues that optimal transfers can either be too high to be politically feasible or they may not even exist. In Chapters 3 and 4, therefore, I depart from the abstract concept of optimal transfers and instead consider federal transfers applied in political practice to make my results applicable to actual policymaking.
Multilevel climate policies Two general perspectives on multilevel or coexisting state and federal policies can be distinguished. First, in the context of environmental regulation, multilevel policies are sometimes considered inefective and costly, as they constitute quasi double regulation for the same objective. In a partial equilibrium model, Böhringer and Rosendahl (2010) and Böhringer and Rosendahl (2011) show, for instance, that a green quota in coexistence with an emissions trading system may increase overall emissions, relating this to the EU emissions trading and member state policies. Second, in the context of fscal federalism, climate policy at one level of government may have an impact on government revenue at the other level, and vice versa. In this case, vertical fscal externalities occur. The precise defnition of a vertical fscal externality is when the upper and lower levels of government tax the same tax base, and these taxes afect revenue raised by the other level. Vertical fscal externalities have been extensively studied without considering environmental regulation, and often found to induce state governments to overtax the local tax base (Böhringer, Rivers, and Yonezawa, 2016; Bruellhart and Jametti, 2006; Dahlby and Wilson, 2003; Keen, 1998; Keen and Kotsogiannis, 2002). Vertical fscal externalities also occur in the context of emissions pricing, whether through taxation or emissions trading at multiple levels, as all levels of government regulate and draw revenue from pricing emissions. Consequently, the emissions pricing of one level may have an impact on the emissions pricing revenue of the other level. Böhringer, Rivers, and Yonezawa (2016) were the frst to assess the importance of vertical fscal externalities in the context of environmental regulation, albeit without considering strategically acting governments. Applied to the Canadian Federation, they fnd that the vertical fscal externality causes a state to implement environmental policy at low cost to itself, at the expense of the other states. Their result shows that the vertical
10 1.2 The European Union as a laboratory for multilevel climate policy
governmental dimension of emission pricing is an important dimension to consider, as it delivers shifts of advantages and disadvantages among states. This thesis sheds light on the vertical fscal externalities of multilevel climate policy in a game-theoretic framework. Instead of focusing on efciency losses of overlapping regulation, I concentrate on the benefcial coexistence of multilevel policies that facilitate more stringent climate policies than decentralized state solutions. In this context, I depart from previous literature, because the novel model approach, developed in Chapter 3, requires that federal policy coexists with state policy, and attains Pareto improvements to the decentralized solution. In this way, I ensure the viability of federal policy, while maintaining the sovereignty of member states. With regard to vertical fscal externalities, Chapter 3 shows that the efectiveness of federal policy depends on the federal transfer design4 and on whether or not states internalize the vertical fscal externality. If this is the case and if federal transfers are based on actual emission payments by states to the federal government (juste retour), I fnd that federal policy becomes inefective. This is a novel result, as juste retour transfers were considered efective in model frameworks that did not consider the internalization of the vertical fscal externality (cf. d’Autumne et al., 2016; Shiell, 2003).
1.2 The European Union as a laboratory for multi- level climate policy
In the previous section, I presented lines of research and results from the economics of climate change and public economics. I explained how they relate to the key issues addressed in this thesis, and argued that these strands accumulate in the research on federalism. In this section, I survey climate policymaking in the EU to show how studying this entity can enable the EU to be used as a “laboratory” to gain more knowledge about feasible and efective multilevel policies in multilateral, federal-like systems. To allow a direct link to the research discussed in Section 1.1, I follow the same structure of subsections.
1.2.1 Multilateral climate policy Climate policy was frst introduced in the EU in the 1990s, but signifcantly gained momentum around the year 2000, when the Kyoto protocol was approved (cf. Delbeke and Vis, 2016). The EU ETS, introduced in 2005, is the main EU climate policy instrument. Emitters covered by the EU ETS must hold emission allowances before they may emit greenhouse gases. These allowances are either allocated free of charge (grandfathering) or must be purchased on the emissions allowances market (auctioning).
4For an more detailed explanation of federal transfer rules considerd, see Chapter 2 and Chapter 3, Table 3.1.
11 1. Introduction
The European Commission estimates that 57% of all allowances were auctioned during 2013-2020, while the remaining 43% were allocated free of charge (EC, 2020a). By auctioning allowances, the EU ETS generates a uniform price signal for all entities participating in emissions trading. The majority of auction revenue is distributed to member states based on their historical emission shares (88% until 2019 and 90% from 2020 onwards). The literature refers to this distribution rule as the sovereignty rule (e.g Böhringer, Rivers, Rutherford, et al., 2015). The remaining auction revenue is allocated to the least wealthy EU member states. The EU ETS applies to all EU member states, and covers 11,000 heavy energy-using installations and airlines operating between these countries. As a result, it covers around 45% of the EU’s greenhouse gas emissions (EC, 2020b). There are several reform proposals or considerations for the EU’s future climate policy.5 These range from specifc policy proposals, such as an EU-wide minimum carbon price, a price foor for the EU ETS, or the use of coexisting fanking measures by member states (cf. Burtraw, Keyes, et al., 2018; Edenhofer et al., 2017), to a large-scale transition of the European economy to achieve carbon neutrality by 2050, as outlined in the European Green Deal (EC, 2019). Protests by the Yellow Vest movement are also expected to further enhance the rationale of relief for poor or disadvantaged population groups at the member state and European level (e.g Douenne and Fabre, 2020). This thesis contributes to the understanding of the feasibility of EU climate policies. Chapter 2 argues that a price foor can counteract shortcomings of the EU ETS. Chapter 3 shows how an efective minimum carbon price can be found that is endorsed by all member states. Chapter 4 uses European data and shows that such a minimum price, together with sovereignty transfers, is progressive.
1.2.2 Voluntary participation of member states in contribut- ing to EU climate targets In the case of the EU, I consider a state willing to voluntarily participate or contribute to a federal public good if it can agree to federal policy. In many cases, EU policy must achieve a qualifed majority of member states. However, "sensitive matters" such as EU fnances and taxation require the unanimous agreement of member states (cf. EC, 2020d; Talus, 2013). Although the unanimity rule on the EU level ensures the voluntary participation of all states, it may also become an obstacle to new EU policies. For example, Poland, a country with a relatively low income and a carbon-intensive energy sector, vetoed the European Commission’s Energy Tax reform proposal several times. As a result, the EU energy tax reform has been pending since 2011. In order to implement an EU carbon
5In 2021, the EU ETS enters into phase 4, which undertakes some major reforms such as the market stability reserve, which allows for a more fexible cap-adjustment. For a discussion of this, see, for instance, Perino (2018).
12 1.2 The European Union as a laboratory for multilevel climate policy
price, such as a carbon tax, the EU would also require unanimity, whereas an ETS requires a two-thirds majority vote (Talus, 2013). In practice, this has meant that the quantity-based ETS was the only politically feasible carbon-pricing instrument at the time of its inception (Skjærseth and Wettestad, 2010; Talus, 2013). Chapters 3 and 4 focus on unanimity. In Chapter 3, I investigate how wealth diferences across states in general can hamper federal policy. I fnd that rich member states in particular place limits on federal policy stringency because they become the largest net donors of federal revenue. Chapter 4 provides the most comprehensive analysis of this thesis with regard to EU climate policy. Its results suggest that Poland and Germany are the countries that are most likely to be the frst ones to veto more stringent EU climate/energy policies. The rationale is that they are both large countries with
relatively large CO2 emission intensities, such that a uniform federal carbon price would impose on them a relatively large per capita burden.
1.2.3 Burden-sharing and transfers The Yellow Vest movement, Poland’s ongoing EU energy tax veto, and the fear of a transfer union have in common their underlying concern about an unfair burden, all of them seeking to hinder EU policy feasibility. Burden-sharing, and especially transfers between member states, are a sensitive topic in EU politics. Two prominent positions, albeit in a much more general context than climate policy, are represented by French President Macron, stating that “If you don’t want transfer payments, you don’t want a common Europe” 6, in contrast to the title published in the Economist “We don’t want no transfer union” — referring to German opposition 7. The EU draws its budget, among other things, from member states’ contributions in relation to gross domestic product. In 2017, major parts of the budget are used to support growth, creating jobs and reducing economic gaps (EC, 2020c). Redistributive transfers to poorer entities are also relevant for implementing climate policy, both from a theoretical perspective, as discussed in Section 1.1, and from a practical perspective, as the Yellow Vest movement shows. Member states, however, often request juste retour transfers from the EU (Warleigh, 2004). Juste retour literally means "fair return". It presumes that the payments made by a state to the EU give that state the right to receive EU revenue transfers equal to that payment. The majority of EU ETS auction revenue is distributed based on member states’ historical emission share levels, as outlined in Section 1.2.1.
6Karin, F., 2015. Frankreich hält EU nur als Transferunion für überlebens- fähig. WirtschaftsWoche, [online] Available at: https://www.wiwo.de/politik/europa/ europa-frankreich-haelt-eu-nur-als-transferunion-fuer-ueberlebensfaehig/12257362.html [Published 31 August 2015] [Accessed 14 March 2020]. 7The Economist, 2010. [online] Available at: https://www.economist.com/europe/2010/12/02/ we-dont-want-no-transfer-union [Published 2 December 2010] [Accessed 14 March 2020].
13 1. Introduction
The results of this thesis show that sovereignty transfers are an plausible choice for EU climate policy. In Chapter 3, I demonstrate that they make federal policy feasible and efective, irrespective of the size of wealth diferences across states. In contrast, equal per capita (egalitarian) transfers may hamper feasibility if wealth diferences become too high. Juste retour transfers become inefective if states internalize the vertical fscal externality (cf. 1.1.4), but perform similarly to sovereignty transfers if the vertical fscal externality is not internalized. Chapter 4 shows that sovereignty make federal carbon pricing progressive, while egalitarian transfers retain the regressive efect.
1.2.4 The federal European Union A federation represents a union of quite heterogeneous member states, which can be exemplifed by considering the EU. EU member states difer, for instance, in terms of technological factors, demography and wealth. Figure 1.1 shows diferences in technological factors in terms of aggregate capital 8 stock (A), CO2 emission intensity (B), and labor supply (C) . Since the labor supply is roughly proportional to population size, (C) also refects the demographic diferences in population size. The ratio of aggregate capital stock to population size shows the diferences in wealth measured in capital per capita (D). The choice of colors in Figures 1.1 A-D divides EU member states into two groups, with one above the respective mean value (white) and the other one below it. These four dimensions (A-D) of diferences between member states show that there is a signifcant diference between northern and western EU member states and those in the east. Northern and western EU states have a higher level of per capita capital (D), a less CO2 emission-intensive production (B), and, in many cases, a larger labor supply and population size (C) than eastern EU member states.
8Data based on World Development Indicators (WDI), Eurostat, and Berlemann and Wesselhöft (2017) and Berlemann and Wesselhöft (2014). For details, see Chapter 4.
14 1.2 The European Union as a laboratory for multilevel climate policy
Figure 1.1: Diferences of EU member states - The fgure shows selected diferences between EU member states. White shading corresponds to the mean. Data based on Eurostat, World Development Indicators and Berlemann and Wesselhöft (2017) and Berlemann and Wesselhöft (2014), for details see Chapter 2, Section 4.4.1.
15 1. Introduction
EU legislation ensures that diferences between member states are taken into account, for instance by applying the principle of subsidiarity9. As also discussed in Section 1.1.3, this principle presupposes that states are closer to their local citizens often giving them a better understanding of local needs and enabling them to address these needs more efciently at the state level than at EU level. The EU only intervenes if objectives cannot be achieved adequately by the member states and could be better achieved at EU level. In this thesis, I address the principle of subsidiarity both explicitly (Chapter 2) and implicitly (Chapter 3 and 4) by modeling coexisting state federal policies, where federal policy intervenes only if it can deliver improvements to the states’ solutions. I consider heterogeneous states throughout Chapters 2-4. Chapter 2 relates to the theory of Chichilnisky and Heal (1994), whose results concentrate on diferences in wealth (measured by diferences in consumption). The numerical partial equilibrium of Chapter 2 entails a detailed representation of the member states’ power generation facilities, and compares investment in generation capacities subject to diferent carbon price scenarios. Chapter 3 concentrates on diferences in wealth and populations size, and their implications for the feasibility of federal policy. Chapter 4, containing the most comprehensive analysis, considers diferences in wealth, population and labor size, and CO2 emission intensity. As Figure 1.1 B and D shows, wealthy EU states are also those with lower CO2 emission intensities. The results of Chapter 4 show that this characteristic threatens to make an EU-wide carbon price regressive if no appropriate redistributive transfers are implemented.
National climate policies and how they consolidate in the EU In this section, I provide a brief overview of states’ energy and climate policies coexisting with the EU ETS, as well as efective aggregate carbon price rates per state, to give an idea of the varying stringency across member states. I show that there is a coexistence of member states’ and EU climate policy, with a high degree of diferences in policy stringency across states. The review in this section, however, is not all-encompassing — for an extensive overview thereof, see, for instance, OECD (2013) and OECD (2018), and from a legal perspective Talus (2013). With regard to the EU ETS, some member states complement the EU ETS carbon price with additional measures such as a price foor (UK), or renewable energy support policies (e.g. France, Germany). For those sectors that are not covered by the EU ETS, some states have already implemented carbon taxes or prices (e.g. Sweden) or are considering to do so (e.g. Germany). Also, some states implicitly support their emission-intensive sectors by allowing reductions or exemptions from taxes and levies (e.g. for Germany see Rosenberg et al. (2011)).
9For the legal basis, cf. Article 5(3) of the Treaty on European Union (TEU), Pro- tocol (No 2) on the application of the principles of subsidiarity and proportionality, and https://www.europarl.europa.eu/factsheets/en/sheet/7/the-principle-of-subsidiarity.
16 1.2 The European Union as a laboratory for multilevel climate policy
The OECD (2018) study has calculated efective carbon prices rates for all EU member states over all sectors. Figure 1.2 uses the OECD’s results and shows how far
a member state’s efective carbon price rate is from a carbon price of 60 EUR/tCO2 (carbon price gap). The diferent shading in the fgure illustrates how strongly states difer in their climate policy stringency. The larger the carbon price gap, the lower the state’s efective carbon price and the darker orange its color. Similarly, the smaller the gap, the greener the color. Luxembourg’s and France’s efective prices, for example, are 10 close to 60 EUR/tCO2, while the Czech Republic and Poland exhibit the largest gap. The fgure also suggest that carbon price rates largely difer across member states. Even
if a country a comes close to 60 EUR/tCO2, this price may still not be high enough. The German Environment Agency, for instance, estimates that the German carbon price
should be roughly 180 EUR/tCO2, (cf. Umweltbundesamt, 2019). For estimates of the social costs of carbon, as calculated by governmental organizations and scientists, see, for example, Agency (2016), Carbon Pricing Leadership Coalition (2017), IPCC (2014a), and Nordhaus (2019) or an overview provided in Mattauch et al. (2020).
Figure 1.2: Efective carbon price stringency - The fgure depicts member states’ diferences in local carbon price stringency, measured by the diference between each state’s
efective carbon price rate to 60 EUR/tCO2, as calculated by OECD (2018). The fgure shows that the efective carbon price rates across member states vary signifcantly.
Concerning vertical fscal externalities in the EU, no study exists to date. A similar probability of signifcant magnitude as for the Canadian situation analyzed by Böhringer, Rivers, and Yonezawa (2016), however, can be assumed. It is also unclear whether EU member states would internalize the vertical fscal externalities. As there are not yet
10It may come as a surprise that Sweden exhibits a large gap, even though it has one of the highest carbon prices in the world. OECD (2018) calculated the efective carbon rates based on emissions from combustion. Sweden has priced many emissions from fossil fuels, but hardly prices emissions from biomass.
17 1. Introduction any studies on these two aspects, they should be analyzed in future research to better understand the incentives of EU member states to ensure efective EU policies. The results of this thesis contribute to understanding and assessing EU multilevel policies. Chapter 2 shows that a minimum price for the EU ETS could make member states’ companion policies more efective by eliminating the so-called waterbed efect Perino (cf. 2018), which represents emission leakage between states subject to state policies. Chapters 3 and 4 focus on carbon prices at all governmental levels. In Chapter 3, I compare cases when states do or do not internalize vertical fscal externalities. My result shows that the internalization yields inefective EU policies if juste retour transfers are used. In such a case, juste retour transfers are not recommendable for EU carbon price revenue distribution. The results of Chapter 4 in particular, which uses a model calibrated to EU data, demonstrate that large emission intensities make the economies of Eastern Europe vulnerable concerning the burden of carbon pricing.
1.3 Objectives and outline
In this thesis, I argue for a new perspective on coexisting climate policies at multiple levels of government. Due to the complexity of the interests of individual and overlapping policies, sound policy analysis in search of feasible and efective solutions must take into account multilateral climate policies and their related equity concerns, the voluntary participation of states in federal policy, and the vertical interaction of multilevel policies. The objective of this thesis is thus to take frst steps towards bringing polycentricity into public economics by designing a feasible and efective multilevel climate policy in federations. It pursues the following four main research questions:
1. How do diferences between member states afect the search for common, uniform federal carbon prices?
2. How does revenue distribution by commonly used transfers help multilevel policies to function efectively?
3. Which member states become donors of federal transfers, and which become recipients?
4. What makes a federal carbon price acceptable and what does this imply for the EU?
A federation with multilevel policies and diferent member states is a rather complex system. Consequently, many characteristics of a federation are mutually dependent, so that the answers to these questions are often intertwined. In order to analyze such a complex system in a structured way, this thesis model increases model complexity chapter by chapter.
18 1.3 Objectives and outline
I next summarize the central elements of the subsequent chapters, highlighting how they build on each other in their respective analyses. Based on the facts described in the previous two sections, throughout the remaining chapters, I focus on a uniform carbon price at the federal level and the distribution of federal carbon price revenue. All chapters suppose diferences in wealth across states, and investigate how multilevel policies function. These chapters, however, difer in their complexity, originality, and novelty of modeling. Furthermore, Chapters 3 and 4 focus exclusively on the second-best settings, consisting of a uniform federal carbon price and commonly used transfer rules to improve knowledge of existing policy and how it can be improved with simple heuristics. The objective of Chapter 2 [Agreeing on an EU ETS price foor] is to identify shortcomings of the EU ETS and to show how a price foor represents a remedy. The normative design principles of solidarity and subsidiarity are discussed, and it is shown how they could guide EU policy design. Furthermore, the efciency implications of a uniform ETS price are studied, especially on its implication for optimal transfer design. It is also shown how intra-ETS leakage can be mitigated with a carbon price foor (cf. “waterbed-efect” Perino (2018)). A numerical exercise is provided to quantify the cost efects of the EU ETS price foor proposal in coexistence with unilateral state policy within the European power sector. This chapter reviews previous literature fndings of shortcomings of the low EU ETS price, and discusses the interaction between heterogeneous member states and EU climate policies by drawing on existing literature and extending with own my insights. As the target audience of this chapter is academics and policymakers who do not necessarily have a formal model background, this chapter uses and explains formal model arguments in an easy-to-grasp format with little use of technical terms. Chapter 2 employs three models. The frst modeling exercise is based on an EU application of the static general equilibrium model of Chichilnisky and Heal (1994). I call this a simple application, because I use equal social welfare weights to study the efciency implication of an ETS on optimal transfers between EU member states in the simplest form. The Chichilnisky and Heal (1994) model gives frst-best solutions, as it supposes the existence of optimal lump-sum transfers. The second model employed is a partial equilibrium to explain how the waterbed efect can be mitigated by a binding ETS price foor. The third model is a dynamic partial equilibrium model to study cost-efcient pathways into long-term investment decisions for Europe’s electricity system (LIMES-EU as documented in Nahmmacher et al. (2014)). Based on the application of the model by Chichilnisky and Heal (1994), I argue that policy design seeking to develop efcient solutions is of little assistance in fnding feasible climate policies because it neglects the dimension of unanimity. This motivates the next chapter: Chapter 3 [Make or brake] revolves around feasibility. To approach the feasibility of federal policy, I choose the setting which has the least risk of state veto. In other words, I suppose that federal policy must ensure the unanimity of member states, while, at
19 1. Introduction the same time, member states set additional emission policies (subsidiarity principle). I depart from considering optimal transfers and instead concentrate on commonly used transfers, namely federal revenue distribution based on equal per capita (egalitarian) transfers, transfers in proportion to states’ historical emission levels (sovereignty), and states’ actual emission payments to the federal authority (juste retour). This chapter develops a novel theory of voluntary participation of diferently wealthy states in federal emissions pricing as follows: I use a static multilevel general equilibrium model which represents a federation. Member states can difer in their capital stock and population size. Population and capital are immobile across states. The entire population consists of identical consumers within a state. Each consumer rents out its capital endowment to the domestic frm. Consumers receive a transfer from the revenue of state and federal emission prices. Firms pay for the emission of harmful transboundary emissions during the production of the fnal good. Each consumer derives utility from the consumption of a private good and disutility from emissions. But to produce the private good, frms must use emissions and capital. State and federal governments choose optimal emission prices that strike a balance between emissions and private good consumption, and recycle the revenue back to the population. Each state government charges a price on domestic emissions, and distributes the revenue equally among its population. The federal government sets a uniform emission price, in addition to state prices, in case this leads to a Pareto improvement relative to the decentralized state policy solution. It distributes revenue based on the egalitarian, sovereignty, or juste retour criteria. Since the distribution rules are given, the federal solution is a second-best optimum. The model uses a game-theoretic approach. At the frst stage, the federal government sets a federal uniform emission price. Its objective is to make at least one state better of, while no other state is worse of compared to the decentralized solution (Pareto improvement). The federal government has information on individual heterogeneous interests and reactions of its member states and thus acts as a Stackelberg leader. At the second stage, based on the reactions of frms and consumers, and taking the federal price as a given, each state government non-cooperatively sets a price on domestic emissions. Its objective is to maximize the utility for the local population. At the third stage, consumers and frms in each state solve their optimization problems, taking all prices and transfers as given. I solve the model analytically using some simplifying assumptions to arrive at the main arguments and relax these in the numerical analysis to test the robustness of the analytical results. Chapter 4 [Technology beats capital] complements the dimension of diferences in wealth (vertical inequality) with diferences in states’ CO2 emission intensity (horizontal inequality). It is motivated by the observation that public support for environmental tax reforms is closely linked to the tax burden imposed on consumers. Public support can be improved by strategically complementing tax reforms by transfer rules that recycle
20 1.3 Objectives and outline
Table 1.1: Overview of the main objectives, challenges covered, and methodological characteristics for the main chapters of the thesis.
Chapter 2 Chapter 3 Chapter 4 Objectives Policy coexistence and subsidiarity ✓ ✓ ✓ Uniform federal carbon price ✓ ✓ ✓ Optimal transfers ✓ Feasibility (unanimity) ✓ ✓ Commonly used transfers rules ✓ ✓ Progressivity ✓
Challenges addressed Diferences in wealth ✓ ✓ ✓ Diferences in population size ✓ ✓ Diferences in CO2 emission intensity ✓ Capital mobility ✓
Methodology Theoretical ✓ ✓ ✓ Numerical ✓ ✓ ✓ multilevel ✓ ✓ ✓ Model development ✓ ✓ Game-theoretic ✓ ✓ First-best ✓ Second-best ✓ ✓ ✓ EU-calibrated application ✓ ✓ tax revenue to consumers. This chapter traces consumers’ burden of uniform federal tax payments to member states’ diferences in wealth and technological CO2 emission intensity for two commonly used transfer rules in a simple static general equilibrium which builds on the model developed in Chapter 3. The model formulation difers from the model in Chapter 3 as follows. In addition to capital endowments, consumers also rent out their labor to frms. I contrast the case of immobile to perfectly mobile capital, while labor is considered immobile. To produce the private good, the representative frm in each state uses capital, labor, and emissions. Firms difer in their abatement technology, which I model using an emission-augmenting factor. I prove the main mechanism underlying my argumentation of Chapter 4 analytically, and solve the full model numerically calibrated to EU data. Table 1.1 provides an overview of the main objectives, challenges addressed, and methods used in the main chapters of the thesis.
21
REFERENCES
References
Agency, E. P. (2016). Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866. Interagency Working Group on Social Cost of Greenhouse Gases, United States Government, [Accessed 2020-05-06]. Barrett, S. (1994). Self-Enforcing International Environmental Agreements. Oxford Economic Papers, 46, 878–894. Bergstrom, T., Blume, L., & Varian, H. (1986). On the private provision of public goods. Journal of Public Economics, 29, 25–49. Berlemann, M., & Wesselhöft, J. (2017). Aggregate Capital Stock Estimations for 122 Countries: An Update. Review of Economics, 68 (2), 1158. Berlemann, M., & Wesselhöft, J. (2014). Estimating aggregate capital stocks using perpetual inventory method - new empirical evidence for 103 countries -: Working Paper Series. Review of Economics, 65, 1–34. Bezos, J. (2020). Today, I’m thrilled to announce I am launching the Bezos Earth Fund. Climate change is the biggest threat to our planet. I want to work alongside others both to amplify known ways and to explore new ways of fghting the devastating impact of climate change on this planet we all share. This global initiative will fund scientists, activists, NGOs — any efort that ofers a real possibility to help preserve and protect the natural world. We can save Earth. It’s going to take collective action from big companies, small companies, nation states, global organizations, and individuals. I’m committing $10 billion to start and will begin issuing grants this summer. Earth is the one thing we all have in common — let’s protect it, together. - Jef [Posted 2020-02-17. Accessed 2020-04-06]. Böhringer, C., Rivers, N., Rutherford, T., & Wigle, R. (2015). Sharing the burden for climate change mitigation in the Canadian federation. Canadian Journal of Economics, 48 (4), 1350–1380. Böhringer, C., Rivers, N., & Yonezawa, H. (2016). Vertical fscal externalities and the environment. Journal of Environmental Economics and Management, 77, 51–74. Böhringer, C., & Rosendahl, K. (2010). Green promotes the dirtiest: on the interaction between black and green quotas in energy markets. Journal of Regulatory Economics, 37 (3), 316–325. Böhringer, C., & Rosendahl, K. (2011). Greening Electricity More Than Necessary: On the Cost Implications of Overlapping Regulation in EU Climate Policy. Schmollers Jahrbuch 131 (2011), 469 – 492 Duncker & Humblot, Berlin. Bruellhart, M., & Jametti, M. (2006). Vertical versus horizontal tax externalities: An empirical test. Journal of Public Economics, 90 (10), 2027–2062.
23 1. Introduction
Burtraw, D., Keyes, A., & Zetterberg, L. Companion Policies under Capped Systems and Implications for Efciency—The North American Experience and Lessons in the EU Context (RFF Report, Ed.). Ed. by RFF Report. 2018. Burtraw, D., & Toman, M. (1992). Equity and International Agreements for CO2 Containment. Journal of Energy Engineering, 118 (2), 122–135. Carbon Pricing Leadership Coalition. (2017). Report of the High-Level Commission on Carbon Prices. Cazorla, M., & Toman, M. (2001). International Equity and Climate Change Policy. Climate change Economics and Policy, 235–247. Chichilnisky, G., & Heal, G. (Eds.). (2000). Environmental Markets. Equity and Efciency. Columbia University Press. Chichilnisky, G., & Heal, G. (1994). Who should abate carbon emissions? Economics Letters, 44, 443–449. Cramton, P., Ockenfels, A., & Stoft, S. (2015). An International Carbon-Price Commitment Promotes Cooperation. Economics of Energy & Environmental Policy, 4 (2), 51–64. Dahlby, B., & Wilson, L. (2003). Vertical fscal externalities in a federation. Journal of Public Economics, 87 (5-6), 917–930. d’Autumne, A., Schubert, K., & Withagen, C. A. (2016). Should the carbon price be the same in all countries? Jounal of Public Economic Theory, 18, 709–725. Delbeke, J., & Vis, P. (Eds.). (2016). EU Climate Policy Explained. European Union. Douenne, T., & Fabre, A. (2020). French attitudes on climate change, carbon taxation and other climate policies. Ecological Economics, 169, 106496. EC. Auctioning [Accessed 2-May-2020]. https://ec.europa.eu/clima/policies/ets/ auctioning_en. Accessed 2-May-2020. 2020. EC. EU Emissions Trading System (EU ETS) [Accessed 2-May-2020]. https://ec.europa. eu/clima/policies/ets_en. Accessed 2-May-2020. 2020. EC. (2020c). How the EU budget is spent [Accessed 2020-05-06]. EC. (2019). The European Green Deal. [Accessed 6-May-2020]. EC. Voting system: Unanimity [Accessed 6-May-2020]. https://www.consilium.europa. eu/en/council-eu/voting-system/unanimity. Accessed 6-May-2020. 2020. Edenhofer, O., Flachsland, C., Wolf, C., Schmid, L. K., Leipprand, A., Koch, N., Kornek, U., & Pahle, M. Decarbonization and EU ETS Reform: Introducing a price foor to drive low-carbon investments. Berlin, 2017. Engström, G., & Gars, J. (2015). Optimal Taxation in the Macroeconomics of Climate Change. Annual Review of Resource Economics, 7 (1), 127–150. Franks, M. (2016). Rents, Taxes, and Distribution: Towards a New Public Economics of Climate Change: Ph.D. thesis. Berlin Institute of Technology.
24 REFERENCES
Franks, M., Lessmann, K., Jakob, M., Steckel, J. C., & Edenhofer, O. (2018). Mobilizing domestic resources for the Agenda 2030 via carbon pricing. Nature Sustainability, 1 (7), 350–357. https://doi.org/10.1038/s41893-018-0083-3. Gardiner, S. (Ed.). (2010). Climate Ethics: Essential Readings. Oxford University Press. Goulder, L. (2020). Timing Is Everything: How Economists Can Better Address the Urgency of Stronger Climate Policy. Review of Environmental Economics and Policy, 14 (1), 143–156. Green Climate Fund. (2020). About GCF [Accessed 6-April-2020]. Grubb, M., Sebenius, J., Magalhaes, A., & Subak, S. (1992). Sharing the burden. In I. M. Mintzer (Ed.), Confronting Climate Change: Risks, Implications and Responses (pp. 305–322). Cambridge University Press. Gruber, L. (2000). Ruling the World: Power Politics and the Rise of Supranational Institutions. Princeton University Press. IPCC, ed. Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change: [Edenhofer, O., R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J. Savolainen, S. Schlömer, C. von Stechow, T. Zwickel and J.C. Minx]. Cambridge, United Kingdom, New York, NY, USA, 2014. IPCC. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (Core Writing Team, R.K. Pachauri and L.A. Meyer, Ed.). Ed. by Core Writing Team, R.K. Pachauri and L.A. Meyer. Geneva, Switzerland, 151 pp., 2014. IPCC. Global Warming of 1.5◦C. An IPCC Special Report on the impacts of global warming of 1.5◦C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and eforts to eradicate poverty ([Masson-Delmotte, V., P. Zhai, H.-O. Pörtner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Péan, R. Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfeld], Ed.). Ed. by [Masson-Delmotte, V., P. Zhai, H.-O. Pörtner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Péan, R. Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfeld]. In press, 2018. Keen, M. (1998). Vertical Tax Externalities in the Theory of Fiscal Federalism. IMF Staf Papers, 45. Keen, M., & Kotsogiannis, C. (2002). Does federalism lead to excessively high taxes? The American Economic Review, 92 (1), 363–370.
25 1. Introduction
Kornek, U. (2015). Designing International Climate Agreements: An Economic Analysis of Free-riding Incentives: Ph.D. thesis. Berlin Institute of Technology. Kverndokk, S. (2018). Climate Policies, Distributional Efects and Transfers Between Rich and Poor Countries. International Review of Environmental and Resource Economics, 12 (2-3), 129–176. Kverndokk, S., & Rose, A. (2008). Equity and Justice in Global Warming Policy. International Review of Environmental and Resource Economics(2), 135–176. Luderer, G., Vrontisi, Z., Bertram, C., Edelenbosch, O., Pietzcker, R., Rogelj, J., de Boer, H., Drouet, L., Emmerling, J., Fricko, O., Fujimori, S., Havlík, P., Iyer, G., Keramidas, K., Kitous, A., Pehl, M., Krey, V., Riahi, K., Saveyn, B., & Kriegler, E. (2018). Residual fossil CO 2 emissions in 1.5–2◦C pathways. Nature Climate Change, 8, 626–633. Mattauch, L., Creutzig, F., aus dem Moore, N., Franks, M., Funke, F., Jakob, M., Sager, L., Schwarz, M., Voss, A., Beck, M., Daub, C., Drupp, M., Ekardt, F., Hagedorn, G., Kirchner, M., Kruse, T., Loew, T., Neuhof, K., Neuweg, I., . . . Wallacher, J. (2020). Antworten auf zentrale Fragen zur Einführung von CO2 -Preisen (Version 2.0). Scientists for Future. Musgrave, R. (1959). The Theory of Public Finance. McGraw-Hill. Nahmmacher, P., Schmid, E., & Knopf, B. Documentation of LIMES-EU - A long-term electricity system model for Europe. https://www.pik-potsdam.de/members/ paulnah/limes-eu-documentation-2014.pdf. 2014. Nature. (2019). The scientifc events that shaped the decade. Nature. Nordhaus, W. (2019). Climate Change: The Ultimate Challenge for Economics. American Economic Review, 109 (6), 1991–2014. Oates, W. (1972). An Essay on Fiscal Federalism. Journal of Economic Literature, 37, 1120–1149. Oates, W. (2000). Fiscal competition and European Union: contrasting perspectives. Regional science & urban economics, 31, 133–145. Oates, W. (2011). Fiscal Federalism. Edward Elgar Publishing Ltd. Oates, W. (2005). Toward A Second-Generation Theory of Fiscal Federalism. Interna- tional Tax and Public Finance, 12 (4), 349–373. OECD. (2013). Efective Carbon Prices. OECD Publishing. %5Curl%7Bhttp://dx.doi. org/10.1787/9789264196964-en%20%7D OECD. Efective Carbon Rates 2018: Pricing Carbon Emissions Through Taxes and Emissions Trading (OECD Publishing, Ed.). Ed. by OECD Publishing. Paris, 2018. Olson, M. (1986). A Theory of the Incentives Facing Political Organizations. Neo- Corporatism and the Hegemonic State. International Political Science Review, 7 (2), 165–189. https://doi.org/https://doi.org/10.1177/019251218600700205
26 REFERENCES
Olson, M. (1965). The Logic of Collective Action: Public Goods and the Theory of Groups (Vol. 20 printing, 2002). Harvard University Press. Olson, M., & Zeckhauser, R. (1966). An Economic Theory of Alliances. The Review of Economics and Statistics, 266–279. Ostrom, E. A Polycentric Approach for Coping with Climate Change: Policy Research Working Paper 5095. 2009. Paterson, M. (2001). Principles of justice in the context of global climate change. In U. Luterbacher & D. Sprinz (Eds.), International Relations and Global Climate Change (pp. 119–126). MIT press. Perino, G. (2018). New EU ETS Phase 4 rules temporarily puncture waterbed. Nature Climate Change, 8 (4), 262–264. Pottier, A., Méjean, A., & Godard, O. (2017). A Survey of Global Climate Justice: From Negotiation Stances to Moral Stakes and Back. International Review of Environmental and Resource Economics, 11 (1), 1–53. Ringius, L., Torvanger, A., & Underdal, A. (2002). Burden Sharing and Fairness Princi- ples in International Climate Policy. International Environmental Agreements(2), 1–22. Roelfsema, M., van Soest, H., Harmsen, M., van VUUREN, D., Bertram, C., den Elzen, M., Höhne, N., Iacobuta, G., Krey, V., Kriegler, E., Luderer, G., Riahi, K., Ueckerdt, F., Després, J., Drouet, L., Emmerling, J., Frank, S., Fricko, O., Gidden, M., . . . Vishwanathan, S. (2020). Taking stock of national climate policies to evaluate implementation of the Paris Agreement. Nature communications, 11 (1), 2096. Rose, A. (1992). Equity considerations of tradable carbon emission entitlements. In United Nations Conference on Trade and Development) (Ed.), Combating global warming: study on a global system of tradeable carbon emission entitlements. (pp. 55–83). United Nations. Rose, A., & Stevens, B. (1993). The efciency and equity of marketable permits for C02 emissions. Resource and Energy Economics, 15 (117-146). Rose, A., Stevens, B., Edmonds, J., & Wise, M. (1998). International Equity and Difer- entiation in Global Warming Policy. Environmental and Resource Economics, 12, 25–51. Rosenberg, A., Schopp, A., Neuhof, K., & Vasa, A. Impact of Reductions and Exemptions in Energy Taxes and Levies on German Industry (Climate Policy Initiative Berlin Brief, Ed.). Ed. by Climate Policy Initiative Berlin Brief. 2011. Sandmo, A. (2004). Environmental taxation and revenue for development. In A. Atkinson (Ed.), New Sources of Development Finance. Oxford University Press. Sandmo, A. (2007). Global Public Economics: Public Goods and Externalities: mis en ligne le 17 octobre 2007, consulté le 30 septembre 2016. Economie publique / Public economics [En ligne], 18-19.
27 1. Introduction
Shiell, L. (2003). Equity and efciency in international markets for pollution permits. Journal of Environmental Economics and Management, 46 (1), 38–51. Skjærseth, J. B., & Wettestad, J. (2010). Making the EU Emissions Trading System: The European Commission as an entrepreneurial epistemic leader. Global Environmental Change, 20 (2), 314–321. Stavins, R. N. (1997). Policy Instruments for Climate Change: How Can National Governments Address a Global Problem? University of Chicago Legal Forum, Article 10, 1997. Talus, K. (2013). EU Energy Law and Policy. A Critical Account. Oxford University Press. Umweltbundesamt. Klimaschutzziele Deutschlands. 2019. UN Economic and Social Council. World Economic and Social Survey 2016: Climate change resilience - an opportunity for reducing inequalities: E/2016/50. https: //wess.un.org/wp-content/uploads/2016/06/WESS_2016_Report.pdf. 2016. UNFCCC. Adoption of the Paris Agreement - Paris Agreement text English. 2016. UNFCCC. (1992). United Nations Framework Convention on Climate Change. Vrontisi, Z., Luderer, G., Saveyn, B., Keramidas, K., Lara, A. R., Baumstark, L., Bertram, C., de Boer, H. S., Drouet, L., Fragkiadakis, K., Fricko, O., Fujimori, S., Guivarch, C., Kitous, A., Krey, V., Kriegler, E., Broin, E. Ó., Paroussos, L., & van Vuuren, D. (2018). Enhancing global climate policy ambition towards a 1.5◦C stabilization: A short-term multi-model assessment. Environmental Research Letters, 13 (4), 044039. Warleigh, A. (2004). European Union: The basics. Routledge. https://doi.org/10.4324/ 9780203331620 Weitzman, M. L. (2014). Can Negotiating a Uniform Carbon Price Help to Internalize the Global Warming Externality? Journal of the Association of Environmental and Resource Economists, 1 (1/2), 29–49. Wiener, J. B. (2007). Think Globally, Act Globally: The Limits of Local Climate Policies. Univ PA Law Rev, (155), 1961–1979. Williams, R. [Roberton3Will]# We need to pay more attention to distributional efects. Economists far too often call the most efcient policy ”best”, even when it has problematic distributional efects. And distribution is a key element of political feasibility. [Tweet]. https : / / twitter . com / Roberton3Will / status / 1081957953907974144. Twitter. Date 2020-01-06, 2019. Williams, R. (2012). Growing state–federal conficts in environmental policy: The role of market-based regulation. Journal of Public Economics, 96 (11-12), 1092–1099.
28 2 Agreeing on an EU ETS Price Floor to Foster Solidarity, Subsidiarity and Efciency in the EU
Ottmar Edenhofer Christina Roolfs Beatriz Gaitan Paul Nahmmacher Christian Flachsland
Published in MIT press as Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland (2017): Agreeing on an EU ETS price foor to foster solidarity, subsidiarity and efciency in the EU. Published in Parry, Pittel, Vollebergh (Eds.): Energy tax and regulatory policy in Europe. Reform priorities. Cambridge MA: MIT press (CESifo seminar series). Published version. DOI: 10.7551/mitpress/10988.001.0001.
29 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only.
Agreeing on an EU ETS Price Floor to Foster 2 Solidarity, Subsidiarity, and Efficiency in the EU1
Ottmar Edenhofer, Christina Roolfs, Beatriz Gaitan, Paul Nahmmacher, and Christian Flachsland
Key Points for Policymakers
• The EU Emissions Trading System (EU ETS) has provided neither credible incentives for long-term investments in low-carbon tech- nologies nor strong near-term mitigation incentives. • Low EU ETS allowance (EUA) prices and the heterogeneity of the EU Member States (MS) have led to a patchwork of national climate policy across MS, with variable and unequal policy stringency. • Under the current EU ETS design these national policies do not achieve additional emission reductions within ETS sectors. Instead, they reduce the EU ETS carbon price and reallocate carbon emissions to MS with weaker national climate policies. • A price floor for EUAs combined with appropriate transfers (the redistribution of EU ETS revenues) allows for the heterogeneity of MS within the multilevel policy structure of the European Union to be addressed. • While the economic literature suggests using optimal transfers across MS to achieve efficiency when a quantity (ETS) or price instrument is employed, the implementation of optimal transfers may not be fea- sible. Nevertheless, there are other transfer schemes that can improve upon the EU’s solidarity and subsidiarity—two well-established EU normative design principles—and the EU ETS’ economic efficiency. • A numerical exercise is provided to quantify the cost effects of the EU ETS price floor proposal within the European power sector.
1. Introduction
Taking into account the heterogeneity of EU Member States (MS), this chapter proposes an EU Emissions Trading System (ETS) price floor as a key element of an EU ETS reform. It links economic efficiency to the
30 2.1 Introduction
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 32 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Figure 2.1 Evolution of EUA price (solid line) and EUA future contracts for the year 2020 (dotted line) Sources: Data for EUA prices for the year 2008 are taken from ECX EUA Futures, Con- tinuous Contract #2, ICE (Quandl 2015b). EUA prices from 2009 to 2015 are based on the settlement prices at the secondary market, EEX (2015). Future contract prices for the year 2020 are taken from the settlement prices in December 2020, ICE (Quandl 2015a).
EU’s principles of solidarity and subsidiarity, and illustrates the cush- ioning effect of an ETS price floor on intra-ETS leakage. While the price floor’s stabilization effect is also identified in chapter 1, this chapter provides an analysis of the role of fiscal transfers to enhance the MS’ agreeability of introducing a price floor. The EU Emissions Trading System (EU ETS) has not yet provided credible incentives for long-term investments in low-carbon technolo- gies. Its credibility has suffered since the year 2008 because the emis- sion cap has been consistently above the EU ETS sectors’ carbon emissions for which the financial crises that started in the second half of the 2000s are blamed for. The subsequent decline in EU ETS allow- ance (EUA) prices from mid-2008 onward, as depicted in figure 2.1, triggered an ongoing and remarkable debate about reforming the EU ETS. EU policymakers attempted to fix the EU ETS by implement- ing a back-loading provision2 and the market stability reserve (MSR).3 Both measures focus on shortsighted fixes of the carbon price decline: they temporarily remove EUA surpluses from the market. However, the EUAs that were temporarily removed will be returned to the market at some point in the future, leaving the cumulative cap (the aggregate supply of permits) unchanged. It was also decided to increase the Linear Reduction Factor—reflecting the annual reduction of the cap— from 1.74 percent to 2.2 percent per year, thereby reducing the cumula- tive EU ETS cap. This has not had a major effect on the EUA price. While there is no clear consensus about the core problems of the EU ETS and the best response options to effectively address them, it is likely that back-loading and the MSR will be insufficient (Knopf et al. 2014).
31 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 33
The debate on structural EU ETS measures launched by the European Commission (EC 2012) is still unresolved and opens a window of opportunity for a debate on long-term reform. In addition to concerns about a lack of the EU ETS’ credibility, fun- damental questions about the coordination of regulatory authorities have been raised. In particular, there is a lively debate about whether MS’ climate polices undermine the cost effectiveness of the EU ETS (IPCC 2014). EU MS have implemented a diverse national climate and energy policies4 with varying stringencies, which affect carbon emis- sions. To illustrate these different stringencies, we derive an aggregate effective carbon price using the OECD’s estimated effective carbon prices for Denmark, France, Germany, and the United Kingdom for different sectors.5 Based on climate and energy policies, the OECD’s study estimates the net social cost paid for each unit of emissions abated for various sectors and countries.6 We weight the OECD’s esti- mated sectoral prices according to the given sector’s share of aggregate emissions in each country. Data on sectoral emission shares is taken from the European Environment Agency (EEA 2013). The result is depicted in figure 2.2. For the sectors and countries we consider, € Germany has the highest effective carbon price of 53 /tCO2e, followed by Denmark, the United Kingdom, and France, with respective effec- € tive carbon prices of 46, 42, and 25 /tCO2e. The variety of instruments implemented across MS and the effective carbon prices presented here indicate that MS prioritize emission mitigation objectives differently and prefer distinct means to pursue those objectives.7 The heterogeneity of the EU MS8 has led to different willingness to pay (WTP) for abatement as reflected in national climate policies. If the different WTP for abatement had been anticipated and taken into account in the design of the EU ETS, current national polices would not have a weakening effect on the EU ETS. Given the evolution of MS policy choices and lessons learned on the interaction among the EU ETS and national policies, it is pressing to revisit and discuss fundamental EU ETS design features. We base our analysis on two normative design principles that are well established within the European Union—the principle of solidar- ity9 and the principle of subsidiarity.10 We use well-known theoretical arguments to conclude that the current EU ETS does not satisfy these principles. We point to solutions provided by economic theory that would help to make the EU ETS more compatible with these principles. We claim that a price floor for EUAs combined with appropriate trans- fers enables the heterogeneity of MS within the multilevel policy
32 2.2 The Main Shortcomings of the Current EU ETS Price Signal
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 34 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Figure 2.2 Estimated effective carbon prices by country derived from the electricity, transport, pulp and paper, and cement sectors Source: Calculation based on a study by the OECD (2013) and weighted according to the sectoral share of emissions data from the year 2010 (EEA 2015).
structure of the European Union to be accounted for, while also allow- ing the principles of subsidiarity and solidarity to prevail. A first-best outcome and an optimal policy design will not be implemented by self-interested MS. It is therefore used as the socially optimal bench- mark—a normative focal point—for our analysis. However, an EU ETS supplemented with a carbon price floor and an appropriate transfer scheme11 is consistent with the self-interests of the MS such that no Member State is harmed and at least some MS are made better off. This policy design approach promises to be a win-win strategy for all MS. The chapter is organized as follows. In section 2, we review previous findings about shortcomings of the current EU ETS price. In section 3, the interaction between heterogeneous MS and EU policies is discussed; in particular the effect on carbon prices. Two normative design prin- ciples and one implementation rule are subsequently suggested, taking into account the second-best reality of EU policy-making. In particular, we argue that the introduction of a carbon price floor is a promising proposal for EU ETS reform. Section 4 illustrates the effects of an EU ETS price floor by means of numerical simulations for the EU power sector. Section 5 states our conclusions.
2. The Main Shortcomings of the Current EU ETS Price Signal
Recent economic literature on climate change favors the use of a carbon tax (price) or hybrid system (a quantity-based instrument with price
33 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 35
stability provisions) over a pure quantity-based instrument such as an ETS without price stability provisions (e.g., Cramton et al. 2015). Those kinds of policy instruments are expected to deliver a more stable price signal and are economically superior to a pure ETS in terms of the ability to avoid price volatility, emission leakage effects, and uncer- tainty about economic costs (Goulder and Schein 2013; Philibert 2009). However, taxation power in the European Union is limited—the col- lection and redistribution of direct taxation is a sovereign right of MS.12 If an EU carbon tax were going to be implemented, it would require a unanimous vote, whereas an ETS requires a two-thirds majority vote (Talus 2013). In practice, this has meant that the quantity-based ETS was the only politically feasible carbon pricing instrument at the time of its inception (Skjærseth and Wettestad 2010; Talus 2013).13 An ETS price floor implemented as an auction reserve price (as done in the Californian ETS) should not be considered a tax in legal terms but builds on the ETS’ political feasibility. Another line of argumentation focuses on the ongoing price decline of the EU ETS which causes specific credibility challenges. Reasons for the EUA price decline are analyzed by Koch et al. (2014; 2016). They find that the global economic recession, renewable support schemes, the inflow of carbon credits, and gas and coal prices can only explain about 10 percent of the price decline in the EU ETS over the 2008–2013 period. They conclude that policy events have a strong influence on EUA price formation and suggest that controversial debates by EU policymakers as well as EU parliament votes—particularly over back- loading—have destabilized the long-term expectations of investors. Not surprisingly, the low price of futures contracts for the year 2020 indicates that traders anticipate only a modest long-term scarcity of emission permits in the market (see figure 2.1). Neither back-loading nor structural reform proposals like the MSR promise to change this expectation, as they only shift the release schedule of a constant cumu- lative EUA budget over time. Elsewhere it has been extensively dis- cussed why a price floor—potentially complemented by a price ceiling, thus yielding a price corridor—at the EU level would help to stabilize price expectations and support long-term credibility (Knopf et al. 2014; Philibert 2006; Wood and Jotzo 2011). Since carbon and energy pricing have a positive impact on clean technology investments (Copenhagen Economics 2010; Eyraud et al. 2011), the low EUA prices from mid-2008 onward together with low prices for future contracts lack the intended incentives for EU-wide clean technology investments.
34 2.2 The Main Shortcomings of the Current EU ETS Price Signal
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 36 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Even in the presence of the EU ETS, MS continue to implement and modify various forms of national energy and climate policies (Strunz et al. 2015; Talus 2013). These additional policies suggest that some MS would rather pursue either a less or more stringent climate policy than is available through the EU ETS. For example, Poland threatened to withdraw from EU climate policy altogether,14 demonstrating its wish for less stringent climate policy. On the other hand, Germany, the United Kingdom, and Sweden have implemented policies demonstrat- ing a wish for more stringent climate policy. Germany implemented the Renewable Energy Act to foster the German energy transition (Ener- giewende). The United Kingdom established the Climate Change Levy, consisting of inter alia a carbon price support rate for EUA. It functions
as a national price floor with a current level of 18£/CO2e (HM Revenue & Customs 2014).15 Sweden established a general carbon tax in 1991, but it made exemptions to some sectors after the EU ETS implementation (OECD 2014).16 Despite these efforts toward more stringent climate policies, national climate policies act to weaken the EUA price as the demand for allowances from MS with more ambitious climate polices decreases (Böhringer et al. 2008). In the following section, we extend the debate about EU ETS reform by considering the heterogeneity of MS. The current EU ETS is sup- posed to equalize marginal abatement costs17 across MS. This can only be efficient without the presence of any unilateral MS policies (Williams 2012). Additionally, the equalization of marginal abatement costs among MS does not account for the federal-like structure between the European Union and its heterogeneous MS, in which EU and MS poli- cies coexist. Instead, climate policy at the EU level could be set in a similar fashion as the EU tax minima for value-added tax in alcohol, tobacco, and energy products. For these, the European Union sets required minimum rates for MS, but they have the flexibility to set higher rates if they wish to for fiscal or other reasons (EP 2014). In the climate context, an EU ETS price floor would not hold back those MS that wish to price emissions more aggressively. The existence of EU minimum rates in other regulatory domains raises the question of whether the European Union’s vertically divided regulatory regime and the MS’ heterogeneity are sufficiently considered in the design of the EU ETS, and if improvements are conceivable. The MS’ heterogeneity has largely been ignored in the EU ETS design, though income heterogeneity is addressed to some extent by certificate allocation provisions.18 As a result, the simultaneous interaction between
35 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 37
the EU ETS and MS climate policies will continue to distort the function- ing of the EU ETS. The policies implemented by individual MS reduce the EUA price and increase the effective national carbon price in the respective MS. Because the cap remains constant, they do not achieve emission reductions. Fixing the EU ETS will be required to ensure an effective, efficient, and ambitious European climate policy. Otherwise, there is a risk that EU climate policy will become further fragmented, ineffective, and costly, and consequently deteriorate over time.
3. Guiding Principles for EU ETS Design with Heterogeneous Member States
In this section we explore the implications of MS’ heterogeneity on the EU ETS price design. For this purpose we consider two types of MS’ heterogeneities. First, MS can differ in income levels. This can stem from differences in factor endowments such as physical capital and human capital, access to fossil resources, and technological differences. In the face of income disparities, the optimal provision of climate change mitigation requires specific transfers (see section 3.1). The use of transfers within the context of climate change has a direct link to the principle of solidarity as described by Hilpold (2015). He relates the EU solidarity principle to the use of transfers to achieve a common goal. Second, MS can be heterogeneous in terms of their preferences for environmental quality and/or how they are affected by climate change. For example, EU countries might expect different effects on their populations from climate change-induced heat waves, droughts, and flooding. Taking these preferences into account plays a fundamen- tal role in the fulfillment of the EU subsidiarity principle (see section 3.2).
3.1 Efficiency, Transfers, and Solidarity Traditional wisdom suggests that, by equalizing the cost increase from reducing a unit of emissions (marginal abatement cost, or MAC) across emission sources, emissions trading always achieves efficiency19 (Coase 1960). However, in the presence of unequal income across countries, a uniform carbon price that equalizes MACs across countries may not be efficient. For example, richer countries may be able to afford more strin- gent national climate policies. The efficiency of MAC equalization across countries was first challenged and refuted by Chichilnisky and Heal (1994), who showed that if a poor country gains more from increases in private consumption than a rich country, the poor country’s WTP20 for
36 2.3 Guiding Principles for EU ETS Design with Heterogeneous Member States
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 38 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
mitigation is lower relative to that of the rich country. For expository reasons, let us consider the case of a poor country in Asia in which a large portion of its population suffers from malnutrition. In that country the gain from increasing private consumption (in particular, food) should be much higher than the gain of a developed country in Europe from increasing private consumption. In such a case, the poor country’s WTP for mitigation is lower and hence it should pay less for emissions mitigation than a developed country in Europe. Despite smaller income gaps among EU MS, similar effects resulting from unequal income levels across countries matter within the European Union. Crucially, Chichilnisky and Heal (1994) demonstrate that an efficient solution to this situation features different MACs across countries. If, however, MACs across countries were to be equalized under an ETS, optimality would require specific transfers from richer to poorer countries.
Optimal Abatement If the optimal transfers are not implemented, Chichilnisky and Heal (1994) show that poorer countries should set lower MACs than richer countries. They find that an efficient outcome is one in which a country’s MAC equals the ratio of the sum of social gains from emission reductions across all countries relative to the social gain from larger private consumption in the respective country. Since countries benefit differently from increasing private consumption, MACs must not necessarily be equalized to achieve efficiency. In such a situation, different carbon prices for each country are an institutional pre-condition for social optimality. However, national carbon prices could lead to a more nationalized and fragmented European climate policy, which would undermine future cooperation within Europe.
Optimality Under an ETS To counteract the fragmentation of Euro- pean climate policy, a uniform carbon price seems preferable to dif- ferentiated national carbon prices. However, a uniform carbon price requires a specific transfer scheme since, without transfers, an ETS imposes MAC equalization across countries but is not efficient. As indicated above, this equalization is not efficient as long as the social gain from increasing private consumption is not equal across countries. Chichilnisky and Heal (1994) point out that optimality within an ETS that employs a uniform carbon price can only be fulfilled by using a transfer scheme that equalizes the social gain from increasing private consumption (SGIPC) across countries. Thus, since poorer countries gain more than richer countries from increasing private consumption,
37 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 39
Box 2.1 Design Principle 1. Efficiency, Transfers, and Solidarity A uniform emissions price at the EU level must be supplemented with appro- priate transfers to ensure economic efficiency. The current EU ETS design in which marginal abatement costs are equalized across all MS is not per se efficient. Efficiency is only obtained if rich MS provide sufficiently large transfers to poorer MS. Such transfers enable higher levels of consumption in poorer MS—hence complying with the EU’s solidarity principle—while significantly increasing the poorer MS’ willingness to pay for mitigation. If the transfers cannot be implemented, rich MS need to have higher marginal abatement costs and therefore abate relatively more than poor MS.
poorer countries must receive transfers leading to the equalization of the SGIPC for all countries. In the face of large income differences across countries, significant transfers must occur. In the current EU ETS, two general types of transfers exist. The first is the redistribution of EU ETS auction revenues to MS. In 2013, 40 percent of all EUA were auctioned for a total auction revenue of about €3.6bn (EC 2015d). Of these revenues, 88 percent were distrib- uted in proportion to historical emissions across MS. Ten percent of the auction revenues were channeled to less wealthy EU MS to promote investments dedicated to carbon intensity reduction and for adaptation to climate change (EC 2013; 2015a; 2015a). The remaining 2 percent (the “Kyoto bonus”) were allocated to nine EU MS that had reduced their emissions by at least 20 percent of their Kyoto Protocol base year or period level by 2005.21 Second, the value of the remaining 60 percent of all EUAs were transferred to firms. If firms are transnational, then it is not clear whether the MS’ population is the full beneficiary of this type of transfer, nor whether the transfer can address differences in wealth as would be needed to equalize the social gains from increased consumption across countries.22 If we consider the EU MS’ per capita gross domestic products as a wealth indicator, we can conclude that differences across MS are rather large. Thus, transfers of 12 percent of the total EUA auction revenues to less wealthy MS are probably insufficient to achieve optimality. In section 3.3, we provide estimates for transfers that would lead to EU ETS optimality (the equalization of the gain from increasing private con- sumption across EU MS) based on Chichilnisky and Heal’s analysis. Box 2.1 summarizes this section and embeds its findings in a design principle.
38 2.3 Guiding Principles for EU ETS Design with Heterogeneous Member States
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 40 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
3.2 National Preferences and Subsidiarity The analysis in the previous section focused on income differences and on one common goal (climate change mitigation). We now address the use of multilevel climate policies driven by heterogeneous MS’ prefer- ences. Heterogeneous preferences can arise due to differing effects from multiple emission externalities, i.e., climate change and air pollution, as well as differing priorities for environmental quality. Even if infor- mation about transboundary and global effects of carbon emissions were perfectly available to regulators at all regulatory levels, MS authorities typically only care for the well-being of national inhabit- ants. By contrast, an overarching regulating layer, such as the European Union, considers the well-being of all inhabitants of all MS, and is better equipped to provide global public goods such as climate change mitigation. In the following section, we focus on the interaction of multilevel regulation for cases in which MS have heterogeneous prefer- ences. Addressing this type of heterogeneity in the context of multilevel policies is important because, as clarified below, it plays a fundamental role in the fulfillment of the EU subsidiarity principle (see also Oates 1972, 2011).23
Heterogeneous Preferences and Strategic Member States A study by Williams (2012) analyzes interactions between government layers in which both sub-level (MS) and top-level (EU) regulating authorities are allowed to regulate emissions simultaneously. He finds that if the top- level regulator implements an ETS, additional MS’ climate policies become either ineffective or may even result in additional costs for the multilevel regulatory system. To attain efficiency with an over-arching ETS, MACs across MS must be equalized and optimal transfers have to be set. However, if MS implement additional policies, MACs can differ. Williams shows that within an over-arching ETS there is no transfer from the top-level regulator to the MS that can achieve effi- ciency as long as MS policies are present. Instead, since the ETS cap is fixed and the ETS price adjusts as MS unilaterally cut emissions leading to increased emissions in other MS (intra-ETS leakage), the top-level ETS cancels out all unilateral abatement efforts.24 Williams also finds that a carbon tax implemented by the top-level regulator is superior to an ETS. This occurs because the top-level and sub-level prices are addi- tive, while quantity instruments are not as the stricter cap is always binding.25
39 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 41
Box 2.2 Design Principle 2. Member States’ Preferences and Subsidiarity Implementing an emissions’ price instrument at the EU level—either by an EU ETS price floor or an EU carbon tax—is consistent with the principle of subsidiarity. With a price-based instrument at the EU level, national policies can prosper as companion policies of the EU ETS. It allows effectively accounting for the MS’ heterogeneous preferences without undermining the EU policy. In contrast, purely quantity-based instruments at the EU level— such as the current EU ETS—would not only make it harder for ambitious MS to become frontrunners with respect to climate policy but would even render their national efforts fruitless.
More specifically, it would be preferable for the European Union to implement an EU-wide carbon tax to address emission leakage effects among MS, and for MS to set national taxes for regulating local emis- sions externalities and/or local preferences. Reflecting Chichilnisky and Heal’s (1994) findings, the multilevel system eventually achieves optimality if the carbon taxes at the two different levels are supple- mented by optimal transfers. Based on Williams’s argument, a uniform EU carbon price combined with MS’ carbon prices and optimal trans- fers can lead to an efficient outcome. On the contrary, the use of a pure ETS—as opposed to one with a binding price floor—precludes the achievement of an efficient outcome, because MS cannot be prevented from implementing national climate polices. In a similar line of research, Roolfs et al. (2016) find that a carbon price set by the top-level regulator in addition to MS’ policies can approximate the first-best outcome, if the top-level regulator employs optimal transfers. They analyze the potential of a top-level regulator to set a union-wide carbon price in coexistence with strategic MS policies while the top-level regulator anticipates how MS’ carbon prices react to the top level’s carbon price. If non-optimal transfers are available, they identify the price floor level that at least comes closer to the first- best outcome while making all MS better off. Box 2.2 summarizes this section and embeds its findings in a design principle. Design Principle 1 and 2 consider different bases for the heterogene- ity of MS (income levels and preferences) but lead to a common result: in a multilayered policy regime, a price instrument implemented at the top level more efficiently allows for heterogeneity to be addressed as long as optimal transfers are employed. However, optimal transfers derived from economic theory are often unviable for policymakers. In the next section, we propose a pragmatic rule that does not achieve the
40 2.3 Guiding Principles for EU ETS Design with Heterogeneous Member States
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 42 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
first-best outcome, but can find consensus across MS such that some MS are better off, while also ensuring that other MS’ well-being remains at least at their original level (Pareto-improvements).
3.3 Institutional Design in a Non-Optimal World The aforementioned design principles are normative focal points derived within an economic, theoretical setting. However, the optimal imple- mentation of both principles may prove difficult in the real world. This may be due to enforcement constraints, to the difficulty or impossibility of overcoming the free-rider behavior of self-interested actors, and/ or—as will be discussed next—to the political infeasibility of the trans- fers that would be necessary to lead to the optimality of an EU ETS. Building on Chichilnisky and Heal’s (1994) findings, we derive the optimal transfers needed to make the EU ETS efficient. To do so, we assume that (a) each country’s well-being is influenced similarly by private consumption, and (b) an upper-level regulator such as the Euro- pean Union weighs all countries equally.26 Given these assumptions, equalizing the social gain from increasing private consumption across countries requires that all countries have an equal level of private con- sumption. To estimate the transfers needed to equalize consumption levels across EU MS, we use private consumption expenditure data (WDI 2015) in purchasing power parity U.S. dollars (PPP$) for the year 2010. Our objective is to find transfers that enable the EU’s population the same level of per capita consumption while making aggregate consumption equal to observed data. The transfer per person in each Member State is the gap between the EU’s and each Member State’s per capita consump- tion levels. Based on consumption data in the year 2010, figure 2.3 shows optimal per capita transfers (per person) across all EU MS that would equalize EU per capita consumption. A negative number indicates that a respective country is not a receiver but it is instead a donor. The popula- tion of Luxembourg, as the richest in the European Union, would be the largest donor (with a negative transfer, a net payment of PPP$7,819 per person). Luxembourg’s population is followed by the populations of Austria, the United Kingdom, and Germany, with respective negative per capita transfers of PPP$4,370, PPP$3,470 and PPP$3,101. The popula- tions receiving the largest transfers would be those living in Bulgaria, Latvia, Estonia, and Romania. The estimated optimal transfers serve to demonstrate the magnitude of the difference in consumption levels across MS. The difference in consumption levels has a large implication for the individual WTP for climate change mitigation.
41 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 43
Figure 2.3 Optimal transfers per person according to private consumption in the year 2010 in thou- sands of dollars of purchasing power parity (PPP$)
Figure 2.4 indicates the aggregate optimal transfers per Member State (the per capita transfers multiplied by the population of each Member State). Transfers of the size depicted in figure 2.4 are very unlikely to be politically feasible. At the same time, current EU ETS transfers equal to 12 percent of the revenues from the EUA auction (0.432bn€ in 2013) seem insufficient. Since a theoretical, first-best outcome of a pure ETS with optimal transfers is very likely to be politically infeasible, we propose the con- sideration of second-best options.27 One such case that is particularly useful is a second-best world in which changes in the EU ETS design make a Member State better off, while also ensuring that other MS’ well-being remains at least at their original level (this is compared to a case in which only MS implement climate policies in a decentralized, uncoordinated setting). From a welfare perspective, this ensures that the joint implementation of climate policies creates winners, while also guaranteeing that there are no losers. In contrast to the normative framework described in sections 3.1 and 3.2, we now consider results from a study that analyzes a setting in which: (a) optimal transfers are not viable; (b) a multilevel policy regime is already established; and (c) a uniform price signal is set at the top level and each Member State sets its own carbon price. This starting point is more similar to the current EU ETS in which the EUA market intends to deliver a uniform price signal to all MS, while MS set additional
42 2.3 Guiding Principles for EU ETS Design with Heterogeneous Member States
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 44 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Figure 2.4 Optimal transfers per country according to private consumption in the year 2010 in bil- lions of dollars of purchasing power parity (PPP$)
climate policies and transfers are given as discussed in section 2. In a comparable setup and with heterogeneous MS, Roolfs et al. (2016) iden- tify MS’ carbon prices and a range of top-level uniform carbon prices, including a price floor level, combined with simple transfers28 that make all countries better off compared to a decentralized setting. When income heterogeneity is considered, Roolfs et al. find that equity-based transfers can put a disproportionate cost burden on the richest Member State. The richest Member State agrees to bear the cost burden of the top-level policy as long as its gain outweighs its costs. Based on the nature of the equity-based transfers, poorer MS carry no burden but benefit by internalizing the emission externalities and by a net income gain. Therefore, the tipping point for the feasibility of top- level policy becomes the consent of the richest Member State and is represented by a carbon price floor. Here, the carbon price floor is the carbon price level that leads to the highest well-being of the richest Member State. As long as the top-level regulator considers the carbon price floor, the top-level policy is compatible with the self-interest of all MS in the sense that all MS are better off. Within the context of their model, Roolfs et al. (2016) show that the price floor based on the richest Member State’s utility and in combina- tion with equity-based transfers works as long as the wealth gap among poor and rich MS is not extreme. Since the price floor ensures that no
43 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 45
Box 2.3 Implementation Rule. Set a Price Floor and Provide Appropriate Transfers A carbon price floor can help to address the challenges associated with the heterogeneity of MS while accepting a non-optimal world. With an EU ETS price floor, transfers must not necessarily be optimal to lead to welfare improvements for all MS.
Member State falls below the welfare level of the decentralized outcome, it also satisfies the principle of subsidiarity. Analogously, Roolfs et al. find that a price floor works for transfers based on the MS’ historical emission levels. However, this transfer scheme—in contrast to equity transfers—can make all MS better off, and it does not always impose restrictions on wealth gaps. They conduct a similar analysis on hetero- geneous preferences for emissions’ externalities on MS’ well-being, and find similar results. In box 2.3 we propose an implementation rule based on this section’s findings by paying tribute to a non-optimal world.
4. Illustration of the Effects of an EU ETS Price Floor
In this section, we provide a twofold sketch demonstrating that national climate policies will not undermine the efficiency of the EU ETS when a price floor is implemented (see also IPCC 2014; Goulder and Stavins 2011). We first give an illustrative description of the cushioning effect of a price floor. We then present results from the European power sector model LIMES-EU.
4.1 Cushioning Intra-ETS Leakage with a Carbon Price Floor Consider a multinational ETS without MS policies. The ETS allowance
price (pETS) is determined endogenously by the ETS market, such that MACs are equalized across all participants. Thus, the MS’ emission
levels (Ei , Ej ) are determined by pETS (see figure 2.5). If Member State i * prefers a lower national emission level Ei than the level that results
from the ETS alone, its WTP for mitigation is above pETS. In order to * τ obtain Ei , Member State i sets an additional national policy , which
results in an effective national carbon price of pMS, such that pMS = τ pETS + . As soon as τ is implemented, Member State i’s firms reduce their
demand for allowances subject to pMS and the ETS allowance price falls
from pETS to p’ETS. If Member State i wants to ensure that its preferred * emission level (Ei ) is reached, it can do so by adjusting its national
44 2.4 Illustration of the Efects of an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 46 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Figure 2.5
Illustration of the cushioning effect of intra-ETS leakage with a carbon price floor (pMIN)
policy (τ′) by means of a so-called variable fee.29, 30 However, Member State j, which has implemented no additional national policy, also faces ′ a decrease on the ETS price (from pETS to pETS). This results in an
increase in Member State j’s emission level from Ej to Ej’. In effect, since the ETS cap is set exogenously, the additional national policy of Member State i has no effect on overall emissions as the emission allowances are used by Member State j’s emitters (100 percent emission leakage).31 From a multinational perspective, the national policy can be considered a disturbance. From a national perspective, the national policy may have beneficial side-effects (e.g., increased national revenues, reduction of local air pollution). However, it fails to reach the goal of total emis- sion reduction due to the intra-ETS leakage effect. The problem of the ETS price decline and the ineffectiveness of national policies can be cushioned by the implementation of an ETS-
wide price floor, pMIN (refer to figure 2.5, in which MS i now implements the variable policy τ″). Since the ETS price decrease is cushioned,
Member State j faces pMIN, which is lower than the initial ETS price (pETS) ′ but above p ETS. Therefore, Member State j implicitly benefits from Member State i’s policy due to the price decrease. However, the emis- sion leakage effect triggered by the national policy disturbance in the
ETS is weakened. The effective emission level of Member State j (Ej’’)
is above the initial emission level (Ej), but below the emission level
without a price floor (Ej’). In the end, Member State i and j are both better off.
45 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 47
4.2 Implications for the European Power Sector: Numerical Simulation In this section, our theoretical analysis is complemented with quantita- tive results from the long-term investment model for the electricity system of Europe (LIMES-EU). The model is a multi-country model32 that simultaneously determines cost-minimizing investment and dis- patch decisions for generation and storage and transmission technolo- gies needed to serve future electricity demand and comply with future energy and climate policies. Its integrated approach, together with an intertemporal optimization in five-year steps from today until 2050, allows for the analysis of consistent and cost-efficient pathways for the future development of the European power system on both aggregate and national levels. The optimal deployment of different electricity generation options strongly depends on future climate and energy policies at EU and national levels. We illustrate the effect of a European price floor plus additional emission reduction efforts in Germany. This is motivated by the current German discussion about how to reach national 2020 climate targets using additional unilateral policies (see, e.g., BMWi 2015). Our analysis is focused on the time span 2015–2030, assuming a common European carbon price from 2030 onward. In the present model frame- work, a price floor on carbon emissions leads to additional costs for the energy system. The revenues and redistribution (transfers) from carbon pricing are not considered in our numerical exercise. Table 2.1 provides an overview of the policy scenarios analyzed. Three different European carbon price floors until 2030 are considered. In the baseline scenarios, the carbon price in Germany is equal to the level of an EU carbon price floor. For the policy scenarios, we imple- ment a variable fee in Germany that raises the effective German carbon € price to 20 /tCO2, a price that is in line with the long-term EU decar- bonization targets (EC 2011; Knopf et al. 2014).33 In all scenarios, aggre- gate European emissions are constrained to be less than or equal to the emission budget that results from the different EU carbon prices (5, 10, € and 15 /tCO2) without an additional Germany policy. From 2030 onward, we consider four scenarios with carbon prices of 20, 25, 30, € and 35 /tCO2 that are effective for all European countries. In order to reflect the energy policies currently in place, the nuclear phase-out in Belgium, Germany, and Switzerland as well as the German renewable energy expansion target are taken into account. Nuclear power investments in other countries are constrained to the expansions
46 2.4 Illustration of the Efects of an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 48 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Table 2.1
Policy Scenarios (all prices in euros per tCO2)
Until 2030 Europe Carbon price floor of €5/€10/€15
Germany No additional policy* or effective carbon price of €20
After 2030 Europe Common European carbon price of €20/€25/€30/ €35 in 2030, subsequently rising by 5 percent per year until 2050
Germany No additional policy*
* If no additional policy is set, the German carbon price is equal to the European carbon price.
already under construction or planned and to the investments needed to replace depreciating capacities. As the future of carbon capture and storage (CCS) is highly uncertain, our policy scenarios are run both with and without the possibility of CCS investments. In total, this leads to 24 baseline scenarios without and 24 policy scenarios with an addi- tional emission policy in Germany. Figure 2.6 summarizes the effects of such an additional policy on carbon emissions in Germany and in Europe on the whole. The results show that an elevated German carbon price reduces German emissions in all policy scenarios (figure 2.6a). This is mostly due to an overall reduction in German electricity production (figure 2.6b). Replacing the missing domestic supply with electricity imports from neighboring countries results in an increase of emissions abroad. In most cases, however, the emission reductions in Germany outweigh the emission increases abroad, leading to an overall reduc- tion of emissions across Europe—implying that it is the EU-wide price floor and not the cap that becomes binding. Figure 2.7 illustrates this effect in the year 2020 for a scenario with a common European carbon € price of 30 /tCO2 in 2030 and different price floors in the years prior to 2030. In these scenarios, the European-wide emission reductions
vary between 0 and 0.67tCO2 per ton of emission reduction in Germany. The reductions in German electricity production when there is an EU € price floor of 5 or 10 /tCO2 result from reductions in the use of lignite and hard coal, while the increase in neighboring countries is predomi- nantly based on natural gas and renewables. The lower emission inten- sity of newly installed foreign power plants reduces the total European
47 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 49 price in Germany 2 emissions and (b) electricity production due to a higher CO emissions and (b) electricity production 2 Note: The gray bar indicates the range over all scenarios and outlined box median scenarios. Figure 2.6 Change in (a) annual CO
48 2.4 Illustration of the Efects of an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 50 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Figure 2.7 Change in annual electricity production (a) in European countries except Germany and
(b) in Germany in the year 2020 due to a higher CO2 price in Germany Notes: Although the use of CCS is possible, it is not yet deployed in 2020 in these sce- narios. The annual electricity production of nuclear power plants does not change in any of the three scenarios.
€ emissions. When there is a 15 /tCO2 European price floor, the German € carbon price of 20 /tCO2 is not high enough to induce a considerable
change in the electricity production pattern, nor in CO2 emissions. Overall, the results suggest that an EU-wide carbon price floor allows for the introduction of more ambitious national carbon prices with a net reduction effect on overall emissions. In our numerical model framework, such additional efforts increase total system costs. This increase in costs34 depends heavily on the level of the European € € price floor. It varies between 12bn (in the case of a 15 /tCO2 price € € floor) and 36bn (in the case of a 5 /tCO2 price floor). For a price floor € of 10 /tCO2, costs incurred by the additional German climate policy are around 24bn€. Other scenario variations, e.g., the level of the common carbon price after 2030, have only a very limited effect on overall costs (i.e., +/–1bn€), with lower cost differences for higher future carbon prices. Our analysis for the German case can only serve as an illustration. Additional analyses focusing on other countries are needed and should be an interesting subject for further research.
4.3 Some Implementation Issues There are several challenges that go beyond the scope of this chapter that are associated with the implementation and operation of a price floor in terms of the detailed design. An extensive analysis can be found in Wood and Jotzo (2011). For the operational implementation of a price floor, they suggest an auction reserve price (which is implemented in the California and Regional Greenhouse Gas Initiative ETSs), a variable fee (as implemented in the United Kingdom), or the buy-back of allow- ances by the regulating authority. To avoid excess allowances being sold at a price below the price floor, the regulator should be willing to buy
49 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 51
back and cancel excess allowances at the level of the price floor, as pointed out by Goulder and Schein (2013). This would imply additional costs for the regulator. Another option to avoid buy-back necessities would be to use a price instrument only (no ETS) (Goulder and Schein 2013). For a detailed discussion on cancelling allowances, see, e.g., Kollmuss and Lazarus (2010). In terms of the newly proposed MSR, a price floor could also be used as a signal indicating when allowances should be withdrawn—i.e., when the ETS price floor is binding. Some analysts might argue that an EU ETS price floor is unnecessary. Another decentralized alternative for MS with higher WTP for mitiga- tion reduction is the unilateral purchase and retirement of allowances (P&RA). This procedure could be carried out in the current legal EU structure as EU responsibilities and the ETS design would remain unchanged. To express a significantly higher WTP for mitigation, MS (such as Germany or the United Kingdom) would need to purchase and retire significant amounts of EUAs. If huge amounts of EUAs were withdrawn and retired, the first effect that would be observed on the ETS market is an EU ETS price increase, due to a reduction in the total EUA effectively available. This procedure could result in dissent from other MS as they may face a higher ETS price that might not be compat- ible with their (comparably lower) WTP for mitigation. Therefore, com- pensation or side payments may become necessary. This brings back the question of transfer design and the role of transfer coordination. The role of coordination could be effectively carried out at the EU level if an EU ETS price floor and appropriate transfers are set. There are consequently two drawbacks to the P&RA. First, national funds need to be used as side payments to generate agreeability among MS on the use of P&RA. Second, the Member State that purchases and retires EUAs must use national revenues that could otherwise be used in other programs, including climate programs (Bianco et al. 2009), thereby causing budgetary disturbances. Given current fiscal pressures, it seems unlikely that a country would do this. A case in point is the United Kingdom, whose government is under intense political pressure to moderate fiscal austerity.35 Therefore, it seems politically infeasible to divert revenues from the national budget to purchase EUAs, which would effectively divert revenues from the UK Treasury to allowance holders in other EU MS. In order to avoid the use of governmental revenues, a country could force companies to retire allowances as Germany attempted through the implementation of a “climate levy” (BMWi 2015). This might not affect the national budget but it hurts
50 2.4 Illustration of the Efects of an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 52 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
some MS due to a higher EU ETS price and causes the same problems as discussed above. From an individual government’s perspective, there are additional potential advantages of a price floor. A price floor not only ensures a more stable price signal for market-participants, but also more stable revenue flows for MS and EU ETS funds. If the carbon price floor is binding in the longer term, it can be a substitute for income and cor- porate taxation, ameliorating the effect on government revenues derived from tax competition (Heinemann et al. 2009) and counteract- ing distortionary effects of taxation. It can therefore be used as a more efficient source of public finance (Edenhofer et al. 2015). Sweden’s environmental national policies exemplify the successful implementa- tion of a carbon price36 and the shift of the fiscal burden from labor to carbon emissions (OECD 2014). Parry et al. (2014) present an extensive analysis of multiple incentives—besides climate change mitigation— for countries to put a price on carbon emissions, subsequently extended to a discussion of climate regime design based on co-benefits by Eden- hofer et al. (2015). Cramton et al. (2015) highlight that the commitment to a uniform multinational carbon price is less risky for individual countries than the commitment to a quantity instrument. They argue that future business-as-usual emissions and abatement costs are both highly uncertain. Due to these uncertainties, the financial risk for coun- tries agreeing on quantity commitments becomes much larger than commitments to a price. A carbon price floor can also entail benefits for the operation of the EU ETS market. Burtraw (2014), for example, emphasizes that a carbon price floor is a non-discretionary and transparent signal by policymak- ers about the level of climate policy ambition, which allows market participants to better anticipate future developments. Burtraw also points out that a price floor may be used as a signal for cap adjust- ments. The more often the price floor binds, the higher the likelihood that the cap was set too loose. As a result, the regulator would need to withdraw and retire allowances or tighten the future cap schedule more often. Therefore, a price floor explicitly and transparently addresses the objective implicitly intended by the MSR—stabilizing the ETS market. More importantly, if credibility about an ETS is lacking, as is the case in the EU ETS, the use of a carbon price floor would increase policy credibility as an additional signal and commitment by policymakers to a certain level of policy ambition.
51 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 53
5. Concluding Remarks
This chapter proposes a price floor, in combination with appropriate— not necessarily socially optimal—transfers, as a key reform option for the EU ETS. A carbon price floor has additional advantages to stabiliza- tion effects. One such advantage that is often overlooked is the ability to address heterogeneity and the policies of MS in vertical regulatory structures like the European Union. There may be gains from multinational climate policy when there is a multilayered governmental structure such as in the European Union. A pure ETS without optimal transfers cannot correctly accommodate MS’ heterogeneity as it neglects differences in income and preferences on carbon emissions. The multilayered structure, however, facilitates a solution. An EU ETS carbon price floor, combined with appropriate transfers, can enable MS to implement national climate policies that are indeed effective. This does not necessarily imply additional changes in EU legislation in terms of an EU revenue system. For example, if allow- ances are auctioned at a price floor, MS can remain in charge of revenue collection. Transfer payments could be coordinated at the EU level while actual payments could be made bilaterally. To conclude, this chapter identifies guiding design principles to reform the EU ETS and depicts why and how the heterogeneity of the MS should be considered. We point out efficiency shortcomings in the EU ETS design particularly in light of the heterogeneity of the MS. We connect economic theory to the solidarity principle and discuss why the traditional EU ETS as a pure cap-and-trade system—in which MACs are equalized—is not efficient per se. A higher WTP by MS and subsequent transfers from richer to poorer MS should play a role. To achieve social optimality, significant explicit transfers from richer to poorer MS would need to be deployed. In a model that departs from an EU ETS, differentiated carbon prices could be implemented in each MS, which would implicitly function as transfers. However, different national carbon prices could lead to a fragmentation of the EU climate policy damaging the cooperation among MS. When embedded in a multilevel governmental system, a quantity- based instrument (such as an ETS) at the upper governmental layer leads to inefficient outcomes if it coexists with MS’ policies. Hence, the current EU ETS is inconsistent with the principle of subsidiarity. The MS’ policies distort the long-term optimal carbon price within the EU ETS. Therefore,
52 References
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 54 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
we have suggested that an EU-wide price floor combined with optimal transfers for the EU ETS could approximate efficiency. To cope with the MS’ heterogeneity if neither optimal prices nor optimal transfers are attainable, we propose an EU ETS price floor as a useful tool. Besides the price stabilization effect, an EU ETS price floor allows willing MS to implement national climate policies without decreasing the dynamic efficiency of the EU ETS. It allows MS’ policies to be integrated without undermining EU ETS-wide emission reduc- tions. Our numerical simulations of the European power sector indicate that an EU ETS price floor ameliorates leakage within MS and can
achieve additional EU-wide emission reductions of up to 50MtCO2, if a Member State, such as Germany, also implements a carbon price. Regardless of the benefits of carbon pricing, the window of opportu- nity for a debate on long-term reform has more far-reaching implica- tions. The European Union can be considered a laboratory for multilateralism and lessons can be learned for implementing global climate policies (Goulder and Stavins 2011; Grubb et al. 2014). If the European Union succeeds with its EU ETS reform, it may prove wrong the accusations of “blame-and-burden” and instead shift attention toward the design of a common climate policy with mutual gains (Grubb et al. 2014). A failure of the EU ETS may send a negative signal about the plausibility of multinational cooperation to non-EU countries trying to implement an ETS. The debate about EU climate policy and the EU ETS reform also interacts with the international climate policy process beyond the twenty-first Conference of the Parties (COP 21) in Paris. At the COP 21 the European Union and its MS committed to a 40 percent EU-wide GHG emission reduction by the year 2030 compared to 1990 levels (UNFCCC 2015). According to the EC’s Impact Assessment document of alternative EU ETS reform options, these EU-wide GHG emission € reductions would require EUA prices of 40 /tCO2e in the year 2030, € and 264 /tCO2e in 2050 (EC 2014). The agreement of the COP 21 feeds back into the need to reform the EU ETS. An improved coordination between EU and MS’ climate poli- cies is required to meet the EU’s and MS’ pledge of EU-wide GHG emission reduction. MS that currently seek to phase out coal-fired power plants, e.g., by unilateral carbon pricing schemes, do not achieve any additional emission reductions beyond the EU ETS cap (pricing schemes are already implemented in the United Kingdom and under consideration in Germany). Since coal phase-outs are pressing mea-
53 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 55
sures to meet the EU’s reduction target, unilateral MS’ initiatives should be empowered to accompany EU policy. Currently, MS policies function in the opposite direction. They weaken the EU ETS by lowering EUA prices. Low EUA prices jeopardize the achievement of the EU’s GHG emission reduction target. As this chapter describes, an EU ETS reform that implements a price floor would allow the policies of ambitious MS to prosper as companion policies to the EU ETS and contribute to the EU-reduction target instead of weakening the EU ETS. Socially optimal targets and policies may not be attainable, but as this chapter points out, coordination around an EU ETS price floor and appropriate transfers could at least enable policy reforms that support consensus across MS and increase the level of success of climate policy ambitions. These findings, just as they apply to a regional ETS, can also apply to international instruments for climate change mitigation.
Notes
1. We thankfully acknowledge helpful comments on a preliminary version of this chapter from Ian Parry, Karen Pittel, Michael Pahle, Kristin Seyboth, and Eva Schmid. Any remaining errors and flaws remain in the authors’ responsibility. Financial assistance by the German Federal Ministry of Education and Research under grant agreement no. FKZ 03EK3523B (de.zentral-project) and by the European Union’s Seventh Frame- work Program under grant agreement no. 308481 (ENTRACTE-project) is gratefully acknowledged. The conclusions expressed here do not necessarily represent the views of the above-mentioned institutions. 2. Auctioning of 900 million EUAs was postponed from the years 2014–2016 to 2019–2020. 3. The MSR mechanism withdraws EUAs from auctioning when a certain upper thresh- old of unused EUAs in circulation (allowance surplus) is exceeded and feeds these into the “market stability reserve.” Once a lower threshold is triggered, these EUAs are re- released from the MSR. 4. For example, the UK’s climate change levy, the German Renewable Energy Act (which includes subsidies to renewable energy production) and eco-tax, the Danish and Swedish fuel and carbon taxes, and a variety of funds for energy efficiency measures in various MS (Landis et al. 2012). 5. The sectors considered by the OECD study are electricity generation, road transport, pulp and paper, cement, and households’ domestic energy use.
6. Note that the OECD’s estimated effective carbon prices are based on specific calcula- tions for different policies. It implies neither that all instruments considered (such as carbon taxes and feed-in tariffs) function in the same way, nor that they have the same effect on emission mitigation. 7. We stress that the estimates based on the OECD (2013) serve for illustrative purpose only. There are different methodologies available to calculate implicit carbon prices. A
54 References
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 56 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
discussion of alternative estimates of effective carbon prices can be found in OECD (2013). 8. The heterogeneity of the EU MS can be in terms of, e.g., economic development, environmental objectives, dependency on domestic polluting fuels, and concerns about vulnerability related to the import of energy fuels. 9. According to Hilpold (2015), the EU solidarity principle means that contributions (transfers) across MS or from the EU budget to MS are given with either (a) the hope of receiving counter-contributions at some point in the future or (b) the intent to pursue a common goal. 10. The subsidiarity principle defines the exercise of the EU’s competences to be justifi- able only if the European Union can improve on the MS’ action (EP 2015). 11. We will specify how we define “appropriateness” later in this chapter. 12. See Lisbon Treaty and tax legislation in the European Union (EC 2015c; 2015b). 13. The claim that allowing special treatments, such as grandfathering, is only possible within an ETS is questioned by Goulder and Schein (2013), who argue that when using a carbon tax price, tax exemptions can achieve similar effects as those from grandfathering. 14. See, e.g., Garside (2015b; 2015a). € 15. At the current exchange rate of 1.35, this accounts for approximately 25/tCO2e in € the year 2015 and 40/tCO2e in the year 2020. 16. Under the Swedish carbon tax program, small industrial producers and agriculture and forestry sectors pay lower carbon taxes than do households (OECD 2014). 17. Roughly defined, the marginal abatement cost is the cost increase from reducing a unit of emissions. 18. For example, by assigning a higher proportion of certificates to Eastern European countries, in particularly Poland (Garside 2015b; EC 2013). 19. We refer to “efficiency” in terms of allocative efficiency, not to be confused with cost- effectiveness, which is sometimes also called “cost-efficiency.” 20. See Sheeran (2006) for intuitive details on the modelling work of Chichilnisky and Heal (1994) and their consecutive work in Chichilnisky et al. (2000). 21. The beneficiaries of the “Kyoto bonus” are Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, and Slovakia. 22. During phase I and II of the EU ETS, EUAs were granted for free (grandfathered) to industry and power companies. Many of these power companies are fully or partially state-owned. In such cases, it is likely that domestic consumers indirectly benefited from these free allowances. Additionally, a proportion of EU ETS emissions are generated by non-domestic and/or transnational firms, inside and outside of electricity production, in which case it is not necessarily the domestic consumers who benefited from the grand- fathered allowances.
23. In addition to climate change externality considerations, there are other reasons why a uniform carbon price—which equalizes MACs across countries—may not be efficient. For example, some countries may wish to price emissions more aggressively for fiscal
55 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 57
reasons. If a country has a relatively mobile tax base with respect to broader fiscal instru- ments (e.g., due to a prevalence of informal markets, tax evasion), then implementing carbon prices may be a fiscal alternative to other taxes. 24. In a comparable setup, Santore et al. (2001) arrive at similar findings. 25. See also Goulder and Schein (2013) and Shobe and Burtraw (2012). 26. In technical terms we impose a separable utility function in which the consumption component is identical across countries. We also assume that an upper regulator such as the European Union equally weighs each country within a social welfare function. 27. Within this essay the EU ETS-transfer estimation only depicts optimal transfers. Esti- mates for other appropriate EU ETS transfers—those transfers that achieve (Pareto) improvements for all MS but not necessarily optimality—are subject to our ongoing research. 28. That is, equity-based transfers and transfers based on historical emissions. 29. If the Member State would not use a variable but a fixed fee, its price would drop
below pMS. We suppose that a Member State may anticipate the price drop effect and therefore adjust its policy to meet its preferred emission level. However, both instru- ments—a variable and a fixed fee—in general generate the same effects in this exercise. For a detailed discussion of a variable fee as a national price floor, see Wood and Jotzo (2011). 30. The mechanism is similar to the UK’s carbon price floor for the EU ETS. 31. See also Goulder and Stavins (2012), who discuss the leakage effect in more detail. 32. The model version applied in this paper comprises 26 of the 28 EU MS plus Switzer- land, Norway, and the Balkan region, and excludes Malta and Cyprus. Except for the Balkan region, all countries are represented as individual model regions. Transmission is modelled as a transport problem from the center of one region to the center of a neigh- boring region, with the maximum transmissible amount of electricity being restricted by the installed net transfer capacity. There are 14 different generation technologies and two different storage technologies represented in LIMES-EU. See Nahmmacher et al. (2014) for detailed model documentation. 33. The large model comparison exercise presented in Knopf et al. (2014) showed that a € carbon price of at least 20/tCO2 is needed before 2030 in order to cost-efficiently reach the long-term decarbonization targets by 2050. 34. The total system costs comprise the dispatch and investment costs of all generation, storage, and transmission technologies until 2050. They are discounted to today’s values with a discount factor of 5 percent per year.
35. See, for example, Inman (2015). 36. In this case, it is a tax. However, it is a constant revenue stream—a feature of both a tax and a price floor.
References
Bianco, Nicholas, Jonas Monast, Tim Profeta, and Franz Litz. 2009. “Allowing States to Retire Allowances without Affecting National Allowance Prices: A Straw Proposal”
56 References
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 58 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
(Working Paper, World Resources Institute/Duke University Nicholas Institute for Envi- ronmental Policy Solutions). https://nicholasinstitute.duke.edu/sites/default/files/ publications/allowing-states-to-retire-allowances-without-affecting-national-allowance- prices-a-straw-proposal-paper.pdf. BMWi (Bundesministerium für Wirtschaft und Energie). 2015. “Eckpunkte-Papier, Strom- markt.“ Bundesministerium für Wirtschaft und Energie. http://www.bmwi.de/BMWi/ Redaktion/PDF/E/eckpunkte-papier-strommarkt,property=pdf,bereich=bmwi2012,spr ache=de,rwb=true.pdf. Böhringer, Christoph, Henrike Koschel, and Ulf Moslener. 2008. “Efficiency Losses from Overlapping Regulation of EU Carbon Emissions.” Journal of Regulatory Economics 33 (3): 299-317. Burtraw, Dallas. 2014. “Commentary: Saving Europe’s Key Weapon against Climate Change, 2014.” Resources for the Future 187. http://www.rff.org/research/publications/ commentary-saving-europe-s-key-weapon-against-climate-change. Chichilnisky, Graciela, and Geoffrey Heal. 1994. “Who Should Abate Carbon Emissions?” Economics Letters 44: 443-49. Chichilnisky, Graciela, Geoffrey Heal, and David Starrett. 2000. “Equity and Efficiency in Environmental Markets: Global Trade in Carbon Dioxide Emissions.” In Environmental Markets: Equity and Efficiency, eds. G. Chichilnisky and G. Heal. Chichester, West Sussex, UK: Columbia University Press, 46-67. Coase, Ronald H. 1960. “The Problem of Social Cost.” Journal of Law and Economics 3: 1-44. Copenhagen Economics. 2010. “Innovation of Energy Technologies: The Role of Taxes, No. 0036” (Taxation Studies, Directorate General Taxation and Customs Union, European Commission). http://ec.europa.eu/taxation_customs/resources/documents/common/ publications/studies/taxation_energy_innov.pdf. Cramton, Peter, Axel Ockenfels, and Steven Stoft. 2015. “An International Carbon-Price Commitment Promotes Cooperation.” Economics of Energy & Environmental Policy 4 (2): 51-64. EC (European Commission). 2011. “Energy Roadmap 2050—Impact Assessment and Scenario Analysis.” European Commission: Energy. https://ec.europa.eu/energy/en/ topics/energy-strategy/2050-energy-strategy. EC. 2012. “The State of the European Carbon Market in 2012.” COM(2012) 652 final. http://ec.europa.eu/clima/policies/ets/reform/docs/com_2012_652_en.pdf. EC. 2013. “The EU Emissions Trading System (EU ETS) Factsheet.” http://ec.europa.eu/ clima/publications/docs/factsheet_ets_en.pdf.
EC. 2014. “A Policy Framework for Climate and Energy in the Period from 2020 Up to 2030. SWD(2014) 15 final.” http://ec.europa.eu/smart-regulation/impact/ia_carried_ out/docs/ia_2014/swd_2014_0015_en.pdf. EC. 2015a. “EU ETS Auctioning.” European Commission: Climate Action. http://ec.europa .eu/clima/policies/ets/cap/auctioning/index_en.htm.
EC. 2015b. “EU Tax Policy Strategy.” European Commission: Taxation and Customs Union. http://ec.europa.eu/taxation_customs/taxation/gen_info/tax_policy/index_en .htm.
57 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 59
EC. 2015c. “The Lisbon Treaty and Tax Legislation in the EU.” European Commission: Taxation and Customs Union. http://ec.europa.eu/taxation_customs/taxation/gen_info/ tax_policy/article_6759_en.htm. EC. 2015d. “Allowances and Caps.” European Commission: Climate Action. http://ec.eur opa.eu/clima/policies/ets/cap/index_en.htm. Edenhofer, Ottmar, Michael Jakob, Felix Creutzig, Christian Flachsland, Sabine Fuss, Martin Kowarsch, Kai Lessmann, Linus Mattauch, and Jan Siegmeier. 2015. “Closing the Emission Price Gap.” Global Environmental Change 31: 132-43. EEA (European Environment Agency). 2013. Trends and Projections in Europe 2013—Track- ing Progress towards Europe’s Climate and Energy Targets until 2020. Luxembourg: Publica- tions Office of the European Union. http://www.eea.europa.eu//publications/trends- and-projections-2013. EEA. 2015. “EEA Greenhouse Gas—Data Viewer.” European Environment Agency: Data and Maps. http://www.eea.europa.eu/data-and-maps/data/data-viewers/greenhouse-gases- viewer. EEX (European Energy Exchange). 2015. “Emissions Secondary Market.” European Energy Exchange: Environmental Markets. https://www.eex.com/en/products/environmental- markets/emissions-secondary-market. EP (European Parliament). 2014. “Indirect Taxation.” European Parliament: Fact Sheets on the European Union. http://www.europarl.europa.eu/atyourservice/en/displayFtu.html ?ftuId=FTU_5.11.3.html. EP. 2015. “The Principle of Subsidiarity.” European Parliament: Fact Sheets on the European Union. http://www.europarl.europa.eu/ftu/pdf/en/FTU_1.2.2.pdf. Eyraud, Luc, Abdoul Wane, Changchang Zhang, and Benedict Clements. 2011. “Who’s Going Green and Why? Trends and Determinants of Green Investment” (IMF Working Paper 11/296). https://www.imf.org/external/pubs/ft/wp/2011/wp11296.pdf. Garside, Ben. 2015a. “BRIEFING: In or Out? Poland Approaches EU Climate Cross- roads.” Carbon Pulse. http://carbon-pulse.com/briefing-in-or-out-poland-approaches-eu- climate-crossroads. Garside, Ben. 2015b. “Poland’s Surging Opposition Seeks Talks on EU Emission Opt- out.” Carbon Pulse. http://carbon-pulse.com/polands-surging-opposition-seeks-talks- on-eu-emission-opt-out-bloomberg. Goulder, Lawrence H., and Andrew R. Schein. 2013. “Carbon Taxes Versus Cap and Trade: A Critical Review.” Climate Change Economics 4 (3).
Goulder, Lawrence H., and Robert N. Stavins. 2011. “Challenges from State-Federal Interactions in US Climate Change Policy.” American Economic Review 101 (3): 253-57. Grubb, Michael, Heleen de Coninck, and Ambuj D. Sagar. 2014. “From Lima to Paris, Part 2: Injecting Ambition.” Climate Policy 15 (4): 413-16.
Heinemann, Friedrich, Philipp Mohl, and Steffen Osterloh. 2009. “Who’s Afraid of an EU Tax and Why? Revenue System Preferences in the European Parliament.” The Review of International Organizations 4 (1): 73-99.
58 References
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. 60 Edenhofer, Roolfs, Gaitan, Nahmmacher, and Flachsland
Hilpold, Peter. 2015. “Understanding Solidarity within EU Law: An Analysis of the ‘Islands of Solidarity’ with Particular Regard to Monetary Union.” Yearbook of European Law 34: 257-85. http://dx.doi.org/10.2139/ssrn.2599725. HM Revenue & Customs. 2014. “Carbon Price Floor: Reform and Other Technical Amendments.” Gov.UK: Business Tax—Policy Paper. https://www.gov.uk/government/ publications/carbon-price-floor-reform. Inman, Philip. 2015. “UK Deficit Rises Steeply after Surprise Fall in Tax Receipts.” The Guardian. http://www.theguardian.com/business/2015/sep/22/uk-deficit-rises-steeply- after-surprise-fall-in-tax-receipts. IPCC. 2014. Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY, US: Cambridge University Press. Kahn, Debra. 2015. “California Engaging in Carbon-Trading Discussions with Other Western States.” E&E Publishing. http://www.eenews.net/stories/1060023751.
Knopf, Brigitte, Nicolas Koch, Godefroy Grosjean, Sabine Fuss, Christian Flachsland, Michael Pahle, Michael Jakob, and Ottmar Edenhofer. 2014. “The European Emissions Trading System (EU ETS): Ex-Post Analysis, the Market Stability Reserve and Options for a Comprehensive Reform.” Nota di Lavoro 79, Fondazione Eni Enrico Mattei, Milan, Italy. Koch, Nicolas, Sabine Fuss, Godefroy Grosjean, and Ottmar Edenhofer. 2014. “Causes of the EU ETS Price Drop: Recession, CDM, Renewable Policies or a Bit of Everything? New Evidence.” Energy Policy 73: 676-85. Koch, Nicolas, Godefroy Grosjean, Sabine Fuss, and Ottmar Edenhofer. 2016. “Politics Matters: Regulatory Events as Catalysts for Price Formation under Cap-and-Trade.” Journal of Environmental Economics and Management 78: 121-39. Kollmuss, Anja, and Michael Lazarus. 2010. “Buying and Cancelling Allowances as an Alternative to Offsets for the Voluntary Market: A Preliminary Review of Issues and Options” (OECD Environment Working Paper No. 21). Landis, Florian, Oliver Schenker, Miguel Angel Toar Reanos, Christina Vonnahme, and Sonja Zitzelsberger. 2012. “An Overview on Current Climate Policies in the European Union and Its Member States.” Centre for European Economic Research, ZEW Mannheim. Nahmmacher, Paul, Eva Schmid, and Brigitte Knopf. 2014. “Documentation of LIMES-EU—A Long-Term Electricity System Model for Europe.” Potsdam Institute for Climate Impact Research, Potsdam, Germany. https://www.pik-potsdam.de/members/ paulnah/limes-eu-documentation-2014.pdf. Oates, Wallace E. 1972. “An Essay on Fiscal Federalism.” Journal of Economic Literature 37: 1120–49.
Oates, Wallace E. 2011. Fiscal Federalism. Cheltenham, UK: Edward Elgar Publishing Ltd. OECD. 2013. Effective Carbon Prices. Paris, France: OECD Publishing. http://dx.doi.org/ 10.1787/9789264196964-en.
OECD. 2014. “Sweden—Highlights.” OECD: Environmental Country Reviews. http:// www.oecd.org/env/country-reviews/highlights.htm.
59 2. Agreeing on an EU ETS Price Floor
©MIT Press. Chapter 2 from Energy Tax and Regulatory Policy in Europe. For evaluation purposes only. Solidarity, Subsidiarity, and Efficiency in the EU 61
Parry, Ian, Chandara Veung, and Dirk Heine. 2014. “How Much Carbon Pricing is in Countries’ Own Interests? The Critical Role of Co-Benefits” (CESifo Working Paper No. 5015). https://www.cesifo-group.de/DocDL/cesifo1_wp5015.pdf. Philibert, Cédric. 2006. “Certainty Versus Ambition: Economic Efficiency in Mitigation Climate Change” (International Energy Agency Working Paper Series). http://philibert .cedric.free.fr/Downloads/rb_certainty_ambition.pdf. Philibert, Cédric. 2009. “Assessing the Value of Price Caps and Floors.” Climate Policy 9 (6): 612-33. Quandl. 2015a. “ICE Futures Europe. ECX EUA Futures, Continuous Contract #2.” https://www.quandl.com/data/CHRIS/ICE_C2-ECX-EUA-Futures-Continuous- Contract-2-C2. Quandl. 2015b. “ICE Futures Europe. ECX EUA Futures, December 2020.” https://www .quandl.com/data/ICE/CZ2020-ECX-EUA-Futures-December-2020-CZ2020. Roolfs, Christina, Beatriz Gaitan, and Ottmar Edenhofer. 2016. “The Richest Wins Them All: Triggering Benevolent Hegemony by Federal Transfers.” (Mimeo). Santore, Rudy, H. David Robison, and Yehuda Klein. 2001. “Strategic State-Level Envi- ronmental Policy with Asymmetric Pollution Spillovers.” Journal of Public Economics 80 (2): 199-224. Sheeran, Kristen A. 2006. “Who Should Abate Carbon Emissions? A Note.” Environmental and Resource Economics 35 (2): 89-98. Shiell, Leslie. 2003. “Equity and Efficiency in International Markets for Pollution Permits.” Journal of Environmental Economics and Management 46 (1): 38-51. Shobe, William M., and Dallas Burtraw. 2012. “Rethinking Environmental Federalism in a Warming World.” Climate Change Economics 3 (4): 33. Skjærseth, Jon Birger, and Jørgen Wettestad. 2010. “Making the EU Emissions Trading System: The European Commission as an Entrepreneurial Epistemic Leader.” Global Environmental Change 20 (2): 314-21. Strunz, Sebastian, Erik Gawel, and Paul Lehmann. 2015. “Towards a General ‘European- ization’ of EU Member States’ Energy Policies?” Economics of Energy & Environmental Policy 4 (2). Talus, Kim. 2013. EU Energy Law and Policy: A Critical Account. Oxford: Oxford University Press. UNFCCC. 2015. “EU Submits its Climate Action Plan Ahead of Paris 2015 Agreement.” http://www4.unfccc.int/submissions/INDC/Published%20Documents/Latvia/1/LV- 03-06-EU%20INDC.pdf.
Williams, Roberton C. 2012. “Growing State–Federal Conflicts in Environmental Policy: The Role of Market-Based Regulation.” Journal of Public Economics 96 (11-12): 1092-99. Wood, Peter John, and Frank Jotzo. 2011. “Price Floors for Emissions Trading.” Energy Policy 39 (3): 1746–53.
60 3 Make or brake — Rich states in voluntary federal emission pricing
Christina Roolfs Beatriz Gaitan Ottmar Edenhofer
Resubmitted to Journal of Environmental Economics and Management
61 3. Make or Brake
Abstract Voluntary participation can improve multilateral environmental governance. We develop a theory of voluntary participation in federal environmental policy by states of difering wealth. Using a general equilibrium model, we formalize voluntary participation by a Pareto-improving federal emission price that coexists with state-level emission pricing. Federal revenues are distributed equally per capita (egalitarian), in proportion to states’ historical emission levels (sovereignty), or states’ actual payments (juste retour). We fnd that the existence of Pareto-improving uniform federal prices depends on the transfer rules, and on whether or not states anticipate them. Sovereignty transfers work in all cases. The efectiveness of egalitarian transfers is hampered by too large diferences in wealth between states. Juste retour transfers render federal policy inefective if states anticipate them. The lowest optimal federal price maximizes the utility of the richest state and represents the voluntarily-feasible federal minimum price. In that sense, rich states brake or make possible voluntary federal policy.
JEL-classifcation: H77, Q58, H23, D62, H87
Keywords: Environmental Regulation, Fiscal Federalism, Pollution Taxes, Transfer Design, Voluntary participation, Pareto-improvements, Minimum price
Reference: Roolfs, Gaitan, Edenhofer (resubmitted to JEEM): Make or brake — Rich states in voluntary federal emission pricing (April 24, 2020). Working paper available at SSRN: http://dx.doi.org/10.2139/ssrn.2941505
62 3.1 Introduction
3.1 Introduction
When facing a common threat, theory predicts that rich people and countries voluntarily contribute more to mitigating this threat than poor ones (cf. Bergstrom et al., 1986; Olson and Zeckhauser, 1966). Environmental pollution, pandemics, and climate change are prime examples of such threats in modern history. When it comes to multinational environmental policy coordination, a uniform carbon price has become a focal point of political and academic discourse — be it in the form of an emission trading scheme or an emission tax (Cramton et al., 2015; Weitzman, 2014).1 With a uniform price, allocative efciency dictates redistributive transfers from rich to poor entities (cf. Chichilnisky and Heal, 1994; Engström and Gars, 2015; Sandmo, 2007). But the necessary transfers may exceed voluntary contributions (e.g. Sandmo, 2007; Shiell, 2003; Stavins, 1997). Surprisingly, the conditions for voluntary participation of states in federal environ- mental policy have not yet been investigated. We study the role of a federal government as a coordinator that uses transfers and a federal emission price to ensure both voluntary participation by the member states and a reduction in emissions for the whole federation. Our framework takes into account the coexisting emission policies at the state and federal level. The policies of one level of government afect the revenues of the other, which results in vertical fscal externalities (cf. Dahlby and L. Wilson, 2003; Keen and Kotsogiannis, 2002; J. Wilson, 2006). In this way, we combine the theory of voluntary public good provision with fscal federalism in the context of environmental regulation. In this framework, we ask: which state becomes a bottleneck for emission mitigation when contributions are voluntary and policy is coordinated by a federal authority? When analyzing which uniform federal prices ensure the voluntary participation and contributions of all states, we consider three diferent transfer rules to distribute federal emission price revenues – equal per capita (egalitarian) transfers, transfers proportional to a states’ historical emission levels (sovereignty), and transfers proportional to actual emission payments (juste retour). An important fnding of our paper is that the richest state takes on the role of the largest net donor under any transfer scheme, and its utility is maximized at the lowest of all optimal federal prices that ensure voluntary participation (minimum price). If the richest state, in its position as the largest donor, demands that federal policy maximizes its utility, it can put brakes on the stringency of federal policy by becoming a bottleneck. At the same time, by virtue of its role as the largest donor, the richest state enables federal policy in the frst place. Requiring voluntary participation or the consent of the member states can be a crucial constraint for the implementation of federal policies. In the federal system of the European Union (EU), EU policy-making requires unanimous or majority consent of the
1European emissions trading also has a uniform price signal. It is the largest emissions trading system in the world. In addition, the European political discourse considers introducing a minimum price there.
63 3. Make or Brake member states. Also, in the EU and the United States, for example, some states have introduced their own environmental policies that coexist with federal environmental policy. They either did so in order to pursue more ambitious goals by establishing accompanying policies, or to mitigate the efects of federal policy on local economies, for example through subsidies (cf. Burtraw, Keyes, et al., 2018; Williams, 2012). Improving the understanding of the obstacles and requirements for voluntary participation in federal environmental policy can i) allow for better policy designs that ensure voluntary participation, ii) allow for assessments of the impact of coexisting policies at the state and federal levels, and iii) provide guidance for revenue distribution in federal systems. In the following, we shall, for the sake of readability, refer to federal emission prices that ensure voluntary participation of the member states and that cannot be Pareto-dominated, simply as “feasible federal emission prices”. We develop a general equilibrium model with coexisting state-level and federal emission pricing, where all governments strive to reduce the damage caused by transboundary emissions. Emissions are an essential input for private good production. While state governments focus on the well-being in their local economies, the federal government cares for the well-being of the entire federation. We suppose that the federal government has a strategic advantage over state governments such that it can infuence state policy choices through its own policy choice (cf. J. Wilson, 2006). We model the federal government as a Stackelberg leader that is informed about the individual heterogeneous interests of its member states and is therefore able to propose federal emission prices guaranteeing their voluntary participation. The notion of voluntary participation in federal policies is captured by the concept of Pareto- improvements relative to the decentralized solution2. Federal revenues are distributed based on egalitarian, sovereignty, and juste retour criteria all of which are well-established in policy practice and theory (for details see Section 3.2). To compare our results with existing multinational climate policy and fscal federalism literature, we contrast the case in which states internalize the vertical fscal externality by taking the federal transfers into account in their own policy-making (henceforth called anticipation), with the case in which they do not internalize the vertical fscal externality (henceforth called no anticipation). Our analytical investigation focuses on state heterogeneity in terms of diferent capital stocks. In addition, we numerically explore in more detail the dimension of population size heterogeneity and discuss how it afects our main fndings. We fnd that if the federal policy is feasible, then the richest state voluntarily pays the bulk of federal revenues. This is good news for environmental policy because from an efciency point of view with a multinational uniform emission price; the richest entity should be the largest donor (cf. Chichilnisky and Heal, 1994; Sandmo, 2007; Shiell, 2003). Our analysis reveals a range of feasible federal emission prices. All else equal, the
2It means that no state is worse of by federal policy compared to its decentralized result, while federal policy makes at least one state better of compared to its decentralized result.
64 3.1 Introduction
lowest feasible federal price maximizes the utility of the richest state. We call this the (feasible) federal minimum price. If feasible, the highest optimal federal price maximizes the utility of the poorest state; this price represents the (feasible) federal maximum price. We fnd that the sovereignty transfer rule is the only rule that always yields feasible federal prices regardless of the capital heterogeneities present. Juste retour transfers perform identically to the sovereignty transfer rule if they are unanticipated by the states. Similar to d’Autumne et al. (2016) and Shiell (2003), whose models do not cover anticipation, we also fnd that juste retour transfers are feasible and efective when they are not anticipated. If the states, however, anticipate federal juste retour transfers, this rule always becomes inefective, so that the outcome corresponds to the decentralized solution. If capital heterogeneity is too pronounced, egalitarian transfers fail to produce Pareto-improvements at any federal emission price. Our analysis shows how a federal system can be designed such that the richest state is willing to be a member and the largest contributor. In this way we connect to the second generation of fscal federalism, which examines the institutional design of federations as central determinant of their viability (cf. Oates, 2005). Our numerical analysis delivers several extensions to the analytical results. Among others, it accounts for two diferent notions of a rich state by introducing diferent population sizes — we can distinguish between a high capital stock in absolute terms and a high capital stock per capita. If egalitarian transfers are used, the minimum feasible federal emission price corresponds to the richest state per capita. If, on the other hand, sovereignty transfers or unanticipated juste retour transfers are used, this role is taken on by the state with the largest capital stock in absolute terms. We also compare the feasible federal price ranges under the various transfer rules. We fnd that the lowest minimum price always occurs with unanticipated egalitarian transfers. However, which transfer rule (with its respective minimum price) creates more emission abatement is ambiguous and depends on parameters such as the sensitivities of output and damages to emissions, the size of the population and also transfer anticipation. The rest of the paper is structured as follows. In Section 3.2, we relate the contributions of our paper to the existing literature on voluntary public good provision, multinational environmental policy and environmental fscal federalism, and link the considered transfer criteria to commonly used equity criteria for burden sharing. Section 3.3 presents and discusses the multilevel general equilibrium model. The impact of diferent transfer criteria is presented in Section 3.4. We divide this section into an analytical part, where we derive the main contributions of this paper and a numerical part which delivers sensitivity analysis and nuances to our analytical fndings. We conclude in Section 3.5.
65 3. Make or Brake
3.2 Literature review
Our analysis of voluntary participation in multilevel environmental policy contributes to the literature on multinational environmental policy, and environmental fscal federalism by combing key aspects of four related strands of literature. We adopt the multilevel governance structure of environmental fscal federalism to refect the decision processes of federations such as the European Union, Canada, Switzerland, Germany and the United States. We draw on the literature on voluntary public good provision to represent sovereign, self-interested decision making of member states in a federation. Since a voluntary multilateral environmental policy may only be feasible if it is regarded as fair by the participants, we use fndings of literature on equity and burden-sharing. This section reviews relevant contributions from the respective three strands of literature in turn. A large part of the literature on multinational environmental policy abstracts from unanimous decision making, focusing instead on efciency and equity concerns of a uniform multinational environmental policy in frst- or second-best settings. These studies investigate top-down regulation from the perspectives of a social planner and/or centralized policy (Chichilnisky and Heal, 2000; Chichilnisky and Heal, 1994; Sandmo, 2007; Shiell, 2003), or a top-level government with delegation authority over all lower levels of government (e.g. d’Autumne et al., 2016). Chichilnisky and Heal were the frst to show that for a global public good, like carbon emission mitigation, equity and efciency cannot be considered separately. In their seminal 1994 paper, they show that achieving efciency with a uniform multinational price, multinational emission mitigation requires transfers from rich to poor states because of decreasing marginal utility of consumption. Engström and Gars (2015) provide a recent overview of this literature strand. These theoretically optimal transfers, however, can be very large and encounter the resistance of states or countries following their self-interest (cf. Edenhofer et al., 2017; Gruber, 2000; Sandmo, 2004; Sandmo, 2007; Stavins, 1997; Wiener, 2007). The top-down view of regulation is complemented by the bottom-up perspective on multinational transfers found in the literature on international environmental agreements (cf. Barrett, 2005). The contrast between central and decentralized regulation is one key motivation for the research agenda of fscal federalism. Indeed, the frst generation of the literature on fscal federalism is particularly interested in determining which regulatory tasks to centralize and which tasks to leave to decentralized decision-making, particularly for the provision of (local) public goods. Drawing mainly on efciency and equity arguments, and similar to the multinational environmental policy literature, this literature fnds that the main responsibility for federal public good provision should be left to the federal government (e.g. Musgrave, 1959; Oates, 1972; Oates, 2000). See also Oates (2005) for an overview of the literature. Early studies on fscal federalism assume that the federal government plays a passive role, e.g. by containing horizontal fscal
66 3.2 Literature review
externalities resulting from tax competition between states. In this context, the "race to the bottom" is often referred to, in which states undercut each other by tax reductions in the competition for mobile factors and thus a locally larger tax base. If, however, states and the federal level compete for the same tax base, vertical fscal externalities can become as central to the analysis of federal systems as horizontal externalities (cf. Keen, 1998; J. Wilson, 2006). Vertical fscal externalities have been extensively studied absent of environmental regulation (Böhringer, Rivers, and Yonezawa, 2016; Bruellhart and Jametti, 2006; Dahlby and L. Wilson, 2003; Keen, 1998; Keen and Kotsogiannis, 2002) and can induce state governments to overtax the local tax base (cf. Keen and Kotsogiannis, 2002). Böhringer, Rivers, and Yonezawa (2016) were the frst to assess the importance of vertical fscal externalities in the context of environmental regulation. Applied to the Canadian Federation, they fnd that due to vertical fscal externalities, a state can implement environmental policy at low cost to itself and the expense of other states. A literature review on fscal federalism, and the research on decentralization of environmental policies is given by Dalmazzone (2006). The second generation of fscal federalism is concerned with the viability of federal institutions (Oates, 2005). One aspect of these studies is that a federal government can have only limited control or competence over the political actions of its states. The EU, for example, is severely limited when it comes to EU-wide tax reforms, since they require the unanimous consent of the member states (Talus, 2013). Few studies exist that deal with limited environmental control in federal systems. Williams (2012) examines limited control by considering the simultaneous existence of state and federal emissions policies. In terms of efciency, he fnds the superiority of a federal tax over quantity controls since the additivity of taxes prevents a mutual overruling of state and federal policies. Depending on the federal policy instrument used, he fnds that sometimes there are no optimal transfers. The literature on voluntary provision of public goods shifts the focus away from efciency concerns. Instead, it investigates voluntary contributions to a public good by decentralized entities. It was pioneered by Olson (1965) and Olson and Zeckhauser (1966) and formalized in the model by Bergstrom et al. (1986). In this context, voluntary participation can be seen as another constraint that multinational policy needs to work with. Olson and Zeckhauser (1966) show that the USA, as the wealthiest NATO member, also contributed the most to the NATO expenditures in the arms race with the Warsaw Pact. Similarly, Olson (1986) discusses how benevolent, yet hegemonic states tend to create multinational systems for public good provision. While the public good benefts all states belonging to the system, the hegemonic state voluntarily bears a disproportionately large cost share of the public good expenses. Redistribution of wealth can, however, have a negative impact on the level of voluntary public good provision, as Bergstrom et al. (1986) show. This result has its origin similar to the decreasing marginal utility of consumption causing the richest entities to be most willing to contribute to the
67 3. Make or Brake public good. Via this simple mechanism, redistribution results in lower overall public good contributions, as entities that are not consumption-saturated would rather use their income for consumption than public good contributions. Such an analysis clearly neglects the potential positive welfare implications of decreasing the inequality across entities, as demonstrated by Chichilnisky and Heal (1994) and others. Our setting takes both considerations into account, as changes in inequality are limited by the requirement of a Pareto-improving federal policy. Multilateral environmental policy that is considered equitable opens the space for voluntary participation. A broad spectrum of equity criteria is developed and discussed in the equity literature (e.g. Burtraw and Toman, 1992; Cazorla and Toman, 2001; Grubb et al., 1992; Kverndokk, 2018; Kverndokk and Rose, 2008; Pottier et al., 2017; Rawls, 1971; Ringius et al., 2002; Rose, 1992; Rose and Stevens, 1993; Rose, Stevens, et al., 1998). Voluntary participation in multinational policy is considered one such equity criterion. In this literature it is referred to as the "compensation" or "Pareto" criterion (e.g. Cazorla and Toman, 2001; Kverndokk and Rose, 2008). Similar to the present paper, it is covered by the multinational (federal) objective of improving the welfare of all consumers in relation to the outcome of decentralized policy. There is, however, no consensus on the "best" equity criterion (cf. Kverndokk, 2018). Recent surveys of the literature can be found in Kverndokk (2018), Kverndokk and Rose (2008), Paterson (2001), and Pottier et al. (2017). We consider three diferent criteria for federal transfer rules that are both well established in the equity literature and applied in federal policy in practice. These are the egalitarian, sovereignty and juste retour criteria. Transfers based on the egalitarian criterion presume an equal ownership of a common resource (e.g. atmosphere) implying that everyone should get an equal share of its revenues. Posner and Sunstein (2008) argue that many people fnd the per capita approach attractive because of its simplicity and appeal to fairness (see supporters also in Grubb et al., 1992; Klenert et al., 2018). In the federal context it is applied, for instance, by the Swiss Federal government which equally distributes part of the revenues from the Swiss CO2 levy back to all Swiss residents (FOEN, 2016). The (ex- ante) sovereignty criterion assumes that past emissions give a right to future emissions (e.g. Böhringer, Rivers, Rutherford, et al., 2015; Grubb et al., 1992). It rewards past higher emission levels and can therefore be considered to be more attractive to richer countries with past higher levels of economic activity. In the literature it is also referred to as “status quo” criterion (e.g. Grubb et al., 1992). In practice the EU’s ETS revenue distribution, for instance, largely takes into account states’ emission levels before the EU ETS (EC, 2015; EC, 2013). While the previous criteria can be determined before a federal policy is introduced (ex-ante), the juste retour criterion accounts for the actual level of emissions. It is therefore an outcome or an ex-post-based criterion (cf. Böhringer, Rivers, Rutherford, et al., 2015). Juste retour literally means “fair return”. It presumes that the actual emission payments of a state to the multinational (or federal) government grant
68 3.3 The model the state the right to federal revenue transfers equal to that payment. The literature also refers to this type of transfers as “no intercountry” transfers (e.g. d’Autumne et al., 2016; Shiell, 2003). As Shiell (2003) puts it, a state that feels relatively poor might not be willing to pay transfers to relatively richer states and might articulate this concern in its negotiation position. In practice, juste retour transfers are often requested from the EU by EU Member States (Warleigh, 2004). For a more technical overview of the transfer rules, see Table 3.1. We combine the four previously discussed strands of literature to contribute to the understanding of environmental policy in multilateral systems and especially federations. Studies on multinational environmental policy fnd that an efcient multinational emission price requires redistribution from rich to poor states. The equity literature presents a variety of fairness criteria, among which rank Pareto-improvements, and egalitarian, sovereignty, and juste retour transfers. The literature on fscal federalism focuses on the efciency and viability of federations and a multilevel policy structure, fnding that policy interactions across multiple levels can incentivize states to override federal policy or to pass on the costs of its local (environmental) policy to other states. The literature on voluntary public good provision examines the willingness of self- interested entities to create or to contribute to a public good. It fnds that a benevolent, yet hegemonic state is willing to create a multinational system, and that rich entities voluntarily contribute more to public good provision than poor entities. Conditions for voluntary participation by the states in federal environmental policy have not yet been examined. This paper combines the theories of voluntary provision of public goods and fscal federalism. We consider insights into multinational environmental policy design and equitable burden-sharing since we consider a transboundary emissions damage (the mitigation of which is a public good). Specifcally, we contribute the environmental focus to the second generation of fscal federalism, which examines the institutional design of federations as a central determinant of their viability.
3.3 The model
The model represents a federation of m member states. Member states can difer in their capital stock and population size. Population and capital are immobile across states. The entire population consists of identical consumers within a state. Each consumer rents out its capital endowment to the domestic frm. Consumers receive a transfer from the revenues of the state and federal emission prices. Firms pay for the emission of harmful transboundary emissions during the production of the fnal good. Each consumer derives utility from the consumption of a private good and dis-utility (damage) from emissions. But to produce the private good, frms must use emissions and capital. State and federal governments choose optimal emission prices that strike a balance between emissions and private good consumption and recycle the revenues back to the
69 3. Make or Brake
Criterion Defnition Operationalized Formula rule ex-ante Egalitarian Equal ownership of Every person gets the = 1 si ∑ n the atmosphere in same share of federal m j which emissions are emission price revenues. stored. E0 Sovereignty Past emissions Federal revenues are = 1 i si n E0 grant a right to distributed among the i actual emissions. states in proportion to each decentralized emission levels. ex-post Juste Retour Actual emission Federal revenues are = 1 Ei si n E (no payments grant a distributed among i intercountry right to federal consumers in a state in transfers) revenues. proportion to the actual emission level of that state.
Table 3.1: Transfer criteria and operationalized federal transfer rules. In the formulas si represents the per capita transfer share of each consumer in state i = 1, ..., m, where i and j index the states. ni is the number of consumers in state i. Emission levels of state i from the decentralized state policy and the state–federal policy solutions are 0 and , respectively. Similarly, 0and denote the aggregate federal emission levels. Ei Ei E E Table adapted from Böhringer, Rivers, Rutherford, et al. (2015), Cazorla and Toman (2001), Grubb et al. (1992), Kverndokk and Rose (2008), and Ringius et al. (2002). population. Each state government charges a price on domestic emissions and distributes the revenue equally among its population. The federal government sets a uniform emission price, in addition to state prices, in case this leads to a Pareto improvement relative to the decentralized state policy solution. It distributes revenues based on the egalitarian, sovereignty or juste retour criteria. Since the distribution rules are given, the federal solution is a second-best optimum. In Table 3.1 we provide an overview of the transfer criteria considered. We operationalize these as transfer rules in our model following the existing literature. We use si to represent the federal transfer share to each consumer in state i = 1, ..., m, while ni denotes the number of consumers in state i. For ex-ante criteria, the operationalization is independent of the federal policy outcome. We use the benchmark scenario without federal policy as the point of reference instead of historical data (cf. decentralized equilibrium). Decentralized emission levels of a state are denoted by 0 and the total Ei decentralized emission level is E0. Actual emission levels, when state and federal policy coexist, are denoted by Ei and E. The structure of the model can be summarized as follows. In the frst stage, the federal government sets a federal uniform emission price. Its objective is to make at
70 3.3 The model least one state better of, while no other state is worse of compared to the decentralized solution (Pareto-improvement). The federal government has information on individual heterogeneous interests and reactions of its member states and thus acts as a Stackelberg leader. In the second stage, based on the reactions of frms and consumers and taking the federal price as a given, each state government non-cooperatively sets a price on domestic emissions. Its objective is to maximize the utility for the local population. In the third stage, consumers and frms in each state solve their optimization problems, taking all prices, and transfers as given.
3.3.1 Private sector agents
3.3.1.1 Firms
In each state i a representative frm produces a homogeneous fnal good Yi which is identical across states. The fnal good is used as a numéraire. To produce Yi the i frm in state i uses a continuously diferentiable production function Y (Ki,Ei) that is homogeneous of degree one where Ki and Ei are capital and emissions, respectively. Let i i( ) and i 2 i( ) . Production increases in both inputs Yx ≡ ∂Y x, z /∂x Yxz ≡ ∂ Y x, z /∂x∂z with diminishing marginal returns, i.e. Y i < 0 < Y i and Y i < 0 < Y i . EiEi Ei KiKi Ki Taking prices as given, frm i chooses Ki and Ei to maximize its profts. The rental rate of capital in state i is denoted by ri, and the per unit cost of emission price is the sum of the state emission price pi and the uniform federal emission price P . Since the fnal good’s price is numéraire, frms maximize profts by setting their marginal cost of production equal to one, i.e. mci = 1, and by setting the marginal product of capital (Y i ) and emissions (Y i ) equal to their respective per unit cost, i.e. Y i = r Ki Ei Ki i and Y i = p + P . Ei i
3.3.1.2 Consumers
Each state i is populated by ni identical consumers. Each consumer derives utility from consuming the fnal good. Aggregate federal emissions, given by = ∑m , negatively E i=1 Ei afect each household’s utility. We assume an additively separable utility function. The i utility function of the representative consumer of state i is given by u (ci,E), where c denotes fnal good consumption. We assume that ui > 0, ui ≤ 0, ui < 0, and i ci cici E i 0, which implies that the higher emissions are, the greater the marginal dis-utility uEE ≤ from emissions. Consumers take prices, emissions, policies and transfers as given. The representative consumer of state i chooses the level of consumption ci that maximizes her utility subject to her budget constraint
Ki Ei ci = ri + pi + siPE (3.1) ni ni
71 3. Make or Brake
where Ki is the aggregate capital endowment in state i and riKi/ni is the per capita return to capital , and piEi/ni and siPE are state level and federal level transfers to each consumer of state i that stem from coexisting state and federal emission pricing revenues. The federal per capita transfer distributes federal emission price revenues PE based on the transfer rule si as introduced in Table 3.1. Since each consumer takes emissions as given, the solution to each consumer’s optimization problem reduces to setting consumption equal to income, equation (3.1).
Zero profts imply that (riKi/ni + piEi/ni) = Yi/ni − PEi/ni. By substituting this into equation (3.1), state i’s consumption becomes
Yi ( Ei ) ci = + siE − P. (3.2) ni ni
Therefore, state i’s consumption departure from local production Y i is determined by the net federal transfer, (siE − Ei/ni) P .
3.3.1.3 Market clearing and reaction of private sector
Capital market clearing in each state implies that capital demand Ki equals the aggregate capital endowment in state i, i.e. Ki = Ki. Market clearing in fnal goods is given by ∑ = ∑ i m nici m Y . Using market clearing conditions together with the solutions to the problems of consumers and frms (private-sector agents) allows their optimal solutions to implicitly be expressed as functions of state and federal prices. We use bold letters to represent these functional forms.3 These solutions can be considered reaction functions of the private sector agents that the state level and federal level governments are aware of and take into consideration in policy making. We report and discuss these reaction functions in 3.C.2 and 3.I.2 for the specifc Cobb-Douglas and general CES production function, respectively.
3.3.2 Multilevel emission policy State and federal governments both deal from their own perspectives with emission reduction. They are confronted with the problem of fnding the balance between consumption and emissions on diferent levels (state vs federal). Since both levels set a price on and draw revenues from the same emissions, the emission price set at one level of government can have an impact on the revenues of the other level (vertical fscal externality). Previous environmental policy literature considers the unanticipated (d’Autumne et al., 2016; Shiell, 2003) (absent of strategic interactions across governmental
3Note, for example, that Y i (K ,E ) = p + P solves for E as a function of p and P ; and since Ei i i i i i i Y (Ki,Ei) is homogeneous of degree one then mci = 1 solves for ri as a function of pi and P , which i i ( ) we respectively denote by Ei (pi,P ), and ri (pi,P ). Similarly, Y (pi,P ) = Y Ki, Ei (pi,P ) and ( ) = ∑ ( ) with = ( ). E p, P i Ei pi,P p p1, ...pm
72 3.3 The model
layers) as well as the anticipated case (Williams, 2012). For the taxation of wage income, Dahlby (2008) considers it likely that changes in federal revenue induced by state policy are ignored (unanticipated) by the states. In such a case, this ignorance can lead to overtaxation at the state level (Boadway and Keen, 1996; Boadway, Marchand, et al., 1998; Dahlby, 1996; Keen, 1998). We contrast these two cases:
Defnition 1 (Transfer Anticipation). Federal transfers are anticipated, if each state takes into account the efect of domestic policy on the federal transfer revenues received (indicated with *). Federal transfers are unanticipated, if each state does not take into account the efect of domestic policy on the federal revenues received. Formally:
∂(s P E) i ̸= 0 for all i = 1, ..., m "anticipated" federal transfers, (3.3a) ∂pi ∂(s P E) i = 0 for all i = 1, ..., m "unanticipated" federal transfers. (3.3b) ∂pi Our comparison between the anticipated and the unanticipated case allows to cover three diferent federal governance. First, if the state is one among many or if it is small, state policy would have a negligible efect on the amount of transfers received from the federal government. In this case and under the egalitarian and sovereignty transfer rules, unanticipated transfers can be a good approximation of the full decision problem. Unanticipated transfers also have the added beneft of being more analytically tractable. Second, as we will show later in this paper, the federal juste retour transfer performs very diferently in the anticipated and unanticipated cases. Unanticipated juste retour transfers have already been discussed in d’Autumne et al. (2016) and Shiell (2003). Third, for large or few member states, anticipated transfers are the more appropriate framework for policy analysis, because individual states would then have a strong impact on federal transfers and would therefore likely incorporate this fact in their decision making.4
3.3.3 State policy Each state government cares about the well-being of domestic consumers. The government
of state i chooses the emission price pi that maximizes the utility of its consumers while
taking the federal emission price P and all other state-level emission prices pj for all j ̸= i as given. The government of state i considers the solution of all consumers and frms optimization problems together with the market clearing conditions in order to arrive at the reaction functions necessary for its own optimization problem. We can thus rewrite i i consumer i’s utility in terms of p = p1, ..., pm and P as u (p, P ) ≡ u (ci (p, P ) , E (p, P )).
4Our discussion on unanticipated transfers can be conceptually linked to the small economy discussion in the literature on international trade, where individual countries are assumed to be unable to impact international prices and policy.
73 3. Make or Brake
The problem of the local government in state i is formalized by
i max niu (p, P ) given pj ∀j̸=i and P. (3.4) pi
The frst-order condition that characterizes the solution to (3.4) is ui = (ui ∂c /∂p + pi ci i i i ) = 0.5 After several algebraic manipulations6, we obtain uE ∂E/∂pi
∂E p ∂E −ui i = ui i i unanticipated, (3.5a) E ci ∂pi ni ∂pi ∂E ( p ∂E ∂(s P E)) −ui i = ui i i + i anticipated. (3.5b) E ci ∂pi ni ∂pi ∂pi
We can see how the equilibrium state emission price is a balancing act between diferent efects. The left-hand side of both equations (3.5a) and (3.5b) represents the marginal dis-utility of emissions i 0, and the change in local emissions induced uE < 7 by state policy, ∂Ei/∂pi. From Lau and Yotopoulos (1972) follows that ∂Ei/∂pi < 0. Therefore, state policy has a positive efect on utility through the marginal dis-utility. The right-hand side accounts for the impact of state policy on the marginal utility of consumption ui > 0, and the state policy-induced change in consumer income, which ci essentially reduces to emission price revenue transfers, cf. equation (3.1). If federal transfers are unanticipated by states, equation (3.5a), then each state only considers the marginal consumer income changes its policy induces on state transfers, pi/ni∂Ei/∂pi.
Since ∂Ei/∂pi < 0, an increase in state policy always has a negative impact on the size of state transfers. If states anticipate the federal transfers, equation (3.5b), each state additionally considers state policy-induced marginal changes in the federal transfer,
∂(siP Ei)/∂pi. Note that to comply with the frst-order condition, the term in parenthesis of equation (3.5b) must be negative as the left-hand side is negative and ui is positive. ci The efect of state policy on federal transfers can be ambiguous, depending on the sign of ∂(siP E)/∂pi. Suppose that ∂(siP Ei)/∂pi < 0 (this is indeed the case for all transfer rules considered in this paper, as we will show later). Then the state policy unambiguously generates a negative efect on marginal utility of consumption via lowering income from federal transfer revenues. For any federal ex-ante transfer criteria, one can 8 see immediately that ∂(siP Ei)/∂pi < 0 because of ∂E/∂pi = ∂Ei/∂pi < 0 implying
∂(siP E)/∂pi = siP ∂Ei/∂pi < 0.
5Note that n cancels out since n ui = 0 implies that ui = 0. i i pi pi 6Provided in 3.A. 7 The result ∂Ei/∂pi < 0 is also derived in 3.C.2 for Cobb-Douglas and in 3.I.2 for CES production. 8See 3.A.
74 3.3 The model
Rearranging the m frst-order conditions (one for each state) for both cases, i.e. equations (3.5a) and (3.5b), implicitly results in the reaction functions of state policy, which depend solely on the federal emission price:
ui p (P ) = −n E unanticipated, (3.6a) i i ui ci ( ui ∂(s P E ) ∂E ) p∗ (P ) = n − E − i i / i anticipated. (3.6b) i i ui ∂p ∂p ci i i
In case of unanticipated federal transfers, equation (3.6a), the state always chooses a positive emission price p > 0 because ui < 0 and ui > 0. In the case of anticipated i E ci federal transfers, equation (3.6b), the policy choice of the state can be positive or negative, depending on whether the term in parenthesis is positive or negative. If
∂(siP Ei)/∂pi < 0, then the state knows that it can positively infuence the size of federal transfers by reducing its local policy stringency. In doing so, it provides the federal government with a larger revenue base (tax base) ∑ . As a result, it might i Ei even be optimal for a state government to subsidize local emissions to increase the federal revenue base, which, in turn, leads to higher federal transfers. Argued from the perspective of fscal federalism, the state government takes into account the negative vertical fscal externality of the local emission price, which consists of the fact that a price increase by the state government leads to a reduction in federal revenues (cf. Keen, 1998). States’ policy choices can difer from each other due to diferences in population
size ni, in the marginal rate of substitution between aggregate emission reduction and consumption MRS ≡ −ui /ui > 0. In case of anticipation, they can additionally E,ci E ci difer due to the state policy-induced change in the federal transfer. Ceteris paribus, a larger population size and a larger marginal dis-utility of federal emissions both lead to a more stringent state policy, whereas the opposite is true for a larger marginal
utility of consumption. Ceteris paribus, if state i is richer than state j, i.e. Ki > Kj, then the marginal utility of consumption in state i is lower than the marginal utility of consumption in state j, i.e. ui < uj. Therefore, equations (3.6a) and (3.6b) suggest that ci j a rich state sets a higher domestic emission price than a poor state, implying that rich states voluntarily contribute more to emission mitigation. This relationship between marginal utility of consumption and emissions matches the fndings in Bergstrom et al. (1986) and Chichilnisky and Heal (1994) who argue that larger gains from consumption (large ui ) increase the optimal level of emissions which is here refected by a lower state ci price.
75 3. Make or Brake
3.3.4 Decentralized policy equilibrium
Defnition 2 (Decentralized policy equilibrium). The decentralized policy equilibrium 0 0 0 0 0 0 = 1 is defned by the quantities ci ,Ei ,Ki ,Yi and prices pi , ri for i , ..., m, such that 0 0 0 0 each ci solves the optimization problem of each consumer in state i; Ei ,Ki and Yi 0 solve the problem of the frm located in state i; pi solves the problem of state i’s government; the market clearing conditions in capital and fnal goods hold with Ki = Ki ∑ 0 = ∑ 0 = 0 and m nici m Yi , respectively; and P . Setting P = 0 in equation (3.6a) and (3.6b), the emission price chosen by the government of state i reads
⃓ ui ⃓ p0 = −n E ⃓ for all i. (3.7) i i ui ⃓ ci ⃓P =0 In the decentralized case, the optimal local emission price in state i is an increasing function of the domestic population (consumers) size and of the . The MRSE,ci MRSE,ci decreases in capital endowment if the marginal utility is decreasing, ui < 0. Since cici frms set the marginal product of emissions equal to the emission price, the policy of state i internalizes the local damage from domestic emissions and weighs it against the consumption losses from reducing emissions. Formally, we have Y i = p0 = n MRS . Ei i i E,ci According to the Bowen-Lindahl-Samuelson Condition, the optimal level of emission reduction would be achieved if each state sets 0 = ∑ . It follows that in the pi i niMRSE,ci socially optimal case, the state prices would have to be the same, 0 = 0 . Suppose pi p that all utilities are weighted equally in a social welfare function9, then in the case of diferently wealthy states, all ui would need to be equalized by lump-sum transfers from ci rich to poor states (cf. Chichilnisky and Heal, 1994). Therefore, purely decentralized policies as in equation (3.7), fail to consider the spillover efects of emission damages to neighboring states and can be improved upon by a joint federal policy.
3.3.5 Federal policy The aim of the federal government is to improve upon the inefcient decentralized equilibrium by setting a Pareto-improving uniform federal emission price that cannot be Pareto dominated. When the federal price makes no state worse of than the decentralized outcome and at least one state does better, voluntary participation is guaranteed and we consider the federal emission price feasible. Feasibility is facilitated by the recycling of federal revenues, which the federal government distributes according to a transfer rule si (cf. Table 3.1). All federal revenues are distributed so that = ∑ . To capture P E i nisiP E the federal government’s coordinating role, we assume that it anticipates the reaction of
9A distributional statement must be made to determine the socially optimal outcome. Other weights are also admissible. We refer here to equal weights because it is the most simple one to illustrate the argument.
76 3.3 The model
the member states, i.e. the response of consumers, frms, and state governments, acting as a Stackelberg leader for these economies. Formally, Pareto-improving policies are found by maximizing any member state’s welfare subject to Pareto-constraints for the other member states:
{ i ⃓ j 0j } max niu (p(P ),P ) ⃓u (p(P ),P ) ≥ u ∀ j ̸= i (3.8) P ⃓
with ( ) ( ( ) ( ) ( )). Since consumers are identical a state, p P ≡ p1 P , p2 P , ..., pm P within the Lagrangian function related to problem (3.8) simplifes to
( ) Li (P, λ) = ui (p(P ),P ) + ∑ λ uj (p(P ),P ) − u0j (3.9) j̸=i j
where λj̸=i are the m−1 Karush-Kuhn-Tucker multipliers related to the utility constraints in problem (3.8). We provide detailed derivation of the frst order conditions in 3.B.10 j 0j If some P satisfes u > u for all j ≠ i, this implies that λj ∀j̸=i = 0. If such a case exists, which we show to be true in the preceding section, matters would be greatly simplifed, and further analytical insights could be attained. In such a case and using equation (3.2), the federal government’s frst-order conditions are reduced to11
ui dc dE dc 1 dY i d(s E − E /n )P − E = i / with i = + i i i . (3.10) ui dP dP dP n dP dP ci i ⏞ ⏟⏟ ⏞ ⏞ ⏟⏟ ⏞ (a) (b) Equation (3.10) indicates that the federal government sets the of each MRSE,ci consumer in state i (left-hand side) equal to its marginal change in consumption due to P relative to a marginal change in aggregate emissions due to P . Using equation (3.2), we see that the marginal change in consumption comprises of the marginal change of domestic per capita income of a consumer in state i (a) and the net federal transfer to a consumer in state i (b).
Defnition 3 (Multilevel policy Stackelberg equilibrium). The multilevel policy
Stackelberg equilibrium with federal transfer rule si is defned by the quantities cˆi, Yˆi,Kˇi, Eˆi and prices rˆi, pˆi, Pˆ such that cˆi solves the optimization problem of each consumer in state i; Yˆi, Kˇi and Eˆi solve the problem of frm i; pˆi solves the problem of the state government i; Pˆ solves the problem of the federal government; the market
10Because of the Pareto constraints on maximizing utility, our Pareto approach does not add up the utility functions of all states, which would mean that their utilities are comparable. Our approach only compares the utility levels of each state between its decentralized and multilevel outcome. However, with the formulation of the federal problem as in equation (3.8), we make use of a traditional concept: The formulation of a Pareto improvement is equivalent to maximize a social welfare function given specifc weights (cf. Krepps, 1990; Sheeran, 2006). For each minimum level assigned to a consumer j, uj0, when maximizing the utility of consumer i in the Pareto-improvement form as in equation (3.8), there is a set of social welfare weights = 1 with ∑ = 1 which produces the same Pareto µi i , ..., m j µj result when maximizing a social welfare function of all consumers and with the similar transfer rule. 11 To get this result, note that for λj ∀j̸=i = 0, the frst-order condition equals ( i ) ui 1 dY + d(siE−Ei/ni)P = ui dE . Rearranging yields equation (3.10). ci ni dP dP E dP
77 3. Make or Brake
= ∑ = ∑ clearing conditions of capital and fnal goods hold with Ki Ki and m nicˆi m Yˆi, + = respectively; and the balance of payments condition Yˆi nisiPˆEˆ − PˆEˆi nicˆi is satisfed for all i.
3.4 Impact of transfer rules
Having specifed the model and its decision structure, we now proceed to investigate the feasibility of federal policy-making when applying specifc transfer rules. We divide this into an analytical results Section 3.4.1, in which we derive the main argument of this paper, and a numerical Section 3.4.2 in which we explore the results for a plausible parameter space and check the robustness of our main fndings to the simplifying assumptions we made in the analytical part of the paper. 12
3.4.1 Analytical results In order to analytically explore the mechanics of the model developed we make the following simplifying assumptions. First, we assume production by Cobb-Douglas technology i( ) = αK αE . The parameters 0, 0 are the output Y Ki,Ei AKi Ei αK > αE > elasticities of capital and emissions, respectively, with αK + αE = 1, and A > 0 is an efciency parameter. Second, population size is equal across states, i.e. nM ≡ ni = nj. Third, we set i γ u (ci,E) = ci − gE (3.11) where g and γ are constant with g > 0 and γ ≧ 1.13 We report and discuss the reaction functions in 3.C.2 for the Cobb-Douglas production function.
Let κi denote the capital share of state i as a fraction of the total capital in the federation, ∑ , such that Also, let denote the average K ≡ i Ki κi ≡ Ki/K. Kav ≡ K/m capital endowment in the federation and
1 m + γ − αE κEG ≡ unanticipated (3.12a) m 1 + γ − αE 1 m + γ − α − 1 ∗ E anticipated (3.12b) κEG ≡ m 1 + γ − αE − 1/m
We call and ∗ the . We use these assumptions κEG κEG capital-homogeneity-restriction and notation in the propositions to come.
12In this section, we suppose that the production technologies have equal technological parameters across states. 13We recognize that the assumption of linear consumption may seem odd at a frst glance. Combining linear consumption with an emission externality where emissions are an input in production in the presence of a fxed capital stock guarantees a utility function that is concave in emissions and the existence of interior solutions. The resulting optimization problem, thus has features similar to those observed in settings with log-utility or power utility, while maintaining analytical traceability. In our numerical analysis, we assume log-utility. The main fndings remain similar.
78 3.4 Impact of transfer rules
3.4.1.1 Main fndings
Proposition 1 (Juste retour (no inter-state transfers) — anticipated). If si =
1/nM Eˆi/Eˆ (juste retour) and federal transfers are anticipated by the states, then dpi/dP = −1; and the federal government cannot achieve Pareto-improvements relative to the decentralized solution.
Proof. See 3.D. Since the federal policy addresses the efect of transboundary emissions, one would expect that each consumer could be made better of by the federal policy. However, when anticipated, juste retour transfers fail to produce Pareto-improvements. Since dpi/dP = −1, an increase in the federal price leads to a 100% reduction in state prices so that state and federal prices are perfect strategic substitutes in case of anticipated juste retour transfers. In this case, the states fully internalize the vertical fscal externality, but they internalize the transboundary emission externality only to the degree that corresponds to their decentralized policy solution. The consideration of the optimal state price, equation (3.6b), already gives us an idea why anticipated juste retour transfers cannot deliver Pareto-improvements by federal emission policy. The partial derivative of the federal revenue transfer (siPE) with regard to pi is ∂(siP E)/∂pi = P/nM ∂Ei/∂pi. Substituting this result into (3.6b) we get
ui p∗ (P ) = −n E − P. (3.13) i M ui ci Expressing (3.13) in terms of the efective emission price as the sum of state and federal price, we can see that the efective emission price that frm i pays under the juste retour transfer rule equals the one it pays in the decentralized solution, i.e. Y i = p + P = n MRS . This already hints that the policy choices at the state level Ei i M E,ci perfectly ofset the federal policy. We show this in more detail in 3.D. We can conclude that the juste retour criterion renders federal policy inefective and therefore infeasible as soon as it is anticipated by the states. As in Chichilnisky and Heal (1994) and Sandmo (2007), we fnd that Pareto optimality cannot be established in the absence of interstate transfers. In our setting it is also the case that, in the absence of interstate transfers and when states anticipate federal transfers, not even Pareto-improvements are possible, despite the presence of a strong federal government (Stackelberg leader). Therefore, the case of anticipated juste retour transfers highlights that it is important that when a federal government chooses a transfer rule it knows whether its states anticipate the federal transfers.
Proposition 2 (Feasible federal policy). The federal government’s policy leads to a Pareto-improvement relative to the decentralized solution either if
1. si = 1/n (egalitarian), federal transfers are anticipated by the states (unanticipated), ∗ = 1 and κi < κEG (if unanticipated κi < κEG) for all i , ..., m, or
79 3. Make or Brake
2. = 1 0 0 si /nM Ei /E ( sovereignty), and federal transfers are unanticipated, or
3. si = 1/nM Eˆi/Eˆ ( juste retour), and federal transfers are unanticipated.
If also K1 ≤ ... ≤ Km, then the lowest uniform federal price that is not Pareto-dominated m solves maxP u (p(P ),P ).
Proof. See 3.E.
If states anticipate federal sovereignty transfers, we can show the existence of Pareto- improving federal prices for γ = 1:
Proposition 3 (Feasible federal policy: sovereignty — anticipated). Let γ = 1. If = 1 0 0 (sovereignty), and federal transfers are anticipated by the states, si /nM Ei /E then the federal government’s policy leads to a Pareto-improvement relative to the decentralized solution.
Proof. See 3.F.
To provide intuition, let us focus on Proposition2 case i) and explain the federal optimization requirements for the case of m = 2 states. Figure 3.1 illustrates the basic arguments of the underlying proofs. Let i = rich, poor where Kpoor < Krich. Using the state reaction functions, equations (3.6a) and (3.6b), available to the federal government we can express state utility in terms of the federal price P alone, i.e.ui(p, P ) = ui(p (P ) ,P ). At P = 0 the level of ui equals the decentralized utility level u0i. To obtain a maximum for a positive P , two conditions must hold: First, the slope of ui must be positive at P = 0. This is ensured by the capital-homogeneity-restrictions involving ∗ and for egalitarian transfers. Since , only κi < κEG κi < κEG κpoor < κrich the rich state can potentially violate the inequalities or ∗ In κrich < κEG κrich < κEG. terms of interpretation, if these restrictions are not met, the utility of the richest state cannot be improved for any positive federal emission price, which makes voluntary participation impossible. Second, ui must be strictly concave in P to ensure a maximum. Both conditions together imply that there exists a bounded range of positive federal prices P for which both utilities ui are greater than the decentralized level u0i.14 If these conditions hold, then each state i reaches its maximum at P i, P i is thus the preferred uniform federal price of state i. We fnd that the federal price P rich associated with maximizing the utility of the richest state is always the lowest feasible price, while the price P poor associated with maximizing the utility of the poorest state is always the largest price15, P rich < P poor. It may happen, that P poor is not feasible which we will discuss with Corollary1 subsequently.
14 i γ When considering diferences in preferences with u (ci,E) = ci − giE and that gi ≠ gj, we fnd that the federal government is also able to attain Pareto-improvements. This proof is available upon request. 15Maximizing the utility of the poorest state corresponds to the the equity criterion developed by Rawls, also known as maximin criterion (cf. Rawls, 1971).
80 3.4 Impact of transfer rules
The intuition behind the ranking of the federal prices is that the largest burden of the federal policy is carried by the richest state. We defne rich states to be those states that have a capital stock share larger than the capital stock average share (∑ = 1 ), i κi/m /m
i.e. 1/m < κrich. Similarly, for poor states, it holds that κpoor < 1/m. Let Spoor denote
the subset of states with capital endowments shares κi ≡ Ki/K smaller than the average
share (1/m) and let Srich denote the subset of states with κi larger than average.
For egalitarian transfers, state prices across poor and rich states are equal, ppoor = prich (cf. 3.36 and 3.56). Since = , it also follows that 0 n mnM < EiϵSpoor /n < EiϵSpoor /nM < implying the net federal transfers to poor and rich E/n < EiϵSrich /n < EiϵSrich /nM , states satisfy
(E E ) (E E ) − poor P > 0 and − rich P < 0. (3.14) n nM n nM This implies that rich states are net transfer donors, and their beneft from federal policy stems solely from the reduction of emission damage. Poor states, however, beneft from federal policy in two ways. First, the emission price at the federal level reduces the damage of emissions in the same way as for rich states. Second, poor states beneft from being net transfer recipients (cf. also equation (3.2)). We describe the optimal federal price range by again using Figure 3.1. Any federal price smaller than P rich is Pareto-dominated by P rich, as P rich would make every consumer in the federation better of than those prices below P rich. Therefore, P rich = rich is the lowest feasible federal price, i.e. min rich. Let rich 0 defne the price at Pˆ P ≡ Pˆ Pind > which the utility level of the rich state equals the utility level in the decentralized solution, 0rich = rich( ( rich) rich). If the federal price poor associated with maximizing the u u p Pind ,Pind P utility of the poorest state is smaller than rich i.e. poor rich, then the largest feasible Pind , P < Pind federal price is P poor = Pˆpoor since any price larger than Pˆpoor is Pareto-dominated by poor. This case is depicted by the middle dashed line in Figure 3.1. If poor rich, Pˆ P > Pind then the largest feasible federal price is rich as the rich state falls below its decentralized Pˆind utility for any P above that. This case is depicted by the dotted line in Figure 3.1. For the m state case, we generalize this fnding:
Corollary 1 (Feasible federal price range). The federal government’s policy solution space is the interval of uniform federal emission prices that satisfes
[ m min 1 2 m ] P ∈ Pˆ , {Pˆ , Pˆind, ..., Pˆind} .
Based on efciency grounds, Chichilnisky and Heal (1994) fnd that poorer states should be net recipients while richer states should be net donors. Our work links their work to Olson (1986; 1965; 1966) (formalized by Bergstrom et al. (1986)) by accounting for the self-interest of the states. Olson argues that a benevolent hegemonic state is
81 3. Make or Brake
Figure 3.1: Stylized illustration of the proof structure of Propositions2 and3 and Corollary1 for two states i = rich, poor where Kpoor < Krich. The lowest federal emission price, which cannot be Pareto-dominated, corresponds to the utility maximum of the richest state ( Pˆrich). The second vertical dashed line indicated the case when the largest federal price, which cannot be Pareto-dominated, maximizes the utility of the poor state (Pˆpoor). The dotted line and light gray area corresponds to the case in which the rich state falls back to its decentralized solution, such that rich becomes the largest federal price Pˆind which cannot be Pareto-dominated.
82 3.4 Impact of transfer rules
willing to create a multinational system. We show how a multinational (federal) system can be designed such that the richest state is willing to be a member and the largest contributor. In this way we connect to the second generation of fscal federalism, which examines the institutional design of federations as central determinant of their viability (cf. Oates, 2005). Our results show that the richest state simultaneously holds two roles in the feasibility of federal policy. As the largest net donor, the richest state becomes the enabler of federal policy. But precisely for this reason, the richest state prefers the lowest of all federal prices, which also makes the richest state the bottleneck for federal policy stringency. In 3.E, we provide further equations such as the closed form solutions of the federal prices P i in equations (3.49), (3.63), and (3.72).
3.4.1.2 Capital-homogeneity-restriction of egalitarian transfers As Proposition2 indicates, only egalitarian transfers impose a restriction on capital stock diferences across states. We now explore these restrictions in more detail. Recall the capital-homogeneity-restrictions stated in (3.12a) and (3.12b) for the unanticipated and anticipated case, respectively. Both depend on the production elasticity of emissions 16 αE, the externality-elasticity parameter γ and the number of states in the federation m. Boundary cases for the number of states in the federation are m = 2 and m → ∞. Let m = 2. Setting m = 2 in equations (3.12a) and (3.12b) the capital-homogeneity-
restrictions can be expressed in terms of only on γ and αK . Taking the derivative w.r.t. 17 αK and γ we get:
⃓ ⃓ ∂κ ⃓ ∂κ ⃓ EG ⃓ = EG ⃓ 0 unanticipated (3.15a) ⃓ ⃓ < ∂γ ⃓m=2 ∂αK ⃓ ⃓ ⃓ ∂κ∗ ⃓ ∂κ∗ ⃓ EG ⃓ = EG ⃓ 0 anticipated (3.15b) ⃓ ⃓ < ∂γ ⃓m=2 ∂αK ⃓m=2
This indicates that, ceteris paribus, larger αK and γ make the capital-homogeneity- restrictions stricter in the sense that the two states must be more homogeneous in terms of their capital stocks. Let m → ∞. The limit of equations (3.12a) and (3.12b) is
1 ∗ lim κEG = lim κ = (3.16) m→∞ m→∞ EG αK + γ Thus, in the limit, for a very large number of states in the federation, the capital-
homogeneity-restriction approaches 1/(αK +γ) for the anticipated and the unanticipated
case, and is both decreasing in αK and γ. Equation (3.16) indicates that anticipation
16The parameter γ can be interpreted as an elasticity since ∂D/∂E E/D = γ, where D corresponds to the size of the dis-utility from emissions. 17Derivatives can be found in 3.H.
83 3. Make or Brake plays an increasingly smaller role for egalitarian transfers the more states there are in the federation. Equation (3.16) also suggests that it is easier to ensure the voluntary participation of the richest state when αK and γ are low. Ceteris paribus, we give some simple intuition for the result of equations (3.12a),
(3.12b) and (3.16) as all decrease in γ and αK . Since αK = 1 − αE, both equations increase in αE. Note that a state with a large capital endowment carries out more production than a state with a small capital endowment. Therefore, a large capital endowment contributes more to the emission damage but also the state’s gross emission payment to the federation (PEi). All states beneft from emission reductions. Emission reduction occurs through emission pricing. Emission pricing reduces the emission damage but also reduces production volumes in the states. If emissions are more important in production than capital (large αE), emission payments to the federal government decline more strongly. Consequently, in the case of egalitarian transfers where federal revenues are distributed equally, the net donations of the richest state are also lower
(payment minus transfers). Thus, when the capital diference is large, but also αE is large, it becomes easier to fnd an agreeable federal carbon price because then, the capital diferences are less important than the share of emissions in output. If the states are not very vulnerable to emissions (low γ), there is not such a great gain by federal emission mitigation. Thus, the federal price tends to be lower and so are the payments from states to the federation. Therefore, the federal payments of the richest state are lower compared to the case of a large γ and its net donations under egalitarian transfers are less, so that fnding an agreeable federal price is less strongly limited by large capital diferences. We plot the capital-homogeneity-restrictions together with the average capital share
κav ≡ 1/m in Figure 3.2 with αK = 0.97 and γ = 2. The fgure shows that state anticipation decreases the gap between both restrictions. By considering the distance to κav, we also see that the more member states there are, the looser the restrictions become.
3.4.1.3 Transfer anticipation and state policy in multilevel equilibrium Using the state prices from the general case, equations (3.6a) and (3.6b) and using the simplifying assumptions from the beginning of Section 3.4.1 allows us to express equilibrium state prices as follows. Under unanticipated transfers, state prices for the transfer rules egalitarian (EG), sovereignty (SO), and juste retour (JR) are
⃓ ( )γ−1⃓ p = n gγ Eˆ ⃓ for all i and rule = EG, SO, JR ˆi,rule M ⃓ Pˆrule where ∑ . Eˆ ≡ j Eˆj
For the anticipated case note that si,EG = sEG = 1/(mnM ) = κav/nM . Using the transfer defnitions in Table 3.1 and the emission levels from the decentralized
84 3.4 Impact of transfer rules
0.6
κEG−restriction 0.4 κ∗ EG−restriction
Capital−share 0.2
κav
0.0 0 10 20 30 40 50 Number of member states (m)
Figure 3.2: Comparison of capital-homogeneity-restriction for diferent num- bers of member states. To ensure that federal policy is feasible (Pareto-improving) when using egalitarian transfers, the capital share of all states κi with i = 1, ..., m must be below (solid line) in the unanticipated case, or ∗ (dashed line) in anticipated κEG κEG case. κav is the average capital share which equals 1/m.
policy equilibrium, we also get that = 0 ( 0 ) reduces to = = si,SO Ei / E nM si,SO κi/nM 18 Ki/(KnM ). If states anticipate the federal transfers, equation (3.6b), then each state sets its optimal emission prices as follows:
⃓ ∗ = ∗ = ( ∗)γ−1⃓ ∗ egalitarian (3.17a) pˆi,EG pˆEG nM gγ Eˆ ⃓ − κavPˆEG. ⃓ ∗ PˆEG ⃓ ∗ = ( ∗)γ−1⃓ ∗ sovereignity (3.17b) pˆi,SO nM gγ Eˆ ⃓ − κiPˆSO. ⃓ ∗ PˆSO The second term in both equations (3.17a) and (3.17b, ceteris paribus, shows that federal transfer anticipation reduces state i’s domestic emission price compared to the unanticipated case by the federal transfer per unit of emissions its consumers receive
(nM siP ). In this case, each state government takes into account the negative vertical fscal externality, namelythat a local emission price increase leads to a reduction in federal revenues and in turnreduces federal transfers. This observation indicates a potentially undesirable outcome of federal transfer anticipation: In the case of egalitarian transfers, the federal transfer per emission unit received by consumers in a state is higher, the
fewer states are in the federation (because κav = 1/m). For sovereignty transfers, the more wealthy a state is in comparison to other states, the higher the federal transfer per
unit of emissions its consumers receive (measured by κi). In the case of anticipation, therefore, states reduce their domestic emission prices due to their expectation of getting higher federal transfers. As we will highlight in Section 3.4.2, the federal emission price responds to lower state prices and will be higher in the multilevel equilibrium of the anticipated case than in the unanticipated case (cf. Keen, 1998; Keen and Kotsogiannis, 2002, on overtaxation in the non-environmental context). It also implies that states
18This result is obtained by using equations (3.30) and (3.32) and substitution of 0 = and pi ,ni nM P = 0.
85 3. Make or Brake individually refrain from a stronger internalization of emission damages and leave this to the federal government if they subsequently receive more transfers from the federal government.
3.4.1.4 Aggregate emission reduction at the federal minimum price We now analyze the relative emissions mitigation subject to the diferent federal transfers.
We can derive analytical insights only for the unanticipated case and with nM = 1. The more states in the federation, the less the anticipated and unanticipated cases difer (cf. equation (3.16)), which implies that our analytic results here would hold approximately for the case of many member states in a federation. Nevertheless, the assumption nM = 1 is a limiting one and we, therefore, examine the role of the population more closely in our numerical analysis.
Proposition 4 (Aggregate emission reduction). Let K1 < ... < Km, and nM = 1. If i)
κi < κEG; ii) the respective federal minimum price is set; iii) state governments do not anticipate the federal transfer; and iv) κm > αE/γ, then sovereignty and juste retour transfers achieve a higher aggregate emission reduction than egalitarian transfers. If
κm < αE/γ then egalitarian transfers achieve a higher aggregate emission reduction.
Proof. See 3.G.
The inequality in Proposition (4) suggests that κm, αE and γ are critical parameters for assessing which federal transfer rule is superior to the others in terms of aggregate emission reduction. A large capital share of the richest state κm as well as a large elasticity of emission damage γ tend to give sovereignty and juste retour transfers the upper hand, whereas the opposite applies for a large output elasticity with respect to emissions αE. In Section 3.4.2, we numerically investigate the likelihood that sovereignty and juste retour criteria are superior to egalitarian by exploring plausible parameter ranges and relaxing some of our simplifying assumptions. In particular, we explore plausible ranges of αE and γ , consider cases where ni ≠ nj, introduce decreasing marginal utility of consumption and the general CES production function.
3.4.2 Numerical analysis In the formal analysis in Section 3.4.1, we found that the richest state is both the enabler and the bottleneck for federal policy, and that its interests manifest themselves in the federal minimum price. In our numerical extension we complement the notion of absolute wealth of states (in terms of their capital stock) by introducing population diferences, that allow us to consider capital per capita, which is a more pertinent measure of well-being and economic development. Regarding the two diferent notions of a rich state, we fnd in this numerical analysis that in the case of federal egalitarian transfers, the
86 3.4 Impact of transfer rules federal minimum price is determined by the utility maximum of the richest state in terms of capital per capita. In contrast, both under sovereignty transfers and unanticipated juste retour transfers, the minimum price corresponds to the utility maximum of the richest state in terms of the largest capital stock in absolute terms.
3.4.2.1 Assumptions and specifc functional forms We relax the following assumptions: the numerical model distinguishes two types of rich states and two types of poor states, i = poor small, poor large, rich small, rich large.
We fx the capital and population size of the poor small state at Kpoor = 1 and nsmall = 1 and scale the population and capital endowments of the other states accordingly. A consumer in a rich state owns 1.2 times as much capital per capita as a consumer in a poor state. Twice as many people live in large states than in small states.19 Table 3.2 gives an overview of state endowments in our model economy, where each entry in parenthesis refects the population and capital of state i as follows (ni, Ki).
Kpoor Krich nsmall (1, 1) (1, 1.2) nlarge (2, 2) (2, 2.4)
Table 3.2: Population-capital endowment matrix.
i To capture decreasing marginal utility of consumption, we assume u (ci,E) ≡ γ log(ci) − E . Production is modeled by a constant elasticity of substitution technology, σ−1 σ−1 σ i( ) ( σ + σ ) σ−1 where is the substitution elasticity between Y Ki,Ei ≡ αK Ki αEEi σ capital and emissions. Based on own estimates or taken from other studies, we set
αE = 0.03, γ = 2 and σ = 0.5. See details on these estimates in 3.J and reaction functions for the CES production function in 3.I.2. Propositions2 and4 reveal that under egalitarian transfers it matters how rich the richest state is, and that αE and γ are critical parameters for the level of emission mitigation. Since we also introduce population size heterogeneity and decreasing marginal utility of consumption, we provide sensitivity analysis in these fve dimensions, which Table 3.3 summarizes.20
19For the illustration of our analytical fndings, we start with relatively small asymmetries. This assumption allows the broadest discussion of transfer rules as all federal prices that maximize the utility of each state i are feasible (except for juste retour anticipated, which is never feasible). If we take the capital per capita levels of EU Member States and divide these into a poor and a large region then a consumer in the rich EU region owns three times as much capital than a consumer in the poor EU region, based on capital stock data from Berlemann and Wesselhöft (2017) and Berlemann and Wesselhöft (2014) and Census data from the year 2011 code cens_11r provided by Eurostat. Similarly, eight times more people live in large EU states than in small small EU states. 20We report and discuss the reaction functions in 3.I.2 for the CES production function.
87 3. Make or Brake
i parameter/function αE γ Krich nlarge σ u (ci) variation (0, 0.3] [1, 3] [1, 20] [1, 15] [0.4, 0.98] log(ci) and linear
Table 3.3: Sensitivity analysis on parameters and the assumption of linear or log utility in consumption.
Sovereignty* GG GG G G feasible federal prices anticipated*
Sovereignty GGGG G G unanticipated
Juste retour* x size of state G large
G Juste retour GGGG G G small
Egalitarian* G GG GG G capital of state G poor
G rich Egalitarian G GGG GG
1.0 1.5 2.0 2.5
Federal price normalized (P Pnorm)
Figure 3.3: Feasible federal price range from minimum (left) to maximum federal price (right end of gray bar). Each circle corresponds to the normalized uniform federal price which maximizes the utility of consumers in that respective state ( i ). Pˆrule/Pnorm Anticipated juste retour transfers cannot provide Pareto-improvements and are thus infeasible (x). Also compare to Figure 3.1.
3.4.2.2 Results We focus the discussion on results that complement and elaborate the fndings in our analytical analysis. Further sensitivity analyses are provided in 3.K.
Federal price range Figure 3.3 presents the feasible federal prices as price ranges (gray bar) from the minimum (left end) to the maximum price (right end). To ease explaining results, all prices are expressed in relative terms of the lowest minimum federal price across the diferent rules considered. The lowest minimum price corresponds to that of the rich small state under the egalitarian and non anticipated case. Let richsmall. If feasible federal prices exist, the range is non-empty and in all cases Pnorm ≡ PˆEG the minimum federal price corresponds to the utility maximum of a rich state (dark color). If the range is empty, no feasible federal prices exist which is denoted by an x in the fgure and relates to anticipated juste retour transfers and Proposition1. Figure 3.3 confrms the fnding of Proposition2 and complements the fnding of Proposition3, i.e. that the richest state’s utility maximum determines the minimum price. Regarding our two diferent notions of a rich state, i.e. rich in terms of aggregate versus per capita capital, we fnd that in case of egalitarian transfers, the minimum price is determined by the utility maximum of the rich small state (hence the largest
88 3.4 Impact of transfer rules
per capita capital with the lowest population size; small dark circle). In contrast, both under sovereignty transfers and under unanticipated juste retour transfers the minimum price corresponds to the utility maximum of the rich large state, that is the one with the largest capital stock in absolute terms (large dark circle). The intuition is as follows. Ceteris paribus, a rich state naturally has a larger per capita emission level than a poor state, making it a larger per capita gross donor of federal revenues, see also (Roolfs et al., 2018). Under egalitarian transfers each state receives in aggregate a federal transfer in proportion to its population size. Because of that, any small state receives a lower aggregate federal transfer than a large state. Also, the emission damage afects small states less than large states as the damage from emissions afects fewer people. Therefore, the small rich state faces a high federal price burden (because it is rich), but low federal transfer receipts and little environmental benefts (as it has a small population size). Accordingly, the utility of the small rich state is maximized at the lowest of all feasible federal prices. As net donors, rich states might demand that their utility is maximized, and thereby the small rich state becomes the bottleneck for feasible policy under federal egalitarian transfers. For sovereignty transfers, the intuition is relatively similar. Sovereignty transfers, however, distributes federal revenues based on decentralized emission levels and not population size. Ceteris paribus, a large state already sets a higher decentralized state price than a small state as it faces a larger emission damage due to its larger population size. A large state, therefore, reduces its decentralized emissions more than a small state (cf. Section 3.3.4). Consequently, under sovereignty transfers, the utility of the large rich state is maximized at the lowest of all feasible federal prices, and it becomes the bottleneck for a feasible policy. Similarly, this reasoning also holds for unanticipated juste retour transfers, since the outcomes under unanticipated juste retour and unanticipated sovereignty transfers are identical.21 The transfers from these two rules are determined by the multilevel or decentralized emission share levels, respectively. And as they are unanticipated, the state policy choices in the multilevel setting are independent of the transfer anticipation term. Comparing the minimum prices of the anticipated (dashed boxes) with the unanticipated case (solid boxes), it appears that anticipated transfers accommodate a higher minimum price. Intuitively, states that anticipate the transfer, know that local consumption also benefts from the transfer payments from the federal government, which raises their acceptance of higher federal prices. They are also able to include federal revenue recycling in their optimization and are, therefore, fnding it optimal to substitute domestic for federal emission pricing in appreciation also of the revenues,
21We show analytically that results are identical for the Cobb-Douglas case in 3.E.3 and 3.E.4, but this result generally holds also for other constant returns to scale production technologies, simply because the transfer shares of these two transfer rules become identical.
89 3. Make or Brake when balancing consumption against damage reduction. This reveals that states are willing to hand over more regulation to the federal government (higher federal price) if they can anticipate the transfer.
In Figure 3.4 we provide a sensitivity analyses by varying αE and γ . The analysis shows that our results on the richest and poorest states defning the minimum and maximum price, as well as more pronounced coordination under transfer anticipation are robust over wide ranges of αE and γ. A sensitivity analysis for σ is provided in 3.K.
Results are similarly robust as for αE and γ.
Efective state emission prices Figure 3.5 presents the efective (consolidated) emission price that each state i faces, i.e. the sum of the federal and state i emission prices in the multilevel policy equilibrium. Each line starts at the federal minimum price and ends at maximum federal price on the x-axis (cf. Figure 3.3). The bisector (45°-line) in Figure 3.5 helps to compare the federal price (x-axis) to the efective price (y-axis). On the bisector, the state price is zero. Above the bisector, a state complements the federal price with a positive state price. In that case, the efective state price is larger than the federal price. On the other hand, an efective state price below the bisection reveals that a state sets a negative state price and thus subsidizes local emissions. The slope of each line in Figure 3.5 refects the response of each state i to the feasible federal prices. For all transfer rules described in A) and B)22, the states lower their price as the federal price increases, i.e. the slope is smaller than unity. All states lower their state prices only slightly if the transfers are unanticipated (solid lines) and strongly if they are anticipated (dashed lines). In the unanticipated case, the efective price always lies above the bisector, which means that states always complement the federal price with a positive state price. In the anticipated case, states react to a federal price with a stronger state emission price decrease. Thus, feasible federal prices are larger in the anticipated case, but the policy response of states gets more pronounced. The intuition is: If states anticipate the federal transfer, they calculate that setting a large local emission price will decrease federal revenues (internalization of the vertical fscal externality), and as such, they will receive a smaller federal transfer. Therefore, states set a relatively small emission price. In turn, the federal authority, acting as a Stackelberg leader, foresees that states set relatively small prices and to compensate for that the federal government sets a relatively larger federal price. For high federal prices, states’ prices with anticipation can turn into local emission subsidies. Figure 3.5 also shows that for any federal minimum price, regardless of which federal transfer is used, the rich large state always faces the highest efective emission price. The reason is that, due to its large population size, it perceives a relatively larger marginal
22We do not report anticipated juste retour transfers as the federal price range is empty and thus infeasible.
90 3.4 Impact of transfer rules
Sensitivity analysis of federal price range
capital of region G poor G rich size of region G large G small
A) egalitarian not anticipated 1.75 1.75 G GGGG 1.50 1.50 G G G GG G G GG GG GGG G G GG G GG GG G G GG G G GG G G GG 1.25 GG G G GG 1.25 G G G GG G GG GGGGG G G G G GG G G GG GG GGG GG G G GG G G GG GGGG G G GG G G GGG GG G G GGG G G G G G G G G GGGG GG G G GG 1.00 G G G G GGG G 1.00 G G G G G GGGGG G G G G G G GGGGGG G G G G G GGGGGG G G G G G G GGGG G G G G G G G GGGG 0.75 G G G G G G G 0.75 G G G G G G G G G G 0.50 0.50 0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 B) egalitarian anticipated
2.5 2.5
G G G GG G G
) G G GG G G
m G G
r G G 2.0 G G G 2.0 G G o G GG G G n GGG G G G G G GGG G GG G G P G GG G G G GG G G G GG GG GG G G G G G G GG GG G GGG G G G GG GGG P 1.5 GGGG G G G 1.5 G G GG ( GG G G GGGG GG G G G G GG G G G GGG GG G G G G GGG G GGG GG G G G G GG G G GGG GG G G G G GGG G G GGGG GG G G G G G G G GGGG GGG G G G G G G G GGGG GGG G G G GGGG GG 1.0 G G GGGG 1.0 G G G
0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 C) juste retour and sovereignty not anticipated
2.0 2.0 G G G G G G G G G G G G G G G G G G G Federal price normalized Federal G G G 1.5 G G 1.5 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G GGG GG G G G G GGGGGG G GGGG G G G G GGGGG GGGGGGG G G G G GGGG G GGGG G G G GGGGG 1.0 G GGG G G G 1.0 GG GG GGGG G G G GGGGGG G GGGG G G G G GGG GGGG G G G G GGGGG G G G G GGGGGGGG 0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 D) sovereignty anticipated
3 3 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G 2 G G G 2 G G G G G G G G G G G G G G G G G GGGGGG G G G G G GGGG GGG GGGGGGG G G G G G GGGG GGGG GGGGG GGGG G G G G G G GGGG GGG GG GGGGGGGG G G G G G G GGG GGGGG G GGGGGGGGGG G G GGGGGG 1 GGGGGGGGGGG 1
0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 alphaE gamma
Figure 3.4: Robustness test of the feasible federal price range with γ and αE variations. The richest state in terms of capital per capita (under egalitarian) or aggregate capital (under sovereignty or unanticipated juste retour transfers) prefers the lowest federal emission price. The gray dashed lines represent our standard parameter assumption. Variations of σ are provided in 3.K.
91 3. Make or Brake emission damage than small states, while it has a lower marginal utility of consumption than poor states. Therefore it chooses the highest state price on top of the respective federal minimum price. The same reasoning applies to poor small state who sets the lowest state price at the federal minimum price.
A) egalitarian transfers B) sovereignty and unanticipated juste retour transfers 2.5 2.5 ) m r o n region P )
P poor large + i
p poor small
( 2.0 2.0 rich large rich small
1.5 1.5 anticipation no yes
Effective price normalized Effective 1.0 1.0
1.0 1.5 2.0 2.5 1.0 1.5 2.0 2.5
Federal price normalized (P Pnorm)
Figure 3.5: Efective (consolidated) emission price per state given by the sum of state i’s price pi and the federal price P in the feasible federal price range. The dotted vertical line corresponds to Pnorm and the dotted 45°-line is the bisector.
In the following, we discuss some sensitivity analyses and provide further ones in 3.K.
Impact of population size on minimum prices and aggregate emissions In Figure 3.6, we report the A) respective minimum prices and B) federal emission reduction in relation to the decentralized solution with varying population size in large states. While we keep the population size constant for small states, we increase the amount of people living in large states nlarge ∈ [1, 15] which corresponds to 50 to 90% of the federal population living in large states,but keep capital per capita as in Table 3.2. Figure 3.6 A) shows that for small population size diferences, all transfers facilitate relatively similar federal minimum prices. This is due to the fact that a larger population increases the weight of the marginal damage caused by emissions. An increase in the minimum feasible federal price internalizes this larger damage. In the range where all minimum prices increases, it is ambiguous whether the minimum feasible federal price in the anticipated case is higher for egalitarian or sovereignty transfers. When population size diferences get larger, i.e. when roughly 80% or more of the federal population lives in large states, the minimum feasible federal price with egalitarian transfers begins to fall and fnally federal policy becomes infeasible (see drop min to P /Pnorm = 0). As before, the reason for that is that under egalitarian transfers,
92 3.4 Impact of transfer rules the smaller the relative size of the state, the less federal transfers it tends to receive and, therefore, as population diferences increase, at some point, the burden becomes too large for voluntary participation in federal policy. The minimum price with sovereignty transfers and unanticipated juste retour transfers is always feasible and thus independent of population size diference. Here, the minimum federal price keeps increasing with larger population sizes. After looking at the ranking of all governmental emission price choices, a fnal comparison of feasible aggregate emission reduction (i.e. level of public good provision) is in order. In Figure 3.6 B) we compare aggregate emissions reduction achieved by federal policy. To do so, we compute the relative changes between the multilevel policy outcome [ 0 0] at the federal minimum price to the decentralized policy outcome, (Eˆ − E )/E | min . Prule We highlight four fndings presented in 3.6 B) to conclude our numerical analysis. First, the larger the population diferences, the smaller the benefts in terms of emission reduction from federal policy as compared to the decentralized solution. With large population diferences, large states’ decentralized policies alone internalize a relatively large portion of emission damages, cf. equation (3.7). Second, anticipation generally results in lower aggregate emission reduction under sovereignty transfers relative to the decentralized solution. Results are ambiguous for egalitarian transfers and become infeasible when population size diferences get too large (roughly at 80% in B)). Third, when the population in large states is signifcantly larger than in small states (with our model assumptions roughly above 70% of the total population), federal sovereignty transfers are superior in terms of emission reduction than egalitarian transfers. Thus, if diferences in population size are too large, sovereignty transfers are preferable because they always guarantee voluntary participation of states in federal-policy making. Fourth, federal transfer anticipation enables states to internalize their vertical fscal externality on federal revenues. However, this fscal internalization can be at the expense of overall emissions reductions, as in the multilevel equilibrium, the cases of unanticipated federal transfers, in many cases, achieve more overall emissions reductions than the case of anticipated transfers (compare the dashed with the solid lines in B)). This fnding translates to the result of the literature of fscal federalism. Pure models of fscal federalism fnd that states tend to locally overtax if they do not internalize their vertical fscal externality (cf. Keen, 1998; Keen and Kotsogiannis, 2002). In our environmental fscal federalism model, there is a fscal and an environmental externality. In this case, the internalization of the vertical fscal externality of each state government results in a reduction of each state’s emission price. As Figure 3.6 B) shows, the internalization of the vertical fscal externality (anticipation) can go at the expense of the internalization of transboundary emission damages.23
23In 3.K, we report sensitivities of aggregate emissions and consumption changes analysis w.r.t.
αE and γ. There, we also provide a sensitivity analysis with larger capital per capita diferences and compare log to linear utility in consumption (cf. Figure 3.7).
93 3. Make or Brake
anticipation no yes
A) minimum price B) aggregate emission relative to decentralized case
) sovereignty 0 ) m
r 5
0 juste retour egalitarian o E n , P E ( n i sovereignty m
P 4 juste retour ( −10
3
−20
2
sovereignty −30 1 sovereignty juste retour
juste retour egalitarian aggregate emission change % relative
Federal minimum price minimum normalized Federal 0 −40 60 80 100 60 80 100
% of federal population in large states (nlarge n*100)
Figure 3.6: Impact of population size diferences on the size of A) feasible federal minimum prices and B) aggregate emission reduction relative to the decentralized aggregate emission level. Intersection of the y-axis at 0 means that the federal policy is infeasible.
3.5 Conclusion
"Given the slowness and confict involved in achieving a global solution to climate change, recognizing the potential for building a more efective way of reducing greenhouse gas emissions at multiple levels is an important step forward" Ostrom (2009)
If we focus solely on efciency, public good provision such as transboundary emission mitigation requires centralized policy solutions. Still, concerns about excessive burdens on individual actors and regions dominate political reality, which in turn jeopardizes voluntary participation in collective action. A prominent example is the mitigation of climate change in the European Union, where policy is lagging far behind what is required to achieve set targets. Creating a multinational political regime based on the voluntary participation of sovereign nations can serve as an entry point for more ambitious policies. This paper examines the prerequisites for such an entry point in terms of the conditions for voluntary participation in a federal setting with heterogeneous capital stocks across states. Rather than seeking the frst-best solution, which would require an omnipotent central regulator, we study three commonly used distribution rules and a uniform price to extract practically relevant insights for policy design. We allow for coexisting state-federal policy so that member states can conduct companion policies individually tailored to their circumstances.
94 3.5 Conclusion
We show that the existence of feasible uniform federal emission prices depends on the wealth diferences across states, the federal transfer rules, and on whether or not the states anticipate the federal transfers. We fnd that the internalization of vertical fscal externalities (transfer anticipation) can go at the expense of internalizing transboundary emission damages. Our study also fnds that the richest state is always the largest net donor of federal revenues, and its utility is maximized at the lowest Pareto-dominant federal emission price (federal minimum price). Since the interests of the richest state manifest themselves in the federal minimum price, the richest state becomes both the enabler of federal policy (donor) and the bottleneck (brake) on the stringency of federal policy. Last but not least, our simple analytical model proposes a structural approach to assess the willingness of individual states to participate in federal environmental policy. Our results contribute to a better understanding of federal and multinational systems by showing i) how to fnd a federal emission price that ensures voluntary participation of all states; ii) how local and federal environmental policies interact in this context; and iii) which states represent a potential bottleneck for federal policy. As our study shows that the richest state can represent a bottleneck, our fndings provide a starting point for informed negotiations on more ambitious federal environmental policies and regional minimum prices. Since the efectiveness of federal policy can be hampered if states adjust their policies in anticipation of the transfer, federal policymakers are better advised to use egalitarian or sovereignty transfers rather than juste retour transfers. Moreover, in case of large diferences in capital per capita across the states, sovereignty transfers are preferable because they always guarantee voluntary participation. There are many ways to extend the current analysis. First, an obvious extension would be to examine the efectiveness of other transfer rules and how they fare against an appropriately specifed social optimum. Second, the efects of mobile capital should be examined to see whether the results from the setting with immobile capital remain valid. Mobile capital becomes particularly relevant if one considers not only short-term entry points but also long-term stable policy-making based on voluntary participation. Third, low-emission sectors could be introduced to allow parts of the production being relocated to less emission-intensive industries. Fourth, states’ membership in the federation could be endogenized, so that a federation of a fexible size can emerge from diferent federal policies, instead of just either having a federal policy or not having one. Fifth, one could change the structure of governmental decisions such that all governments are Nash-players. Sixth, one could investigate the impact of populations being mobile across states, or, what could be even more interesting, consider immigration from outside the federation. Migration could indeed have a substantial impact on the feasibility of federal policy, as it can change the patterns of capital per capita across states and the impact of environmental damages.
95 3. Make or Brake
Acknowledgments and funding
The manuscript has largely benefted from comments of three anonymous reviewers and Till Requate. The authors are also grateful for helpful comments and support of Geir Asheim, Dallas Burtraw, Rick van der Ploeg, Oliver Schenker, Emmanuelle Taugourdeau, and Cees Withagen. We thank Kerstin Burghaus, Jacob Edenhofer, Max Franks, Brigitte Knopf, Kai Lessmann, Linus Mattauch, Thang Dao Nguyen, Michael Pahle, Gregor Schwerhof, and Boyan Yanovski for their helpful comments on previous versions of this manuscript. Funding by the DFG [grant number LE 782/2-1] is gratefully acknowledged. Christina Roolfs’ work was supported by the German Federal Ministry of Education and Research [grant number FKZ 03EK3523B]. The conclusions expressed in the manuscript do not necessarily represent the views of the above-mentioned institutions.
96 REFERENCES
References
Barrett, S. (2005). The theory of international environmental agreements. In K. G. Mäler & J. R. Vincent (Eds.), Handbook of Environmental Economics (pp. 1457–1516). Elsevier. BEA. Industry Economic Accounts, The Use of Commodities by Industry after Redefnitions (1987–2006). Bureau of Economic Analysis. 2008. Bergstrom, T., Blume, L., & Varian, H. (1986). On the private provision of public goods. Journal of Public Economics, 29, 25–49. Berlemann, M., & Wesselhöft, J. (2017). Aggregate Capital Stock Estimations for 122 Countries: An Update. Review of Economics, 68 (2), 1158. Berlemann, M., & Wesselhöft, J. (2014). Estimating aggregate capital stocks using perpetual inventory method - new empirical evidence for 103 countries -: Working Paper Series. Review of Economics, 65, 1–34. Boadway, R., & Keen, M. (1996). Efciency and the optimal direction of federal-state transfers. International Tax and Public Finance, 3, 137–55. Boadway, R., Marchand, M., & Vigneault, M. (1998). The consequences of overlapping tax bases for redistribution and public spending in a federation. Journal of Public Economics, 68 (3), 453–78. Böhringer, C., Rivers, N., Rutherford, T., & Wigle, R. (2015). Sharing the burden for climate change mitigation in the Canadian federation. Canadian Journal of Economics, 48 (4), 1350–1380. Böhringer, C., Rivers, N., & Yonezawa, H. (2016). Vertical fscal externalities and the environment. Journal of Environmental Economics and Management, 77, 51–74. Bruellhart, M., & Jametti, M. (2006). Vertical versus horizontal tax externalities: An empirical test. Journal of Public Economics, 90 (10), 2027–2062. Buchholz, W., Haupt, A., & Peters, W. (2013). International Environmental Agreements, Fiscal Federalism, and Constitutional Design. Review of International Economics, 21 (4), 705–718. Burtraw, D., Keyes, A., & Zetterberg, L. Companion Policies under Capped Systems and Implications for Efciency—The North American Experience and Lessons in the EU Context (RFF Report, Ed.). Ed. by RFF Report. 2018. Burtraw, D., & Toman, M. (1992). Equity and International Agreements for CO2 Containment. Journal of Energy Engineering, 118 (2), 122–135. Cazorla, M., & Toman, M. (2001). International Equity and Climate Change Policy. Climate change Economics and Policy, 235–247. Chichilnisky, G., & Heal, G. (Eds.). (2000). Environmental Markets. Equity and Efciency. Columbia University Press. Chichilnisky, G., & Heal, G. (1994). Who should abate carbon emissions? Economics Letters, 44, 443–449.
97 3. Make or Brake
Cramton, P., Ockenfels, A., & Stoft, S. (2015). An International Carbon-Price Commitment Promotes Cooperation. Economics of Energy & Environmental Policy, 4 (2), 51–64. Dahlby, B. (1996). Fiscal externalities and the design of intergovernmental grants. International Tax and Public Finance, 3, 397–412. Dahlby, B. (2008). The marginal cost of public funds: Theory and applications. Dahlby, B., & Wilson, L. (2003). Vertical fscal externalities in a federation. Journal of Public Economics, 87 (5-6), 917–930. Dalmazzone, S. (2006). Decentralization and the environment. In E. Ahmad & G. Brosio (Eds.), Handbook of Fiscal Federalism (pp. 459–477). Edward Elgar Publishing. d’Autumne, A., Schubert, K., & Withagen, C. A. (2016). Should the carbon price be the same in all countries? Jounal of Public Economic Theory, 18, 709–725. Destatis. (2017). VGR des Bundes - Input-Output-Rechnung - Fachserie 18 Reihe 2 - 2013 (Revision 2014, Stand: August 2016). Federal Statistical Ofce Germany. Dietz, S., & Stern, N. (2015). Endogenous Growth, Convexity of Damage and Climate Risk: How Nordhaus’ Framework Supports Deep Cuts in Carbon Emissions. The Economic Journal, 125 (583), 574–620. EC. (2015). EU ETS Auctioning. European Commission [Accessed 6-April-2017]. EC. (2013). The EU Emissions Trading System (EU ETS). Factsheet. European Commission. Edenhofer, O., Roolfs, C., Gaitan, B., Nahmmacher, P., & Flachsland, C. (2017). Agreeing on an EU ETS price foor to foster solidarity, subsidiarity and efciency in the EU. In I. W. Parry, K. Pittel, & H. Vollebergh (Eds.), Energy Tax and Regulatory Policy in Europe: Reform Priorities. MIT press. Engström, G., & Gars, J. (2015). Optimal Taxation in the Macroeconomics of Climate Change. Annual Review of Resource Economics, 7 (1), 127–150. FOEN. (2016). Swiss federal ofce for the environment [Accessed 6-October-2016]. Grubb, M., Sebenius, J., Magalhaes, A., & Subak, S. (1992). Sharing the burden. In I. M. Mintzer (Ed.), Confronting Climate Change: Risks, Implications and Responses (pp. 305–322). Cambridge University Press. Gruber, L. (2000). Ruling the World: Power Politics and the Rise of Supranational Institutions. Princeton University Press. Hsiang, S., Kopp, R., Jina, A., Rising, J., Delgado, M., Mohan, S., Rasmussen, D. J., Muir-Wood, R., Wilson, P., Oppenheimer, M., Larsen, K., & Houser, T. (2017). Estimating economic damage from climate change in the United States. Science, 356 (6345), 1362–1369. Keen, M. (1998). Vertical Tax Externalities in the Theory of Fiscal Federalism. IMF Staf Papers, 45. Keen, M., & Kotsogiannis, C. (2002). Does federalism lead to excessively high taxes? The American Economic Review, 92 (1), 363–370.
98 REFERENCES
Kemfert, C., & Welsch, H. (2000). Energy-Capital-Labor Substitution and the Economic Efects of CO2 Abatement: Evidence for Germany. Journal of Policy Modeling, 22 (6), 641–660. Klenert, D., Mattauch, L., Combet, E., Edenhofer, O., Hepburn, C., Rafaty, R., & Stern, N. (2018). Making carbon pricing work for citizens. Nature Climate Change, 8 (8), 669–677. Krepps, D. (1990). A Course in Microeconomic Theory. Princeton University Press. Kverndokk, S. (2018). Climate Policies, Distributional Efects and Transfers Between Rich and Poor Countries. International Review of Environmental and Resource Economics, 12 (2-3), 129–176. Kverndokk, S., & Rose, A. (2008). Equity and Justice in Global Warming Policy. International Review of Environmental and Resource Economics(2), 135–176. Lau, L., & Yotopoulos, P. (1972). Proft, Supply, and Factor Demand Functions. American Journal of Agricultural Economics, 54 (1), 11–18. Manne, A., & Richels, R. (1992). Buying Greenhouse Insurance: The Economic Costs of CO2 Emission Limits (1st ed., Vol. 1). The MIT Press. Musgrave, R. (1959). The Theory of Public Finance. McGraw-Hill. Oates, W. (1972). An Essay on Fiscal Federalism. Journal of Economic Literature, 37, 1120–1149. Oates, W. (2000). Fiscal competition and European Union: contrasting perspectives. Regional science & urban economics, 31, 133–145. Oates, W. (2005). Toward A Second-Generation Theory of Fiscal Federalism. Interna- tional Tax and Public Finance, 12 (4), 349–373. Olson, M. (1986). A Theory of the Incentives Facing Political Organizations. Neo- Corporatism and the Hegemonic State. International Political Science Review, 7 (2), 165–189. https://doi.org/https://doi.org/10.1177/019251218600700205 Olson, M. (1965). The Logic of Collective Action: Public Goods and the Theory of Groups (Vol. 20 printing, 2002). Harvard University Press. Olson, M., & Zeckhauser, R. (1966). An Economic Theory of Alliances. The Review of Economics and Statistics, 266–279. Ostrom, E. A Polycentric Approach for Coping with Climate Change: Policy Research Working Paper 5095. 2009. Paterson, M. (2001). Principles of justice in the context of global climate change. In U. Luterbacher & D. Sprinz (Eds.), International Relations and Global Climate Change (pp. 119–126). MIT press. Posner, E. A., & Sunstein, C. R. Justice and Climate Change: Discussion Paper 2008-04. Cambridge MA, 2008. Pottier, A., Méjean, A., & Godard, O. (2017). A Survey of Global Climate Justice: From Negotiation Stances to Moral Stakes and Back. International Review of Environmental and Resource Economics, 11 (1), 1–53.
99 3. Make or Brake
Rawls, J. (1971). A theory of justice. Oxford University Press. Ringius, L., Torvanger, A., & Underdal, A. (2002). Burden Sharing and Fairness Princi- ples in International Climate Policy. International Environmental Agreements(2), 1–22. Roolfs, C., Gaitan, B., & Edenhofer, O. (2018). Assessing EU minimum prices and revenue distribution from a multilevel policy perspective. WCERE 2019 conference paper. Rose, A. (1992). Equity considerations of tradable carbon emission entitlements. In United Nations Conference on Trade and Development) (Ed.), Combating global warming: study on a global system of tradeable carbon emission entitlements. (pp. 55–83). United Nations. Rose, A., & Stevens, B. (1993). The efciency and equity of marketable permits for C02 emissions. Resource and Energy Economics, 15 (117-146). Rose, A., Stevens, B., Edmonds, J., & Wise, M. (1998). International Equity and Difer- entiation in Global Warming Policy. Environmental and Resource Economics, 12, 25–51. Sandmo, A. (2004). Environmental taxation and revenue for development. In A. Atkinson (Ed.), New Sources of Development Finance. Oxford University Press. Sandmo, A. (2007). Global Public Economics: Public Goods and Externalities: mis en ligne le 17 octobre 2007, consulté le 30 septembre 2016. Economie publique / Public economics [En ligne], 18-19. Sheeran, K. (2006). Who Should Abate Carbon Emissions? A Note. Environmental and Resource Economics, 35 (2), 89–98. Shiell, L. (2003). Equity and efciency in international markets for pollution permits. Journal of Environmental Economics and Management, 46 (1), 38–51. Stavins, R. N. (1997). Policy Instruments for Climate Change: How Can National Governments Address a Global Problem? University of Chicago Legal Forum, Article 10, 1997. Talus, K. (2013). EU Energy Law and Policy. A Critical Account. Oxford University Press. van der Werf, E. Production Functions for Climate Policy Modeling: An Empirical Analysis, Nota di Lavoro, NOTA DI LAVORO 47.2007. 2007. Warleigh, A. (2004). European Union: The basics. Routledge. https://doi.org/10.4324/ 9780203331620 Weitzman, M. L. (2014). Can Negotiating a Uniform Carbon Price Help to Internalize the Global Warming Externality? Journal of the Association of Environmental and Resource Economists, 1 (1/2), 29–49. Wiener, J. B. (2007). Think Globally, Act Globally: The Limits of Local Climate Policies. Univ PA Law Rev, (155), 1961–1979. Williams, R. (2012). Growing state–federal conficts in environmental policy: The role of market-based regulation. Journal of Public Economics, 96 (11-12), 1092–1099.
100 REFERENCES
Wilson, J. (2006). Tax competition in a federal setting. In E. Ahmad & G. Brosio (Eds.), Handbook of Fiscal Federalism. Edward Elgar Publishing.
101 3. Make or Brake
Appendix
Appendix 3.A First-order conditions of states
( ) Note that Y i K ,E = p + P solves for E as a function of p and P , which we denote Ei i i i i i i with bold letters by Ei (pi,P ). Moreover, since Y (Ki,Ei) is homogeneous of degree one, i i the cost of producing Y is linear in output. Profts, thus, equal (1 − mci (ri, pi,P )) Y i where mci (ri, pi,P ) is marginal cost of producing Y . Zero profts in turn, imply mci = 1, which solves for ri as a function of pi and P , which we denote ri (pi,P ). Similarly, i ( ) = i ( ( )) and ( ) = ∑ ( ) with = ( ) Y pi,P Y Ki, Ei pi,P E p, P i Ei pi,P p p1, ..., pm . Consumption can thus be written as 1 ( i ( ) ) ci (p, P ) = Y Ki, Ei − P Ei + siEP. (3.18) ni Since ( ) = ∑ ( ) and = the derivative of (3.2) with E p, P i Ei pi,P ∂E/∂pi ∂Ei/∂pi, regards to pi yields
∂c 1 ( ) ∂E ∂s P E i = Y i − P i + i i (3.19) Ei ∂pi ni ∂pi ∂pi and since Y i = p + P then Ei i
( ) ∂ci = pi ∂Ei + ∂ siP Ei (3.20) ∂pi ni ∂pi ∂pi
i i Maximizing niu (p, P ) = niu (ci (p, P ) , E (p, P )) with regard to pi, state i sets
( ∂c ∂E ) n ui = n ui i + ui = 0, (3.21) i pi i ci E ∂pi ∂pi which implies that
∂c ∂E ui = ui i + ui = 0. (3.22) pi ci E ∂pi ∂pi Substitution of (3.20) and using (3.96) yields
∂E ( p ∂E ∂(s P E )) ui = ui i + ui i i + i i = 0 (3.23) pi E ci ∂pi ni ∂pi ∂pi for all i = 1, ..., m. The unanticipated case* is derived by setting ∂(siP Ei)/∂pi = 0. These m frst order conditions (one per state) implicitly defne pi as a function of P , which we denote ( ). pi P
102 3.B Stackelberg-Leader’s frst order conditions
Appendix 3.B Stackelberg-Leader’s frst order con- ditions
Let = ( ( ) ( )) Using equation (3.2) and the reaction function of frms, p p1 P , ..., pm P . consumers and states (indicated with bold letters) from the market clearing and state problem (indicated with bold letters), the frst-order condition when maximizing the utility of state i subject to uj (p(P ),P ) ≧ u0j is
( j ) m dE m 1 d Y − P Ej ds EP − ∑ λ uj = ∑ λ uj + j (3.24) j=1 j E j=1 j cj dP nj dP dP with λi = 1. For any function Fi(p, P ) its total derivative with respect to P equals dp dF i = ∂F i dpi + ∂F i and for G(p, P ) is dG = ∑ ∂Gj j + ∂G and dP ∂pi dP ∂P dP j ∂pj dP ∂P
j 0j λj(u (p,P ) − u ) = 0 for all j ̸= i . (3.25)
Appendix 3.C Cobb-Douglas technology
3.C.1 Firm’s problem Suppose the production function is represented by a Cobb-Douglas technology, i( ) = αK αE . The objective of frm reads Y Ki,Ei AKi Ei i { ⃓ } max ( ( + ) ) ⃓ i = αK αE (3.26) Yi − riKi − pi P Ei ⃓Y AKi Ei . Ki,Ei
The parameters αK > 0, αE > 0 are the output elasticities of capital and emissions, respectively, with + = 1, and 0 is an efciency parameter. Let Ω = αK αE . αK αE A > αK αE A
The marginal cost (mci) of producing good Yi equals
= αK ( + )αE Ω (3.27) mci ri pi P / .
Zero profts imply mci = 1. The frst order conditions of frm i also imply conditional demand levels of capital and emissions as follows;
Ki = αK Yi/ri and Ei = αEYi/ (pi + P ) . (3.28)
3.C.2 Market clearing and reaction functions of frms and consumers
Note that bold letters indicate functional forms solely deepening on emission prices. Substituting equation(3.27) into the zero proft condition (mci = 1) and solving for ri,
103 3. Make or Brake we obtain ( ) 1 Ω αK ( ) = (3.29) ri pi,P αE . (pi + P ) ri is clearly decreasing in (pi + P ) , refecting that if pi or P increase, the remuneration that frms can make to the owners of capital must decrease. Using equation (3.28), α setting i = K αE and solving for it follows that Y AKi Ei Ei
1 ( ) α αEA K Ei (pi,P ) = Ki. (3.30) pi + P and in turn ( α ) 1 α E A αK i ( ) = E (3.31) Y pi,P αE Ki. (pi + P ) Clearly, the return to capital, output and emissions of state i decrease with the per unit cost of emissions pi + P . Aggregate emissions equal
( ) 1 m α A αK E (p, P ) = ∑ E KαK . (3.32) j=1 j pj + P Consumption reads
( ) Y i Ei ci (p, P ) = + siE − P. (3.33) ni ni
Equations equations (3.29) − (3.33), defned in terms of pi for i = 1, ...m and P , are known to all governments and allow them to derive the reaction functions of consumers and frms. Appendix 3.D Proof juste retour anticipated*
Using equation (3.33) and (3.97), and the juste retour transfer criterion (Table 3.1) with anticipation state governments set
1−σ σ 1 (pi + P )A σ ci = Ei. (3.34) ni αE
The m frst order conditions of all states (one per state) form a squared system of equations in terms of pi + P for i = 1, ...m. This system of equations simultaneously solves for pi + P in terms of exogenous parameters captured by hi for i = 1, ...m. Solving for pi implies pi = hi − P implying that dp i = −1 for all i = 1, ..., m. (3.35) dP Hence the state governments react with a state price decrease of proportionally one unit in response to the federal government’s price. This refects a full reversal of the states prices as a reply to the federal price. Moreover one can readily verify that all the equations of the decentralized solution are equal to those of the juste retour with anticipation by setting 0 = + . pi pi P
104 3.E Proofs Proposition2
Appendix 3.E Proofs Proposition2
For ease of reading, we indicate functions with bold letters but drop the dependencies. We use the assumptions as introduced in Section 3.4.1. If not indicated otherwise the following proofs all go, if possible, through the same steps: Step 1: We express the indirect utility function depending solely on P and calculate the frst-order conditions. Step 2: We solve for the federal price P that maximizes the utility of state i, and denote this by P i. Step 3: We evaluate the slope of the utility function of all consumers at P = 0. If it has a positive slope there are Pareto-improving federal prices. We then prove that each indirect utility function is concave in P and thus P i globally maximizes the utility of consumer i . Step 4: We rank the prices P 1,P 2, ..., P m. 3.E.1 Egalitarian unanticipated
Step 1 If state government i does not anticipate the federal transfer this implies that ∂ (siP E) /∂pi = 0. Thus, the emission price of state i is
= γ−1 for = 1 . (3.36) pi nM gγE i , ..., m
Note that pi = pj. Substituting pi from equation (3.36) into equation (3.32) , we get
1 ( ) α = αEA K (3.37) E γ−1 K. nM gγE + P from (3.30) and (3.32) it follows that
Ei = κiE, (3.38)
Substituting for from (3 38) into i = αK αE we get Ei . Y AKi Ei
( E )αE Y = AK . (3.39) i i K
Using (3.38) and (3.39) implies
1 ( )αE ( ) ( i ) Ki E κi ci = Y − P Ei + siEP = A + si − P E. (3.40) nM nM K nM
Rearranging equation (3.37) for P we get
(K )αK P = α A − n gγEγ−1 (3.41) E E M
105 3. Make or Brake
Let U i denote the indirect utility function depending solely on P . Substituting (3.41) into (3.40), the indirect utility function U i equals ( κ ) i i αK αE γ U = A αK + αEsi K E − ((sinM − κi) γ + 1) gE . (3.42) nM Step 2 i Setting si = 1/ (mnM ) = κav/nM , U equals A i αK αE γ U = (αEκav + αK κi) K E − ((κav − κi) γ + 1) gE . (3.43) nM
If for some P and some j all the constraints U j̸=i ≧ u0j̸=i are not binding (that is,U j̸=i > u0j̸=i), then the federal government’s frst-order condition equals
dU i dE = Z =! 0, (3.44) dP i dP where Zi reads
( )αK αEA K γ−1 Zi = (χi − θi) − χigγE (3.45) nM E with
χi = 1 + (κav − κi) γ and θi = χi − (αEκav + αK κi) (3.46)
From the frst-order condition in equation (3.44) follows that either Zi or dE/dP or both must equal zero. Implicit diferentiation of equation (3.37) leads to
dE = E (3.47) − ( )αK . dP K + ( 1) γ−1 αEαK A E nM gγ γ − E
Since αK , αE, A, and g are positive, γ ≥ 1, and E > 0, the numerator and denominator of the right-hand side are positive. It follows that dE/dP < 0.Thus, Zi must equal zero to satisfy the federal frst-order condition (3.44). i Let E denote the federation’s aggregate emissions E that makes Zi equal to zero.To avoid introducing more notation we used Ei to denote the federation’s aggregate emissions that maximize the utility of state i while Ei denotes state i ’s emissions. We set Zi = 0 and solve equation (3.45) for E = Ei which gives
( ) 1 ( ) 1 α A α κ + α κ γ−αE α A χ − θ γ−αE Ei = E E av K i KαK = E i i KαK . (3.48) nM gγ 1 + (κav − κi) γ nM gγ χi
Let P i defne the federal price P that maximizes the utility of a consumer in state i. Substitution of equation (3.48) into equation (3.41) and after some manipulations gives
( ) γ−1 ( ) αK γ−αE γ−αE i αEA αK nM gγ P = θi K . (3.49) χi χi − θi We proceed to show that P i, if it exists, must be positive. Step 3
106 3.E Proofs Proposition2
Let l denote the subset of (low wealth or “poor”) states with capital endowments shares smaller than the average share ( = ∑ = 1 ) and let κi ≡ Ki/K κav i κi/m /m h denote the subset of (high wealth) states with κi larger than average. For χi∈l and θi∈l from equation (3.46), follows that
χi∈l > 1 and θi∈l > 0 and χi∈l−θi∈l > 0. (3.50) Together with equation (3.49), it follows that any P i∈l > 0. Let us examine the behavior of U i∈l on the interval [0,P i∈l) by evaluating the slope i∈l of U at P = 0. We know from equation (3.47) that dE/dP < 0. Substitution of χi, and into from equation (3.45) and some algebraic manipulations yields θi E|P =0 Zi
γ−1⃓ Zi|P =0 = − θigγE ⃓ . (3.51) ⃓P =0 Since the parameters of equation (3 51) are positive for , it follows that 0. . i ∈ l Zi∈l|P =0 < As dE/dP < 0, it follows from equation (3.44) that U i∈l has a positive slope at P = 0. Implying that if U i is concave, only positive federal prices can make poorer states better of relative to the decentralized solution, whereas negative federal prices make them worse of. Consequently, if there is a role for the federal government, then any feasibleP must be positive.
Let us examine what P > 0 implies for states in set h where κav = 1/m < κi∈h. To ensure a Pareto improvement via P > 0 for all i ∈ h the slope of U i∈h must increase at P = 0. As in Proposition (2) let
m + γ − αE κi < κav for i = 1, ..., m (3.52) 1 + γ − αE
i ⃓ then θi∈h > 0 and therefore Zi∈h| < 0, implying that dU /dP ⃓ > 0 for all states P =0 ⃓P =0 i = 1, ..., m. This also implies that for a range of positive federal prices constraints U j ≧ u0j are not binding (U j > u0j). We now prove that U i decreases on the interval (P i, ∞). Let P b > P i and evaluate the slope of equation (3.44) at P b. Using equation (3.48) we get
( )γ−αE α A (χ − θ ) Ei = E i i KαK . (3.53) nM gγ χi
Since dE/dP < 0, then P b > P i implies
⃓ αEA (χi − θi) α Eγ−αE ⃓ < (Ei)γ−αE = K K . (3.54) ⃓ b P =P nM gγ χi Rearranging yields
( )αK α A (χ − θ ) K ⃓ 0 < E i i − χ gγ Eγ−1⃓ . (3.55) i ⃓ b nM E P =P The right-hand side of equation (3 55) is nothing other than and hence . Zi|P =P b i ⃓ Zi| b > 0 implying that dU /dP ⃓ = ZidE/dP | b < 0 for all i. This proves P =P ⃓P =P b P =P that U i is a concave function with maximum P i > 0. Step 4
107 3. Make or Brake
We now rank the diferent P is for i = 1, ..., m. From equation (3.48) one can readily verify that ∂Ei > 0, and from equation (3.47) follows that the higher Ei is the lower P i ∂κi must be. Therefore, the federal prices rank P m < ... < P 1 . 3.E.2 Egalitarian anticipated*
All else is equal as in (3.E.1), except for the assumption that state governments anticipate the federal transfer. If not mentioned explicitly, the steps are similar to the previous proof such that we only provide the equations without description. We omit the asterisk *. Step 1 P p = n gγEγ−1 − (3.56) i M m
Note that pi = pj.
1 αK αK αEAK E = (3.57) γ−1 + (1 1 ) ) nM gγE − m P
α (αK m − 1) κi + αE ((mκi − 1) ) U i = AK K EαE + γ − 1 gEγ (3.58) n − nM m − 1 Step 2
dU i dE = Z =! 0 (3.59) dP i dP where
( ( )αK ) nχi − nM K χi − θi − κiκav γ−1 Zi = mαEA − gγE (3.60) n − nM E nχi − nM and
dE m − 1 E = − α < 0. (3.61) dP m ( K ) K + ( 1) γ−1 αEαK A E nM gγ γ − E
Solving Zi = 0 for E yields
( ) 1 mα A χ − θ − κ κ γ−αE Ei = E i i i av KαK . (3.62) gγ nχi − nM Substituting equation (3.62) into equation (3.57) and solving for P leads to
γ−1 ( ) αK ( ) (α A) γ−αE κ (nχ − n ) γ−αE χ − θ − κ κ P i = E gγ av i M Kγ−1 1 − i i i av . (3.63) (m − 1)κav χi − θi − κiκav χi − κav Step 3 Evaluating Zi at P = 0 yields gγn ⃓ = ( + ( 1) ) γ−1⃓ (3.64) Zi|P =0 − θi κi − κav E ⃓ . n − nM ⃓P =0
108 3.E Proofs Proposition2
Substitute θi from (3.46) to get
0 < θi + (κi − 1)κav = 1 + (γ − αE)(κav − κi) + κiκav − (κi + κav) for i ∈ l. ⏞ ⏟⏟ ⏞ ⏞ ⏟⏟ ⏞ >1 <1 Thus, for follows that 0. Just as argued in the previous proof, it must i ∈ l Zi∈l|P =0 < be that P i > 0 for i ∈ l. Let
m − αE + γ − 1 κi < κav for i = 1, ..., m. (3.65) 1 − αE + γ − κav then also for it follows that + ( 1) 0 and consequently 0 for i ∈ h θi κi − κav > Zi|P =0 < all i. Take equation (3.62) and since dE/dP < 0 and P b > P i, then
⃓ ( )γ−αE αEA χi − θi − κiκav α Eγ−αE ⃓ < Ei = K K . (3.66) ⃓ b P gγnM χi − κav Rearranging, we get
nχ − n ( ( K )αK χ − θ − κ κ ) 0 = i M i i i av γ−1 (3.67) < Z|P b mαEA − gγ E|P b n − nM E|P b nχi − nM and hence 0. Therefore, it follows that i is a concave function with a unique Zi|P b > U maximum at P i > 0. Step 4 The P is can be ranked by considering
i i + ( 1) ∂E = n − nM E αK κav γ − (3.68) ∂κi γ − αE χi − θi − κiκav nχi − nM
From (3.62) it follows that the product (χi − θi − κiκav)(nχi − nM ) is positive and i m 1 ∂E /∂κi is therefore positive. Just as in the previous proof P < ... < P follows.
3.E.3 Sovereignty unanticipated
Note that decentralized state prices equal 0 = ( 0) γ−1. Hence, the pi nM gγ E decentralized state emission prices are all equal and the ratio of state-i’s emissions to aggregate federal emissions equals . In turn, implying that SO = 0 ( 0 ) = . κi si Ei / E nM κi/nM Step 1 Almost all equations are similar to those in Proof 3.E.1, precisely equations (3.36)- (3.38). Consumption and utility equal κ i αK αE ci = AK E nM
κi α U i = AK K EαE − gEγ (3.69) nM
109 3. Make or Brake
Step 2
dU i dE = Z = 0 (3.70) dP i dP ( )αK κi K γ−1 where Zi = αEA − gγE and dE/dP < 0 as it equals equation (3.47). Thus, nM E i Zi must equal zero. Setting Zi = 0 and solving for E we get
( ) 1 α A κ γ−αE Ei = E i KαK . (3.71) gγ nM By substituting the right-hand side of Ei for E in equation (3.37) and into (3.41) , we get 1 ( γ−1 ( )αK ) γ−α i ( αK ) gγnM E P = (1 − κi) αEAK . (3.72) κi Note that all terms in equation (3.72) are positive. Thus, prices that solve the federal government’s problem exist and are independent of any restriction or capital heterogeneity constraints. Step 3 i Consider P on the interval [0,P ). Since dE/dP < 0, using Zi from equation (3.70) and substituting E0 we get
= (1 ) ( )γ−1 0 (3.73) Zi|P =0 − − κi gγ E|P =0 < . i ⃓ i i Thus, dU /dP ⃓ = ZidE/dP | > 0 and U increases on the interval [0,P ). ⃓P =0 P =0 Consider P on the interval (P i, ∞). From equation (3.71) follows that
( )γ−αE α A κ Ei = E i KαK . (3.74) gγ nM ⃓ γ−α Since dE/dP < 0 it follows that P b > P i implies Eγ−αE ⃓ < (Ei) E . Using ⃓P =P b equation (3.74) we get
( )αK ⃓ κ K ⃓ 0 i γ−1⃓ (3.75) < αEA − gγE ⃓ . nM E ⃓P =P b The right-hand side of equation (3 75) is , implying i 0 in the interval . Zi|P =P b dU /dP < (P i, ∞). Hence U i is a concave function with a unique maximum at P i > 0. Step 4 i Consider equation (3.71)and calculate ∂E /∂κi to see that federal prices rank as P m < ... < P 1. 3.E.4 Juste retour unanticipated Note that
1 Ei κi si = = . (3.76) nM E nM and therefore the solution reduces to that of the sovereignty transfer rule under the unanticipated case, refer to (3.E.3).
110 3.F Sovereignty anticipated*
Appendix 3.F Sovereignty anticipated*
All else is equal as in 3.E.3, except for the assumption that each state government anticipates the federal transfer and. Let γ = 1. Step 1 Using (3.6b) and SO = 0 ( 0 ) = state prices equal si Ei / E nM κi/nM
= (3.77) pi nM g − κiP and therefore 1 ( ) α αEA K Ei = Ki, (3.78) nM g + (1 − κi) P rearranging yields
( )αK Ei = αEA (3.79) Ki nM g + (1 − κi) P or ( ( )αK ) Ki 1 P = αEA − nM g . (3.80) Ei 1 − κi Aggregate emissions equal
1 ( ) α = ∑ αEA K (3.81) E + (1 ) Kj j nM g − κj P and consumption (using SO = 0 ( 0 ) = ) reads si Ei / E nM κi/nM 1 ci = (Y i + (κiE − Ei) P ) (3.82) nM Using = ( + ) and = we get Yi pi P Ei/αE pi nM g − κiP ( ) 1 nM g + (αK − κi) P U i = Ei + (κiP − nM g) E (3.83) nM αE Step 2 Taking the derivative of U i and using (3.79) we get
dU i 1 ∑ (1 − κj) Ej = κiE + (−gnM + κiP ) α (3.84) j ( Kj ) K dP nM AαEαK Ej ( )αK Ei gnM + A (αK − κi) αK − κi Ki + − Ei. (3.85) αE AαEαK
Step 3 Evaluate i at = 0 by substituting 0 and 0 to get dU /dP P Ei E ⃓ dU i ⃓ E0 ⃓ = ∑ (1 − κ ) κ > 0. ⃓ j̸=i j j dP ⃓P =0 nM αK
111 3. Make or Brake
Appendix 3.G Proof of Proposition4
Let nM = 1 and K1 < ... < Km. Suppose that αE/γ < κi. Consider the defnitions of χi and θi given by χi = 1 + (κav − κi) γ and θi = χi − (αEκav + αK κi). After some algebraic manipulations, we get that αE/γ < κi implies
χi − θi κi < . (3.86) χi Consider the emission levels in closed form with unanticipated federal transfers from equations (3.48), and (3.71) to see that if inequality (3.86) holds, then i = i i . ESO EJR < EEG Proceed in a similar manner to prove that implies i = i i αE/γ > κi ESO EJR > EEG. Appendix 3.H Capital-homogeneity-share consider- ation
⃓ ⃓ ∂κ ⃓ ∂κ ⃓ 1 1 EG ⃓ = EG ⃓ = − < 0 unanticipated (3.87a) ⃓ ⃓ 2 ( + )2 ∂γ ⃓m=2 ∂αK ⃓m=2 αK γ ⃓ ⃓ ∂κ∗ ⃓ ∂κ∗ ⃓ 1 EG ⃓ = EG ⃓ = − < 0 anticipated (3.87b) ⃓ ⃓ (2( + ) 1)2 ∂γ ⃓m=2 ∂αK ⃓m=2 αK γ −
Appendix 3.I General CES-function
3.I.1 Firm’s problem Suppose the production function is represented by a CES technology. The objective of frm i reads
{ ⃓ σ } ⃓ ( σ−1 σ−1 ) σ−1 max ( ( + ) ) ⃓ i = σ + σ (3.88) Yi − riKi − pi P Ei ⃓Y A αK Ki αEEi Ki,Ei ⃓ where parameters αK > 0, αE > 0 are the output elasticities of capital and emissions, respectively, with αK + αE = 1, σ is the substitution elasticity, and A > 0 is an efciency parameter. Due to zero profts marginal cost (mci) of producing good Yi equals
1 = 1 = ( σ 1−σ + σ ( + )1−σ) 1−σ −1 (3.89) mci αK ri αE pi P A . The frst order conditions of frm i also imply conditional demand levels of capital and emissions as follows;
σ−1 σ σ−1 σ αK A σ i αEA σ i Ki = Y and Ei = Y . (3.90) ri pi + P
112 3.J Estimation of αE, γ and σ
3.I.2 Market clearing and reaction functions of frms and consumers From equation (3.89) follows that
0 σ 1−σ = 1−σ σ ( + )1−σ ( ) (3.91) < αK ri A − αE pi P ≡ ϕi pi,P and σ (1 ) ∂ϕi = ∂ϕi = αE − σ 0 (3.92) − σ < . ∂pi ∂P (pi + P )
1 ( ) 1−σ ( ) = ϕi (3.93) ri pi,P σ αK Using the market clearing condition for capital and Y i = p + P implies Ei i
σ ( )σ ( ) 1−σ αE ϕi Ei (pi,P ) = Ki (3.94) pi + P αK and
( )σ ( ) σ m α ϕ 1−σ E (p, P ) = ∑ E j K . (3.95) j=1 j pj + P αK These equations correspond to the reaction functions of frms and consumers. Since i ϕi decreases with the per unit cost of emissions pi + P , also ri, Y , Eiand E decrease in pi + P , and
1−σ ∂Ei = ∂E = A σEi 0 (3.96) −( + ) < . ∂pi ∂pi pi P ϕi σ ( σ−1 σ−1 ) σ−1 Substituting from (3.94) into i = σ + σ output can be Ei Y A αK Ki αEEi
written as a function , of pi, and P as follows
1−σ σ ( + ) σ i pi P A Y (pi,P ) = Ei(pi,P ). (3.97) αE
Appendix 3.J Estimation of αE, γ and σ
We estimated a proxy for αE in two ways. Since the model assumes perfect competition, αE represents the share of emissions on output. Our frst estimate is the ratio between aggregate expenditure on carbon emission generating raw inputs denoted COG (coal mining and oil and gas extraction), and net value added plus COG as reported in the BEA (2008) 2007 input-output tables for the US. The second estimate is the ratio between aggregate expenditure on emission generating processed inputs denoted PI (petroleum refneries, manufactured petroleum and coal products, petrochemical and gas manufacturing) and net value added plus PI, also from BEA (2008). Both procedures lead to the same estimate of 0.042. Following the same procedure but using German data for the year 2013 from Destatis (2017) leads to an estimate of 0.027 for the former and 0.02 for the latter.
113 3. Make or Brake
The actual dis-utility or damage from emissions and climate change is still subject to ongoing research. Recent studies come up with regional estimates, for instance Hsiang et al. (2017) fnds the value of damages in the US to be quadratically increasing in global mean temperature. In theoretical models, climate change damages were often assumed to be linear or quadratic (e.g. Buchholz et al., 2013; Dietz and Stern, 2015), largely for reasons of analytical tractability. Therefore, we report the sensitivity for γ ranging from 1 to 3. van der Werf (2007), Manne and Richels (1992) and Kemfert and Welsch (2000) fnd the elasticity of substitution between energy and a composite input with capital to range from 0 − 0.7.
Appendix 3.K Sensitivity analysis
3.K.1 Linear vs log utility and capital stock diferences
In Figure 3.7, we vary the capital stock of rich states from Krich ∈ [1, 10] and report the respective minimum prices. In Figure A), we assume log(ci) and in B) we suppose utility to be linear in consumption. Comparison of A) and B) shows that the observed capital- homogeneity-restriction for egalitarian transfer of Section 3.H vanishes, when supposing log(ci), while when supposing linear consumption, egalitarian transfers make federal pricing with increasing Krich soon infeasible. That means, that in case of log(ci) the efect of consumption saturation dominates the restriction from capital stock diferences
(for nlarge = 2). In this case, rich consumers hardly gain from consumption increases in contrast to emission mitigation and are thus are willing to accept larger federal prices. .
anticipation no yes
A) log utility in consumption B) linear utility in consumption ) 4 4
norm sovereignty P
min sovereignty 3 juste retour 3
2 2 sovereignty egalitarian sovereignty juste retour 1 1
0 0 egalitarian Federal minimum price minimum normalized (P Federal 0 5 10 0 5 10
capital per capita in rich states normalized (Krich Kpoor)
Figure 3.7: Sensitivity analysis of capital stock in rich states. Contrasting log to linear consumption utility.
114 3.K Sensitivity analysis
3.K.2 Robustness test for sigma-variation
Sensitivity analysis of federal price range
capital of region G poor G rich size of region G large G small
A) egalitarian not anticipated B) egalitarian anticipated
1.6 1.6
G G G G G G G G G G G G G G G G G G 1.2 G 1.2 G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G 0.8 G 0.8 G G G GG G G G G G G G G G ) G G
m G G G G G r G G
o G G G n G G G G G G G P G G G G G G G G G G G G G P G G G G
( G G G GG G 0.4 0.4
0.4 0.6 0.8 0.4 0.6 0.8 C) juste retour and sovereignty not anticipated D) sovereignty anticipated Federal price normalized Federal G G G G G G G G G G G G G 1.0 G G 1.0 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G 0.5 G G 0.5 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G
0.4 0.6 0.8 0.4 0.6 0.8 sigma sigma
Figure 3.8: Robustness test of the feasible federal price range with σ. The richest state in terms of capital per capita (under egalitarian) or aggregate capital (under sovereignty or unanticipated juste retour transfers) prefers the lowest federal emission price. The gray dashed lines represent our standard parameter assumption.
3.K.3 Consumption changes at minimum prices
We report consumption changes relative to the decentralized outcome for poor and rich states in Figures 3.9 and 3.10, respectively. Consumption changes always decline in αE and increase in γ.
115 3. Make or Brake
Consumption changes in poor states A) poor small state
0.0 0.0
G
G G G G −0.1 −0.1 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G ) G
0 i −0.2 G −0.2 G G c G G G G G G
0 i G G G G G
c G G −0.3 G G −0.3 G G G G G − G G G G
i G G G G G G G G c G G G G G G
( G −0.4 G G −0.4 G G G G
−0.5 −0.5 0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 B) poor large state
0.0 0.0 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G −0.1 G G −0.1 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G −0.2 G G G −0.2 G G G G
Relative consumption change Relative −0.3 −0.3
−0.4 −0.4
−0.5 −0.5 0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 alphaE gamma
transfer G EG JR SO anticipation no yes
Figure 3.9: Sensitivity analysis of consumption changes evaluated at the related minimum prices relative to the decentralized levels in poor states.
Consumption changes in rich state A) rich small state
0.0 0.0
G −0.1 −0.1 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G ) G G G G G G G
0 i −0.2 −0.2 G G G G G G
c G G G G G G G 0 i G G G c −0.3 G G −0.3 G G G G G
− G G G G G i G G G G G G G c G G G G G G ( G G G −0.4 G G −0.4 G G G G G G G G G −0.5 −0.5 0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 B) rich large state
0.0 0.0 G G G G G G G G G G G G G G G G G G G G −0.1 G −0.1 G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G −0.2 G G G G G G −0.2 G G G G G G G G G G G G
Relative consumption change Relative −0.3 −0.3
−0.4 −0.4
−0.5 −0.5 0.0 0.1 0.2 0.3 1.0 1.5 2.0 2.5 3.0 alphaE gamma
transfer G EG JR SO anticipation no yes
Figure 3.10: Sensitivity analysis of consumption changes evaluated at the related minimum prices relative to the decentralized levels in rich states.
116 3.K Sensitivity analysis
anticipation no yes
1.05 1.05 1.05 m r o n P e n l i u m r P less emissions with egalitarian ) at 1.00 1.00 1.00 G
E less emissions with sovereignty E and not anticipated juste retour O S E (
0.95 0.95 0.95 Ratio of federal emissions Ratio of federal
0.90 0.90 0.90
0.0 0.1 0.2 0.3 0.4 0.6 0.8 1.0 1.5 2.0 2.5 3.0 alphaE sigma gamma
Figure 3.11: Ratio of aggregate federal emissions subject to transfer rule at each minimum price with varying αE, σ and γ. In case of unanticipated transfers, juste retour transfers perform similar to sovereignty transfers (SO). Egalitarian transfers are denoted with EG.
3.K.4 Aggregate emission levels at minimum prices We plot the ratio of aggregate emission levels at the respective minimum price and transfer rules in Figure 3.11. When transfers are anticipated, egalitarian transfers facilitate lower aggregate emissions than sovereignty transfers. When transfers are unanticipated, sovereignty and juste retour transfers are superior in terms of aggregate emission reduction.
117
4 Technology beats capital — Sharing the carbon price burden in federal Europe
Christina Roolfs Beatriz Gaitan Ottmar Edenhofer Kai Lessmann
Submitted to and under review at Journal of Public Economics
119 4. Technology Beats Capital
Abstract Passing federal environmental policy reform is a challenge as the approval of interest groups such as consumers and state-level governments is often a prerequisite. Among others, the burden sharing’s progressivity has a large impact on reform approval. We investigate how carbon tax payments by states to a federal authority are infuenced by diferences in technological emission intensity and wealth and show how they can turn out to be at the expense of poor states. We show that a uniform federal carbon tax that is endorsed by all states with equal per capita transfers can theoretically put a higher burden on poorer states than richer states. The opposite applies for transfers based on historical emissions (sovereignty transfers) which reduce the burden of emission-intensive states. We test our results numerically in a general equilibrium model with a vertical federalism governance structure calibrated to the European Union. Our simulations show that a federal minimum emissions tax with sovereignty transfers is twice as high as for equal per capita transfers and also has a progressive efect.
JEL-classifcation: H77, H23, Q58, H87, H62
Keywords: Emission Regulation, Federalism, Unanimity, Transfers, Pareto-improving policy, European Union
Reference: Roolfs, Gaitan, Edenhofer, and Lessmann (under review at JPubE): Technology Beats Capital - Sharing the Carbon Price Burden in Federal Europe (February 6, 2020). Working paper available at SSRN: http://dx.doi.org/10.2139/ssrn.3533251
120 4.1 Introduction
4.1 Introduction
When French President Macron unveiled his fuel tax reform plans to support the French energy transition in autumn 2018, he triggered one of the fercest protests France has seen in 50 years. Demonstrators took to the streets fearing the heavy tax burden on workers and the middle class. After months of protests President Macron suspended his reform, conceded that he had failed to sufciently take into account the burden for low-income earners and promised to improve his reform proposal. The French example is a case in point for the argument to make policies progressive so as to reduce social tensions, avert social resistance, and make their social support more likely (e.g. Chiroleau-Assouline and Fodha, 2014; Fullerton, 2016; Sterner and Robinson, 2018). Much like civil protests at the national level, unanimity rules and veto power of sovereign nations can be an obstacle to new policies in supranational, federal systems such as the European Union (EU). Poland, for example, a country with relatively low income and a carbon intensive energy sector, has vetoed the European Commission’s Energy Tax reform proposal several times. Since EU tax matters require unanimous consent of all member states, the EU energy tax reform is pending since 2011. Studies of the burden and redistribution of carbon pricing and its revenues have so far focused on the single country setting. Within one country, equal per capita revenue recycling make carbon pricing progressive (Boyce and Riddle, 2007; Burtraw et al., 2009; Klenert and Mattauch, 2016; Klenert, Mattauch, et al., 2018; Rausch et al., 2010). Less attention has been paid to the carbon price burden from the point of view of a federal system and the implications of inter-jurisdictional heterogeneities (but see Böhringer, Rivers, Rutherford, et al., 2015; Böhringer, Rivers, and Yonezawa, 2016). With multiple levels of government, federal policies can interact and confict with state interests but also provide an additional way for welfare improvements. We ask how uniform carbon pricing and the distribution of its revenues afect acceptance and burden in a federal policy system. Our research design follows three aspects of the federal structure of EU climate
policy: First, EU member states’ asymmetries of capital per capita and CO2 intensity of production technology run contrary to each other (Figure 4.1). We include both asymmetries to capture their distributional implications. Secondly, many EU policies require unanimity or majority voting. We hence constrain federal choices to Pareto- improving policies. Third, in contrast to the equal per capita transfers frequently considered in the single country context, most of the revenue from the EU Emissions Trading Scheme (ETS) is distributed to members on the basis of the sovereignty rule,
121 4. Technology Beats Capital
Figure 4.1: Capital per capita and CO2-emission intensities in EU member states for the year 2014. The correlation line and coefcient (R) show that capital-poor countries tend to have greater emission intensities. Data based on World Development Indicators (WDI) and Berlemann and Wesselhöft (2017) and Berlemann and Wesselhöft (2014). See detailed description in Section 4.4.1. i.e. on the basis of the historical emission levels. We include both distribution rules in our analysis of the federal carbon price burden.1 The frst part of our analysis develops an intuition how the burden of a uniform federal CO2 price is afected by technological emission intensity and wealth diferences in closed form. The implications of the diferences across member states are intuitive: We show how a large capital stock but also a large emission intensity result in a higher federal tax burden and prove that high emission-intensity in production causes high emission level, if gross complementarity of production inputs is strong and cost shares of capital and labor are high. The opposite is true with low cost shares for capital and labor and better substitution possibilities. The intuition is that since low wealth is associated with high emission intensities for many EU member states, a uniform federal carbon price absent of revenue recycling threatens to place a greater tax burden on poorer states. We investigate federal tax burden adjustments by the redistribution of federal tax revenues. We fnd that an equal per capita redistribution makes the federal carbon tax regressive. Mobile capital reinforces the regressive efect. In contrast, the sovereignty transfer relieves emission-intensive states from their federal tax burden and hence the federal emission tax becomes progressive.
1See on these rules also Cazorla and Toman (2001), Kverndokk (2018), and Kverndokk and Rose (2008). In particular, the sovereignty rule immediately raises the question of justice and historical accountability of large historical emitters and has been discussed extensively elsewhere in the literature. We refer the interested reader to Cazorla and Toman (2001), Grubb et al. (1992), Ju et al. (2019), and Kverndokk and Rose (2008) to mention at least some. Zhou and Wang (2016) gives an overview of the allocation rules that have been studied for many analyses of multinational emission reductions.
122 4.2 Literature
In the second part our analysis, we use numerical simulations to solve the full multilevel equilibrium of the model with strategic state policies and diferences in labor and population size heterogeneities in addition to capital and technological diferences. The properties of three distinct groups among the European member states are decisive for our numerical results, namely being rich in terms of capital per capita while (1) being large in population size or (2) small in population size, or (3) being poor in capital
per capita and small in population size. We fnd that in terms of CO2-intensity, groups 1 (“rich and large”) and 2 (“rich and small”) fall below the average EU country, whereas the “small and poor” group countries have a higher emission intensity than the EU average. We fnd that the net burden of a uniform federal emission tax given equal per capita transfers largely falls on the group of “rich and small” countries and several “poor and small”countries while “rich and large”countries may even face a net beneft. Thus the policy package is regressive. In contrast, sovereignty transfers achieve more progressiveness as the net burden is by and large carried by consumers of countries belonging to group “rich and large”. The sovereignty transfers rule also facilitates a Pareto-improving minimum federal tax level double that of the equal per capital transfers rule. The remainder of the paper is structured as follows. In Section 4.2, we relate our paper’s contribution to the existing literature on progressive environmental policy design and fscal federalism. In Section 4.3 we describe the structure of the federal general equilibrium model. Subsection 4.3.2 conveys the reasoning behind the results that we produce with our more complex numerical model for the EU. Subsection 4.3.3 expands the model set-up to the federal multilevel policy architecture. Section 4.4 contains the numerical simulation. We conclude in Section 4.5.
4.2 Literature
The design of progressive environmental policy from a multinational (and fscal federalism) perspective explores the intersection of environmental taxation and fscal federalism. This section reviews relevant contributions from the respective strands of literature in turn. Previous literature on the distributional implications of environmental taxes and transfer design has frequently focused on the single country context exploring diferent income groups within an economy and policies choices of a single national government. A literature overview is provided by Klenert, Mattauch, et al. (2018). Many empirical studies fnd carbon pricing (absent of revenue redistribution) to be regressive for OECD countries (Dorband et al., 2019). When revenue recycling schemes are considered, the literature is ambiguous about the progressive impact of such policies, as Klenert and Mattauch (2016) point out, except for equal per capita revenue recycling, which tends to
123 4. Technology Beats Capital make carbon pricing progressive (Boyce and Riddle, 2007; Burtraw et al., 2009; Klenert and Mattauch, 2016; Klenert, Mattauch, et al., 2018; Rausch et al., 2010). But even with a clear understanding of welfare and distributional implications, national environmental policies are restricted by concerns about their efect on mobile tax bases and transboundary spillovers (e.g. Musgrave, 1959; Oates, 1972; Oates, 2000)2. The fscal federalism literature hence argues that, in general, relief of poor households and provision of public goods with spillover efects across states, as it is the case for mitigating
CO2-emissions, should be carried out by the federal authority. The EU in particular, however, may not be very well equipped to relieve poor households, as it does not have sufcient power and budget for redistribution (Oates, 2000). So while an EU carbon pricing scheme exists by means of the EU ETS, the adoption of mainly sovereignty based revenue recycling to member states suggests that little or no redistribution was intended or was simply not feasible. For a literature overview of the related environmental fscal federalism we refer to Böhringer, Rivers, and Yonezawa (2016) and Oates (2001). Strategic interaction of federal and state-level environmental policies are addressed in a recent line of literature observing that federal environmental policy is often counter- acted or overruled by state policies (see e.g. Böhringer, Rivers, and Yonezawa, 2016; Knopf et al., 2014; Lutsey and Sperling, 2008; Williams, 2012). In this line, Williams (2012) uses an analytical framework to compare the efciency of emission policies when strategic federal and state policies regulate emissions simultaneously. He fnds the superiority of a federal emission tax over federal quantity controls. A decisive prerequisite for this result is the availability of optimal revenue recycling by means of optimal transfers to the states. Our modeling approach builds on a federal-state policy structure similar to Williams (2012) but extends the analysis by constraining federal regulation to Pareto-improving policy packages. The focus on Pareto-improvements connects our research to the literature on voluntary public good provision and the analytical model of Bergstrom et al. (1986). They show that rich individuals would voluntarily donate more to public good provision than poor individuals. We use a model with a strategic multilevel policy structure which has been developed in a simpler version and absent of technological, and labor size diferences in Roolfs et al. (2018b). Their method identifes a range of possible voluntary (unanimity-ensuring) federal uniform carbon prices, consisting of a minimum and a maximum federal price. Two further studies have investigated the impact of environmental policies for the Canadian Federation. Böhringer, Rivers, and Yonezawa (2016) analyze the unilateral state incentive for overruling federal emission regulation due to the efects of regulating the same tax base on their budgets. In contrast to the fully endogenous policy setting of Williams (2012), they focus on exogenously given policy choices to comply with given
2In fact, there are several narratives for state’s inefciency regarding transboundary spillovers: one is that state governments simply ignore or cannot measure their spillover impact on other states (Oates, 2001). Another one is that states would engage in a ’race to the bottom’ in their environmental policy to attract mobile factors (Zodrow and Mieszkowski, 1986).
124 4.3 The model
emission targets. Their study shows that a state has an incentive to ofset its emission pricing cost to other states by means of VFE. In addition to an analytical treatment, they apply a computational general equilibrium model to the Canadian Federation and fnd that the VFE enable a state to reduce its emissions by up to 20% without bearing the costs itself. In a preceding paper Böhringer, Rivers, Rutherford, et al. (2015) leave out provincial carbon pricing, and examine the federal emission price burden given diferent transfer rules (e.g. equal per capita and sovereignty transfers) across Canadian provinces. They fnd that the burden on the population in the various provinces varies greatly depending on diferent transfer rules and provincial heterogeneity: equal per capita transfers are most burdensome in provinces with high GDP per capita while they are least burdensome for several provinces with low GDP per capita. They fnd emission intensity heterogeneity to be an important factor determining carbon burden sharing. Based on Böhringer, Rivers, Rutherford, et al. (2015)’s fndings, we can argue that equal per capita transfers support a progressive federal carbon price in Canada. Canadian provinces, however, have a positive correlation between GDP and carbon intensity of production while EU states have a negative correlation. For the case of the EU, our study is – to the best of our knowledge – the frst to consider multilevel EU environmental policies with strategic policy choices on all levels. Previous studies have studied the tax burden of i) energy and fuel/transport taxation (for instance Padilla and Roca (2004), and Cambridge Econometrics (2008) as cited in Kosonen (2012)), and ii) emission mitigation and transfer rules (Böhringer and Lange, 2003; Chiroleau-Assouline and Fodha, 2014). Böhringer and Lange (2005), for instance, take the emission reduction commitment from the Kyoto Protocol. Their paper is in a similar spirit as Böhringer, Rivers, Rutherford, et al. (2015) but compares overall costs of emission mitigation subject to diferent transfer rules. They fnd that diferent transfer rules have very diferent efects on overall costs. While the previously mentioned EU-related papers use numerical or econometric models, Chiroleau-Assouline and Fodha (2014) employ an analytical model to the EU context. To comply with EU tax matters, they aim at unanimity-ensuring (Pareto-improving) environmental policy. They fnd that an environmental tax for the EU can always be designed to be unanimity-ensuring (Pareto-improving) if its revenue is used for a wage tax reform. In the following we will combine the insights and recommendations for action from the above-mentioned literature strands. We will show that the federal context leads to a situation where equal per capita transfers no longer have a fundamentally progressive efect.
4.3 The model
We consider a general equilibrium model of a federation. The federation consists of
i = 1, ..., m member states. Member state i is populated by Li consumers. Consumers
125 4. Technology Beats Capital are immobile across states and rent out their labor to the domestic frm. We consider the case of immobile and perfectly mobile capital. If capital is immobile, then consumers only rent out their capital endowment to the domestic frm. If capital is mobile, then consumers can rent out their capital endowment to any frm i = 1, ..., m. Consumers own the atmosphere in which frms store harmful emissions. Governments enforce consumer property rights through emission taxation, so frms pay for polluting the atmosphere. The redistribution of tax payments is stipulated on transfer rules which we will specify below. Each consumer derives utility from consuming a private good and dis-utility from a transboundary emission externality. To produce the private good, the representative frm in state i uses capital, labor and emissions. We suppose an emission augmenting factor to describe the emission efciency of each state’s production technology. The lower its emission efciency, the larger its emission intensity. In addition to heterogeneity in the emission efciency among states, we allow for heterogeneity of capital stocks and population sizes. Federal and state governments set emission taxes. States tax domestic emissions in order to maximize the utility of domestic consumers. State emission tax revenues are returned uniformly to domestic consumers. The federal authority seeks to improve the utility of all consumers living in the federation by choosing a uniform federal tax that satisfes an environmental policy package as follows: i) federal revenues are recycled following a predefned transfer rule (equal per capita or sovereignty), ii) the federal policy package must achieve Pareto improvements relative to the sovereign state outcome3, and iii) comprises only of Pareto-dominant solutions. Thus, if there are solutions to the federal objective, they are second best optima. Moreover, these solutions theoretically ensure the unanimity of member states towards federal policy-making. The decision structure is as follows. In the frst stage, the federal authority acts as the Stackelberg-leader and searches a uniform tax on federal emissions to deliver Pareto-improvements relative to the decentralized (sovereign) state policy outcome as described in the preceding paragraph. In the second stage, each state government non-cooperatively sets a tax on its state emissions taking all other taxes as given. In the third stage consumers and frms solve their optimization problem, taking all prices, taxes and transfers as given.
4.3.1 Economic agents This section formulates the consumers’ and frms’ problems, the damage (dis-utility) and source (production) of emissions and the market clearing conditions. Thereby, it solves the third stage.
3A sovereign state outcome is the outcome that would prevail when states regulate emissions absent of federal policy.
126 4.3 The model
4.3.1.1 Consumers and revenue recycling
In state i live Li identical working consumers. Each consumer in state i is endowed
with capital ki and one unit of labor li (that is li = 1). Since consumers are immobile across states, each consumer in state i rents out its labor to the domestic frm i. When capital is immobile, each consumer in state i rents out its capital endowment to the domestic frm i. If capital is perfectly mobile, each consumer in state i can rent its capital endowment to any frm i = 1, ..., m. In addition, each consumer receives transfers from the recycling of state and federal emission pricing revenues. Each consumer in state
i receives an equal share 1/Li of the revenues from state i’s domestic emission pricing
(tiEi). Revenues from federal emission pricing (TE) are distributed to each consumer in
state i by the federal transfer rule Si. The budget constraint of each consumer in statei equals
Ei ci = riki + wili + ti + SiTE (4.1) Li
where ri and wi respectively denote the rental rate of capital and labor wage rate. Ei denotes the emissions of state and ∑m are aggregate federal emissions. i E ≡ i=1 Ei ti denotes the emission tax rate levied by state, T is the uniform federal emission tax rate
levied by the federal authority and Si is the federal transfer rule. Consumer i derives utility from private good consumption and dis-utility from federal emissions, i u (ci,E) (4.2) which is assumed to be additively separable with frst and second partial derivatives w.r.t. consumption and emissions being ui > 0, ui ≤ 0, and ui < 0 and ui ≤ 0, ci cici E EE respectively. All consumers take emissions, prices, taxes and transfers as given such that the solution to each consumer’s optimization problem reduces to setting consumption equal to income from endowments and transfers, equation (4.1).
4.3.1.2 Firms and emission efciency
In each state i, a representative frm i produces a fnal private good Yi. Emissions are a by product of production which we treat as an input in production. Firm technologies can difer in the emission augmenting factor which we use to describe frm i’s emission
efciency, χi > 0. We set χi in such a manner so that ∂Yi/∂χi > 0, i.e. a greater level of
χi corresponds to a less emission intensive production technology. The production technology is represented by a two-layered function, supposing constant returns to scale. At the top-layer, a constant elasticity of substitution (CES)
127 4. Technology Beats Capital
function combines emissions Ei with a second-layer Cobb-Douglas capital-labor composite input Vi to produce fnal output Yi. The production function is depicted by
σ ( σ−1 σ−1 ) σ−1 = i ( ) = σ + ( ) σ (4.3) Yi Y Vi,Ei A αVi z χiEi with = i ( ) = βK βL are positive efciency parameters. The elasticity Vi V Ki,Li BKi Li . A, B 4 of substitution between Vi and Ei is denoted by σ with 0 < σ ≤ 1. Distribution parameters {α, z, βK , βL} ∈ (0, 1) and satisfy α + z = 1 and βK + βL = 1.
Taking as given the rental rate of capital ri, the wage rate wi and emission tax
τi(ti,T ) ≡ ti + T frm i maximizes profts by choosing capital Ki, labor Li and emissions . Let i denote the marginal product of input and treat the price of the fnal good Ei YX X as numéraire. Proft maximization implies Y i = τ and Y i = p where p denotes the Ei i Vi i i composite price of composite input Vi.
We solve frm i’s conditional demand for Ki,Li, and Ei as functions of output and prices by solving the cost minimization problem of frm i (see conditional demand levels in 4.A). Zero profts imply
σ 1 1 z 1−σ = σ 1−σ + 1−σ = 1 (4.4) mci α pi 1−σ τi , A σ χi where mci denotes the marginal cost of producing output Yi, which is decreasing in the augmenting factor χi.
As we shall see below, the emission augmenting factor χi ambiguously impacts conditional emission demand and thereby, diferences in emission intensities can infuence a state’s emission tax base and impact the federal tax incidence on a state.
4.3.1.3 Market clearing
Aggregate capital and labor supply in state i are Ki ≡ Liki. Labor market clearing is given by Li = Li. In the case of immobile capital, capital markets clear with Ki = Ki. If capital is perfectly mobile across states, capital market clearing implies ∑ = ∑ i Ki i Ki and in such case the rental rate of capital is equal for all states, ri = rj = r. Market clearing in fnal goods is given by ∑m = ∑m . Using the market clearing i=1 Lici i=1 Yi conditions all variables can be expressed as a function of state and federal taxes. Let bold letters indicate these functions which take into account the solutions (frst-order conditions) of consumers’ and frms’ problems. We report the relevant variables in 4.B and 4.C.
4We undertake this assumption since emissions are generally proportional to energy usage. van der Werf (2007), Manne and Richels (1992) and Kemfert and Welsch (2000) estimate the elasticity of the substitution between energy and the composite input to be between 0 and 0.7, see also Carraro et al. (2011).
128 4.3 The model
4.3.2 Technology and capital impact Now that we have characterized all supply and demand plans and all market equilibria, and before we introduce the multilevel policy architecture, we develop intuition about our main results, namely how emission intensity and capital wealth afect the burden from policy packages.
Since the federal authority levies the emission tax payment TEi from the representa- tive frm in state i, and redistributes its tax revenues (TE) so that each consumer in state i receives SiTE, then the per capita net payment N i from state i to the federal authority equals
( ) Ei N i ≡ − SiE T. Li
If N i > 0 state i is a net donor of federal emission tax revenues. On the contrary, if
N i < 0, state i becomes a net recipient. Net payments directly feed into consumption changes as constant returns to scale properties of output and zero profts imply ci =
Yi/Li − N i. Clearly, net transfers impact the incidence of federal policy. In the following, suppose that all states have equal populations sizes normalized to one and no state policies are implemented, i.e. τi = T and T > 0. Then the federal net payment of state i reduces to N i = (Ei − SiE) T, and each frm in each state faces the same emissions tax rate T such that Y i = T . We will relax these assumptions in the Ei numerical analysis. We now analyze how the emissions of state i respond to changes in emission intensity and capital wealth. Recall Section 4.3.1.2 introducing σ as the elasticity of substitution between the capital-labor composite and emissions . Let eli denote the elasticity Vi Ei YE of output Y with regard to E , i.e. eli ≡ Y i E /Y = TE /Y . i i YE Ei i i i i
Lemma 1 (Technology impact with immobile capital). Let χi > χj, Ki = Kj, Li = Lj, and τi = τj = T > 0. If capital is immobile across states and
eli (1 ) (4.5) YE < − σ , then (that is state i’s emissions are lower than those of state j). If eli (1 ) Ei < Ej YE > −σ , then Ei > Ej.
Proof. See 4.E.
Lemma 1 shows that a larger χi confronts the frm in state i (frm i) with the decision whether it is more proftable to use more or less emission input in contrast to the frm in state j (frm j). Two questions play the central role: how sensitive output reacts to a small change in emission input? How easy can inputs be substituted by each other. eli YE and σ in inequality (4.5) refect this decision problem of frm i. See also Figure 4.2 for
129 4. Technology Beats Capital the general relationship between production elasticities and substitution elasticities in output. Let us consider the decision problem of frm i more closely. When capital is immobile the composite input is constant, V i. Suppose that inputs are very complementary (σ → 0). The complementarity implies that increasing one factor without increasing the other does not increase output but would only increase production costs. When χi > χj, then frm i uses less emission input than frm j simply because of the complementary nature of and . To the contrary, a large output elasticity (large eli ) refects a Ei V i YE high sensitivity of output to small changes in . If eli is large and the inputs Yi Ei YE Ei and V i substitute each other well (σ → 1), then it is more proftable for frm i to use more emissions than frm j because the use of emissions is less limited by fxed supply of V i.
Lemma 2 (Technology impact with mobile capital). Let χi > χj, Li = Lj, and
τi = τj = T > 0. If capital is mobile across states and
eli (1 )(1 ) (4.6) YE < − σ − βk , then . If eli (1 )(1 ), then . Ei < Ej YE > − σ − βk Ei > Ej
Proof. See 4.F.
The case with mobile capital is relatively similar to the immobile case of Lemma 1.
If capital is mobile, however, also its elasticity on the composite input, refected by βk in inequality (4.6), comes into play. With mobile capital, the only limited input factor for frm i is labor supply. Hence, frm i faces a similar decision as before but can now freely chose over its input levels of emissions capital . Similar as eli refects Ei and Ki YE the sensitivity of output Yi to small changes in Ei, βK refects the output elasticity of the composite Vi to Ki. If Ei and Vi are very complementary (σ → 0) and/or if Vi is very insensitive to small changes in Ki (βk → 0), frm i will use less emissions than frm j. In contrast, if frm i can easily substitute Ei and Vi (σ → 1), and if small changes in and have a large impact on and (large eli and large ), respectively, frm Ei Ki Yi Vi YE βk 5 i tends to increase its emissions with raising χi. Empirically, it is likely that inequality (4.5) or (4.6) holds: On the left-hand side, we have that eli = i , where i is the cost share paid to emissions, which is currently YE ωE ωE below 10 percent (empirical data cf. Section 4.4)6. Values of the right-hand side are 7 βk ≈ 0.3 for the capital share and σ ∈ (0,0.7) for the elasticity of substitution . With
5We provide an alternative interpretation for the right-hand side by the relative marginal rate of substitution in 4.G. 6 Since Yi is homogeneous of degree 1 it follows thatYi = piVi + (τi + T )Ei. Division by Yi yields 1 = i + i , where i is the cost share paid to the composite. If = 0 then i = . ωV ωE ωV τi ωE TEi/Yi 7Supposing we treat σ similar to the values estimated for the elasticity of substitution between energy input and the capital-labor composite.
130 4.3 The model
Figure 4.2: Schematic sketch of production function indicating interplay of output and substitution elasticities.
Transfer rule Acronym Formula Equal per capita 1 ∑ SEQ / i Li Sovereignty i 1 o o SSO /LiEi /E Table 4.1: Transfer rules. these numbers, inequalities (4.5) or (4.6) hold easily and they would still hold for an emission cost share up to 20 to 30 percent for mobile and immobile capital, respectively. We fnd a simpler relationship for the impact of capital on emission demand:
Lemma 3 (Capital impact). Emissions in state i increase with a marginally larger available capital stock.
For the immobile capital case, consider equation (4.20), substitute βK βL for Proof. BKi Li
V i and take the derivative of Ei w.r.t. Ki to get that ∂Ei/∂Ki > 0. For mobile capital, proceed similar by using equation (4.24).
All else equal, Lemma3 follows by noticing that a larger available capital stock, either because of a larger capital endowment or an infow of capital (in the case of mobile capital) increases the marginal product of emissions and hence emissions increase. Let us combine our insights from Lemma 1, 2, and 3. In case of the EU, low capital stocks frequently coincide with low emission efcient technologies, cf. Figure 4.1. As such, in the absence of appropriate transfers the tax burden of a uniform federal emission tax threatens to be regressive. The federal tax payment can be compensated by redistributing federal tax revenues to consumers. We consider two types of federal transfer rules. The equal per capita transfer rule which is self-explanatory and the sovereignty transfer rule which accounts for a state’s high emission levels before federal policy-making (denoted with superscript o). In Table 4.1 we report the implementation of both rules. The equal per capita transfer rule distributes an equal share to all consumers in the federation and thus, if the federal tax payment is regressive, then also the net federal payment N i retains the federal tax’s regressive efect. The contrary is true for the sovereignty transfer rule.
131 4. Technology Beats Capital
4.3.3 Multilevel emission tax choices State governments and federal authority, i.e. both levels of government, regulate emissions. Emission taxes generate revenues for state governments and federal authority such that their revenue recycling budget reads tiEi and TE, respectively. The emission taxes of state governments now afect the budget of the federal authority, and vice versa (see also
Ei and E in 4.B and 4.C). Recall that the emission tax paid by frm i is the composite of τi = ti + T .
4.3.3.1 State governments
This section presents the second stage in which state government i non-cooperatively chooses the domestic emission tax ti that maximizes the sum of its population utility. Each state government takes all other emission taxes as given and incorporates the solution of the frm’s and consumers’ problem and the market clearing conditions into its optimization. Formally, it implies using (4.1) and (4.2) and substitution of the relevant variables after market clearing as in 4.B or 4.C for the immobile or mobile capital case, in which we now substitute the variables’ dependencies on states’ and federal taxes explicitly, τi = τi(ti,T ) and τ = τ(t1, ..., tm,T ). The indirect utility function then reads i i u (t, T ) ≡ u (ci (ti,T ) , E (t, T )). Since consumers within a state are identical, state i government’s problem is given by i max Liu (t, T ) given tj ∀j̸=i and T . (4.7) ti 8 Since ∂E/∂ti = ∂Ei/∂ti the frst-order condition of state i’s problem equals
∂ui ∂c ∂E = ui i + ui i = 0 for all i. (4.8) ci E ∂ti ∂ti ∂ti
The m states’ frst-order conditions implicitly defne the states’ taxes depending solely on the federal emissions tax which we denote by ti (T ). We defne the vector of all these state taxes as t (T ) ≡ (t1 (T ) , ..., tm (T )). For the case of capital mobility, we adopt the assumption of Zodrow and Mieszkowski (1986) that states do not take their efect on r into account. 9 Suppose that Si is exogenous and constant, then, after some algebraic manipulations of equation (4.8) we get
( ui ) t (T ) = L − E − S T for all i. (4.9) i i ui i ci
All else equal, state i’s tax is positively infuenced by the population size Li and a larger marginal dis-utility from emissions i , i.e. larger marginal damage. The state tax uE 8See also equation (4.28). 9See 4.H.
132 4.3 The model becomes lower with a larger marginal utility of consumption ui , transfer rule S , and ci i federal tax T . Depending on the magnitude of the marginal rate of substitution between total emissions and individual consumption, −ui /ui > 0, in contrast to federal transfer E ci and tax rule (SiT ), the state tax can also become negative (subsidy). For a detailed discussion of this term, we refer to Roolfs et al. (2018a) and Roolfs et al. (2018b). In absence of federal policy, i.e. T = 0, the second stage has an equilibrium solution, the decentralized solution denoted by subscript o. This solution will be used in the sequel to constrain the federal authority’s policy choice for Pareto-improvements and to calculate the sovereignty transfer rule, i . SSO
Defnition 4 (Decentralized policy equilibrium). The decentralized policy equilibrium with = 0 consists of quantities o, o, o o o and prices o, o, and taxes o m , T ci Yi Ki ,Li ,Ei ri wi {ti }i=1 such that for all i o solves the optimization problem of each consumer in state ; o, o ci i Yi Ki , o, and o solve the problem of frm ; o solves the problem of state ’s government; Li Ei i ti i and the market clearing conditions in capital, labor and fnal goods hold.
Let uoi denote the decentralized utility level. Setting T = 0 into equation (4.9) state i’s tax equals10 ui to = −L E for all i. (4.10) i i ui ci Each sovereign state i, absent of federal policy, internalizes the local damage from emissions afecting its population ( i ). Their chosen emission tax levels neglect the LiuE spillover efect of transboundary emissions to other states’ inhabitants, implying that there is potential for improvement beyond the decentralized solution since it lies below the social optimum (Samuelson rule).
4.3.3.2 Federal authority In the frst stage, the federal authority uses a policy package consisting of a uniform federal emission tax and the transfer rule Si as specifed in Section 4.3.2, Table 4.1. We constrain the federal authority to search for uniform federal taxes T which are Pareto-improving, relative to the decentralized solution, such that the federal policy package could be unanimously accepted by states and all consumers. Mathematically, the federal authority maximizes the utility of one consumer in state i such that no other consumers in any other state falls below their decentralized utility levels. Being the Stackelberg-Leader of the federation, the federal authority considers the indirect utility as defned in Section 4.3.3.1 and in addition the states’ policy reactions to the federal tax from equation (4.8). In the case of mobile capital, we suppose that federal authority takes into account its policy impact on r.
10In the case of mobile capital and if states would take their policy impact on r into account, the resulting state tax levels are ambiguous. In the decentralized case, capital importing states would set a higher state emission tax than in the small open economy case. Capital exporters would set lower state taxes. Proof available on request.
133 4. Technology Beats Capital
Figure 4.3: Stylized representation of the minimum and maximum uniform federal tax for two states adapted from Roolfs et al. (2018b). If 1 2 then 1 = max Tind < T Tind T .
The federal authority’s objective is given by
{ ⃓ } max ui (t (T ) ,T ) ⃓uj (t (T ) ,T ) ≥ uoj ∀ j ̸= i . (4.11) T ⃓ Let us give an intuitive explanation of the optimal federal solutions for m = 2 states by using Figure 4.3. In Figure 4.3 we plot the utility of each consumer as a function of the federal tax T . The decentralized utility levels are uo1and uo2. The federal authority seeks to raise the utility level of each consumer in state i above uoi, as long as the level for each consumer of state j does not fall below uoj. In the frst case, indicated with dashed lines, consumer 1 reaches its maximum u∗1 at T 1 and before consumers 2 who attains its maximum u∗2 at T 2. The federal solution space then ranges from T 1 ≡ T min to T 2 ≡ T max. If it happens that the utility of consumer 1 falls below its decentralized level uo1 before consumer 2 has reached its maximum, then the federal solution space ranges from 1 min to 1 max (dotted lines). T ≡ T Tind ≡ T We denote the uniform federal tax that solves equation (4.11) for state i with T i. Suppose, without loss of generality, l federal tax levels are solutions and that they can be ranked as T 1 < T 2 < ...T l. We will refer to T 1 as the minimum tax T min. If l = m we will refer to T m as the maximum tax T max. However, it may happen that the utility of one consumer in statek falls to its decentralized level before other consumers have reached their maxima. Then, T max no longer corresponds to T l but to the federal tax level ( k ) at which k equals ok— this case will indeed occur in some of our numerical Tind u u results. While we provide more technical details in 4.I, let us mention two other features that are important to the federal solutions. First, any uniform federal tax that satisfes unanimity must be positive (Roolfs et al., 2018b). Intuitively, the reason is that in
134 4.4 Numerical application order to carry out transfers, the federal authority must have a positive budget, T E > 0. This is only the case if the federal emission tax is positive. Second, any federal tax in the interval [T min, ...., T max] is a Pareto-dominant solution to the federal problem and denoted by T ∗. It satisfes the multilevel policy equilibrium:
Defnition 5 (Multilevel policy equilibrium). A multilevel policy equilibrium with transfer rule is the quantities ∗ ∗, ∗ ∗ ∗, prices ∗, ∗, and taxes ∗, ∗, such Si ci ,Yi Ki ,Li ,Ei ri wi ti T that for all = 1 consumption ∗ solves the optimization problem of eachconsumer i , ..., m ci in state ; ∗, ∗ ∗ and ∗ solve the problem of frm ; ∗ solves the problem of the i Yi Ki ,Li Ei i ti state government i; T ∗ solves the problem of the federal authority; the market clearing conditions of capital, labor and fnal goods hold; and the balance of payments condition ∗ + ( ∗ ∗ ) ∗ = ∗ is satisfed for all . Yi /Li SiE − Ei /Li T ci i In what follows we use subscripts EQ and SO when reporting equilibrium levels under the equal per capita and sovereignty transfers, respectively. We will refer to T i as the ’optimal uniform federal tax from state i’s perspective’ or in short ’the optimal federal tax of state i’. To compare the numerical results, we normalize T i with the lowest minimum tax { min min}. Tnorm ≡ min TEQ ,TSO
4.4 Numerical application
Section 4.3.2 showed that uniform federal taxes tend to put a higher burden on emission intensive states. Where low capital and low emission efciency coincide, the incidence of a federal policy package becomes an empirical question. In this section, we calibrate the model to the European Union and solve the general equilibrium numerically to determine the range of federal taxes that solves the federal authority’s problem as wells as each state’s net payments and incidence.
4.4.1 Data and calibration We account for three types of heterogeneity between EU countries: Population (and labor), capital stocks, and CO2 -emission efciency. To help isolate the efects of the three heterogeneities the remaining parameters are set symmetrically across all member states.
Population size and labor supply (Li) We used the most recent Census data from the year 2011 code cens_11r provided by Eurostat and extracted the number of persons per country of working age 15-64 years. In this analysis we assume that the population size is equal to the labor supply of the respective country. We will refer to countries with a small population as small countries and countries with a large population as large countries.
135 4. Technology Beats Capital
Aggregate capital stocks (Ki) We took capital stock estimates from Berlemann and Wesselhöft (2017; 2014) for the year 2014. Their estimates rely on the aggregate investment data provided by World Development Indicators (WDI) database. Data are in constant USD 2010. We selected capital stock data for the year 2014 as a compromise to use relatively recent data but keep the impact of the EU ETS low (ETS certifcate 11 prices were stable below 10 EUR per tonne of CO2 during this time) .
Capital per capita (ki) We derived per capita capital by dividing each country’s aggregate capital stock by its population size (Ki/Li). Eastern EU-countries have, in general, lower capital per capita levels, while the largest levels are in small, non-Eastern- European countries (Luxembourg and Sweden). For simplicity, we refer to countries with low per capita capital levels as poor countries and countries with large capital per capita as rich countries.
Emission efciency (χi) For a country’s representative emission efciency, we determined the relative CO2 -emission efciency. We took CO2 -emission intensity data for the year 2014 from WDI (code EN.ATM.CO2E. KD.GD accessed on 1/07/2019) accounting for CO2 -emissions from the burning of fossil fuels, the manufacture of cement and gas fuels and faring. The database provides each country’s CO2 -emission intensity measured in kg per constant 2010 USD of GDP. We calculated the inverse of country’s
CO2-emission intensity as measure of CO2 -emission efciency per country (CEi) and estimated the relative share of country i’s CO2 -emission efciency χi as
m ∗ CE χ ≡ i (4.12) i ∑m CE j j such that the average emission efciency is χi = 1.0. m = 28 equals the number of EU-countries.
Production We set the elasticity of substitution between the capital-labor composite and emissions equal to σ = 0.5, which falls well within the range identifed in empirical work.12 We use UK data from 2004 to specify production parameters. We used data from Ofce of National Statics (ONS) reference number 008744 accessed on 17/12/2018 for emission tax revenues and compensation of employees from ONS code DTWM (accessed on 03/01/2019). Non-residential emissions and gross value added are taken from WDI databases codes EN.ATM.CO2E.KT and NY.GDP.DEFL.ZS.AD (both
11Due to a lack of sufcient investment data for Malta, we approximate it’s capital stock by using the Perpetual Inventory Method for all other EU countries and take the average of their capital stock growth to extrapolate it to Malta. 12Elasticities of substitution between energy and the composite input are estimated by van der Werf (2007), Manne and Richels (1992) and Kemfert and Welsch (2000) and range from 0 − 0.7. Most empirical studies fnd that the elasticity of substitution between capital and labor is larger than the elasticity of substitution between energy and those inputs (Carraro et al., 2011).
136 4.4 Numerical application accessed on 17/12/2018), respectively. Using this data, we fnd the capital share of value added to be equal to βK = 0.43, and the share parameter for the capital-labor input to be equal to α = 0.97. We set B to unity and A = 3.1 using the approximation that = ( ).13 A ˜ Yi/ χiEi
Emission externality and utility We suppose that the utility of each consumer i γ in state i equals u (ci,E) ≡ log (ci) − E . We assume the damage to be quadratic on federal emissions and thus set γ = 2.
Country-clusters We summarize the regional data in Figure 4.4 and plot each country’s population size (y-axis) against its per capita capital levels (x-axis). The plot identifes four quadrants and three country-clusters: The top-right quadrant (group 1) comprises the cluster of large and rich countries with France, Germany, Italy, Spain, and the UK. The bottom-right quadrant (group 2) covers the cluster of small and rich countries (Luxembourg, Sweden, Denmark, Ireland, Austria, Finland, Netherlands and Belgium). In the bottom-left quadrant (group 3) small and poor countries cluster. Half of all EU countries belong to quadrant 3. Only Poland belongs to the upper upper left quadrant (group 4) which contains large and poor countries. Whether a country’s emission efciency is above or below the EU average is indicated by the size of its data point. Small data points indicate a large emission efciency.
Countries are in the average CO2-emission-efciency band for 0.9 < χi < 1.1. The relative emission efciency of almost all countries in quadrant 3 and 4 is below average, while all countries in the right quadrants 1 and 2 have average or above average emission efciencies. Remarkable outliers for the identifed country clusters are Germany, Luxembourg and Poland. Germany and Luxembourg are relatively far away from the other countries in their cluster in terms of population size (Germany) or capital per capita (Luxembourg). Poland is the only country in the large and poor cluster. Additionally, Germany is the only large and rich country with only average emission efciency.
4.4.2 Results immobile capital
Federal tax ranking We now report each of the optimal uniform federal tax levels that would be preferred by each member state (T i). That is, the equivalent taxes as those presented in the stylized illustration of Figure 4.3. With equal per capita transfers, Figure 4.5 shows that lowest optimal uniform taxes belong to several small and poor countries. Among these, Estonia is the country that prefers the lowest tax; the largest optimal tax is preferred by Sweden. Tax levels marked with dark gray bars violate the Pareto improvement restriction as other countries would
13See derivation in 4.J.
137 4. Technology Beats Capital
Figure 4.4: Countries cluster with respect to capital per capital and population size into four groups: rich and large (group 1), rich and small (group 2), poor and small (group 3), and poor and large (group 4) countries. Countries with the lowest per capita capital levels
have also the lowest CO2 -emission efciency.
138 4.4 Numerical application
Ki ki Li χi i 0.18 0.18 0.13 0.68 TEQ i -0.48 -0.16 -0.55 -0.08 TSO Table 4.2: Correlation coefcients between ranking of T i and countries’ heterogeneities. fall below their decentralized results. Comparing the ranking of taxes under sovereignty transfers to the country-clusters of Figure 4.4 shows that large and emission intensive countries rank at the lowest optimal federal tax levels: The lowest optimal tax min TSO belongs to Germany, a rich and large country with EU-average emission efciency, followed by Poland, which is the only poor and large country and has a low emission efciency. When comparing the levels of normalized federal tax ranges under sovereignty and equal per capita transfers, we see that i) the level of the minimum tax under sovereignty transfers is roughly two times larger than under equal per capita transfers. ii) The spread of the minimum to the largest optimal federal tax is also larger under sovereignty transfers, i.e. while max is more than four times larger than min the maximum tax TEQ TEQ max is more than 20 times larger than min. TSO TEQ
Figure 4.5: Optimal uniform federal tax rates T i subject to equal per capita (EQ) and sovereignty (SO) transfers. Each T i maximizes the utility of the respective country i i and provide solutions to the federal problem (red). Each T is normalized by Tnorm ≡ i = EST . Gray bars are truncated tax levels and indicate that the optimal federal min{T } TEQ tax of that respective country is too large and therefore is not a solution to the federal problem (no Pareto improvements).
In Table 4.2 we further disentangle the optimal federal tax ranking by reporting Bravais-Pearson-correlations, between the optimal federal taxes from each country’s perspective and endowments and technological heterogeneities of the countries.
139 4. Technology Beats Capital
The correlations identify CO2 -emission efciency as the driving force of the ranking of federal taxes with equal per capita transfers. Clean production is associated with preference for a high EU wide emissions tax. Capital stocks, capital per capita and population size have a much less positive impact on the ranking of optimal taxes. With sovereignty transfers all correlations are negative. The dominating correlation is found for population size, followed by countries’ capital stocks. Capital per capita and
CO2 -emission efciency have a weak correlation and thus less impact on the ranking of optimal taxes. Comparing the correlations of the two transfer rules it becomes apparent that the signs are always opposite. This indicates that the rules impose contrary burden distribution. In the following we will further examine this at the minimum taxes.
Net federal payments at the federal minimum tax In Figure 4.6 we report the federal transfer to consumers (negative, blue bars), the per capita tax payment to the federation (positive, red bars) and the net payment (black dots) which is simply the diference between payments and transfers. Under equal per capita transfers only the payment side creates diversity in the net payments. Per capita payments are the largest for Estonia and Luxembourg. Many small countries — being emission intensive or rich — face a positive net payment and thus become net donors to federal tax revenues, while several rich countries become net recipients.
Figure 4.6: Federal payment (positive, red) and transfer (negative, blue) and the resulting net payment at the federal minimum tax given equal per capita (EQ) and sovereignty (SO) transfers per capita.
Under sovereignty transfers per capita payments to the federation are relatively homogeneous and similar to the equal per capita case. Whereas transfers from the federation are relatively low for consumers in large and rich countries and vice-versa for small and poor countries. Rich and large countries’ consumers are net donors and
140 4.4 Numerical application
smaller countries’ consumers are net recipients. The top fve countries with the highest net payments are all from the rich and large country cluster. In contrast, almost all other countries become recipients of net payments under sovereignty transfers (only Poland is a exemption) and thus, we have a frst indication that sovereignty transfers tend to turn federal emission pricing progressive. The order of magnitude between the positive and negative net payments with sovereignty transfers is larger than with equal per capita transfers. Why does the federal net payment per capita with sovereignty transfers appear to draw from relatively few countries, while there are large transfers to many others? The
CO2 -emission level of a country represents its emission tax base. As shown in Lemma 3, larger capital stocks imply a larger emission tax base and thus a larger federal tax payment. If the largest net payment lies on rich and large countries then the majority of federal revenues is collected from these largely populated countries. But due to their large population sizes still each of these net donating countries faces a low per capita net payment. Let us consider Figure 4.7 which illustrates the previously identifed tax-base-efect by measuring net payments per country instead of per capita payments. We report the
federal transfer, the federal payment, and the net federal payment (N iLi) per country. The tax payment per country translates to the size of each country’s emission tax base. The larger the bar of the tax payment, the larger is its tax base, and vice versa. Under equal per capita transfer, countries with a positive net payment (net donors) do not necessarily have the largest emission tax base. Whereas the picture changes under sovereignty transfers. Countries with large emission tax bases take the largest positive gross tax payment but also take the largest positive net payment.
4.4.3 Results contrasting mobile to immobile capital This section extends our numerical consideration by introducing perfect capital mobility across states. As we will show, the tendency of the results remains robust to mobile capital but gets more pronounced. With mobile capital this paper connects to the inter-regional tax competition literature which was frst investigated by Zodrow and Mieszkowski (1986), and Wilson (1986). In a similar spirit as the previous literature but motivated by the overlapping political architecture in the EU, Habla and Winkler (2018) investigate the interaction of states’ capital taxation and federal emission policy on the provision of public goods. They fnd that capital mobility can result in too low or too large public good provision in a state depending on the in- or outfow of capital. We refer to their paper also for a more extensive literature review. While that specifc literature focuses on capital taxation, our paper keeps the model setup with only emission taxation but contrasts the result given immobile and mobile capital.
141 4. Technology Beats Capital
Figure 4.7: Federal payment (positive, red) and transfer (negative, blue) and the resulting net payment at the federal minimum tax given equal per capita (EQ) and sovereignty (SO) transfers per country. The size of the red bars refects the size of a country’s emission tax base. With sovereignty transfers, net payments are mostly proportionally increasing with a larger tax base. Equal per capita transfers show no positive correlation with the tax base size.
Ki ki Li χi i 0.40 0 56 0.26 0.88 TEQ . i -0.40 -0.39 -0.42 -0.28 TSO Table 4.3: Correlation coefcients given mobile capital.
We select three metrics to compare the case of mobile to immobile capital. First, we discuss the ranking of the uniform federal taxes T i when capital is mobile. Second, we compare the net federal payment under mobile and immobile capital. Third, we compute the burden (incidence) as change in per capita consumption levels for both capital cases. The basic mechanisms of the following observations and discussed efects base on our statements in Lemma 1, 2 and 3.
Federal tax ranking with mobile capital We fnd an almost similar order of the ranking of optimal federal taxes as in Section 4.4.2 and present a comparable Figure in 4.L. Table 4.3 reports on the correlation of the ranking of the optimal tax levels and the considered heterogeneities when capital is mobile. With equal per capita transfer we see a strong increase in the correlation of all heterogeneities. In particular, the efect of emission efciency χi now dominates the ranking of optimal federal tax levels even stronger. The change in the correlation of state heterogeneities under sovereignty transfers is much less pronounced.
142 4.4 Numerical application
Figure 4.8: Net federal payment under equal per capita transfers ranked by capital per capita. The net payment lies on poor (group 3 and 4, red) and small and rich countries (group 2, white) are net donors.
Comparison of net payments In Figures 4.8 and 4.9 we present the per capita federal net payments. On the x-axis, we plot capital per capita levels per country ranked from the lowest (Bulgaria) to the largest (Luxembourg) level. On the y-axis, the net federal payment per capita indicates if the country is a per- capita net donor (a positive number) or a net recipient (a negative number) of federal tax revenues. We mark the net federal payment per capita for the immobile and mobile capital case with black squares or red circles, respectively. We connect these two data points corresponding to the consumer in state i with a vertical line. A longer vertical line signals a larger diference between the net payment level under mobile and immobile capital. Figure 4.8 shows that under federal equal per capita transfers and mobile capital, consumers belonging to the the small and poor-cluster (group 3) and Poland are net donors of federal emission tax revenues while in the case of immobile capital, fve of these countries are still net recipients. All consumers from countries belonging to the rich and large cluster face a negative net payment. Most rich and small countries’ consumers are net donors in the immobile capital case, but become less strongly net donors or even net recipients with mobile capital. Figure 4.9 suggests that our claim that sovereignty transfers make a federal tax progressive gets more pronounced under capital mobility. But the net diference between mobile and immobile capital assumption are much smaller in almost all cases (length of vertical lines). Capital poor consumers, except for Poland, become net recipients. Poland’s exemptions can be explained by its remarkably large population size which makes it an exception to the cluster of all other low-capital countries. Population size
143 4. Technology Beats Capital
Figure 4.9: Net federal payment under sovereignty transfers ranked by capital per capita. Most of the net payment lies on the large and rich country-cluster (group 1, gray).
Capital Ki ki Li χi N i,EQ immobile -0.31 0.34 -0.40 -0.22 N i,EQ mobile -0.61 -0.69 -0.50 -0.85 N i,SO immobile 0.50 0.07 0.57 0.29 N i,SO mobile 0.49 0.26 0.55 0.38
Table 4.4: Correlation coefcients of net payments and heterogeneous country data. links to labor supply. Thereby, Poland’s large population size acts similar as a large capital stock and increases production. As a by-product emissions increase as well14. Table 4.4 reports more correlations between net payments and heterogeneous country data. Why do our previous observations become stronger with the introduction of mobile capital? The reason is that mobile capital reinforces the efects of the technologies’ emission efciencies. i) Low-capital EU countries tend to have lower emission efciencies in production, leading to higher emission levels compared to other countries that are otherwise identical (cf. Lemma 1 and 2). ii) With the additional capital-infow due to mobile capital15, capital-poor EU countries can take over more of the production of the federal economy (cf. Lemma 3). The capital-infow thus further increases their emission levels.
14Technically speaking, this is due to the gross complementarity between the composite input of capital-labor, and emissions. 15In the absence of capital mobility, the marginal product of capital is larger in poor countries than in rich countries. With capital mobility, capital fows from rich to poor countries as to equalize the marginal product of capital across countries.
144 4.5 Conclusion
In other words, capital-rich countries not only export their capital, but also implicitly export emissions to capital-poor countries and thereby increase the multinational emission tax burden on poor countries. Absent of appropriate revenue transfers, capital mobility retains the observed regressive efect of a uniform federal emission tax.
Federal tax incidence Figure 4.10 reports the welfare changes due to consumer burden resulting from the federal minimum tax. We report the relative percentage change of per capita consumption between the multilevel policy equilibrium relative to the decentralized outcome, ( ∗ o) o. To conceptualize our results, we add the group ci − ci /ci clusters from Figure 4.4. We fnd that consumption in small and rich and all poor countries decreases with federal equal per capita transfers compared to the decentralized case (triangles). The changes in consumption in large and rich countries with equal per capita transfer are almost zero. Under sovereignty transfers (squares), all consumers in large and rich countries are exposed to a larger decline in per capita consumption in contrast to equal per capita transfers but also in contrast to other country clusters. Many poor countries and some rich and small countries’ consumers experience an increase in consumption. While we fnd most consumption changes to be in a similar range as in Böhringer, Rivers, Rutherford, et al. (2015) we fnd larger diferences in direction and magnitude for poor states given sovereignty transfers. We attribute these departures to the characteristic negative correlation between emission intensity and capital per capita in many EU countries. In the case of the Canadian provinces the relation between GDP per capita and carbon intensity is positive. When comparing Figure 4.10 to Figures 4.8 and 4.9 we see that the net payments to the federal authority are an indicator of the resulting tax burden. Countries may be rightly concerned about the regressive tax incidence if they focus on net payments by the federal government.
4.5 Conclusion
When a government plans an environmental tax reform, public support for the reform is closely linked to the tax burden on consumers. Public support can be improved by strategically complementing tax reform by transfer rules that recycle tax revenues to consumers. In practice, policy-makers, scientists, lobby groups and policy advisers are confronted with limitations and diferent opinions about the "best" transfer rule to use (Burtraw et al., 2009; Delbeke, 2017; Kverndokk, 2018; Williams, 2019). Furthermore transfer heuristics often follow rules of thumb based on welfare economics, moral considerations, and state self-interests. In this paper, we trace the consumers’ burden of uniform federal tax payments to member state diferences in wealth and technological emission intensity for two
145 4. Technology Beats Capital
Figure 4.10: Incidence of the federal minimum emission tax. Filled symbols correspond to mobile capital and hollow symbols to immobile capital simulation. commonly used transfer rules in a simple general equilibrium. We fnd that the gross emission tax payment is larger for countries that are wealthier or for those with a large emission intensity of the production technology or both. When countries with a high emission intensity also rank lower on the distribution of wealth, an environmental policy reform threatens to become regressive. We show that equal per capita transfers do not counteract this regressive efect, but a transfer to consumers based on historical emissions (sovereignty) can result in a progressive emission tax reform. For our numerical simulation to the EU, we extend the model with population diference and with coexisting state and federal governments that non-cooperatively choose their emission tax levels to maximize state or federal welfare. We have constrained the federal policy to theoretically ensure unanimous consent of states (Pareto-improvement with respect to their sovereign policy outcome) towards the federal tax and transfer policy. Our fndings contribute to the understanding of how to increase acceptance of federal environmental policies by citizens and state-level governments. We know from earlier studies that in the single country context a uniform tax (or a uniform price) combined with equal per capita transfers are progressive policy for citizens. One might think that this result holds, when it is applied to the member states of a federal CO2 price system.
This paper shows that this is not the case when, as in the EU, wealth and CO2 intensity of production afect the burden in opposite ways. Sovereignty transfers, i.e. the way the
146 4.5 Conclusion
EU has calculated the bulk of revenues from the ETS, produce – perhaps surprisingly – an egalitarian result. We see a number of ways to extend the current analysis. First, the analysis could be extended to cover other transfers discussed in literature and politics. A particular interesting case are strategic transfers that reward consumers in a state according to the state’s mitigation ambition, setting an incentive for more mitigation to receive more transfers. Alternatively, one could include the historical accountability of those who have generated large prosperity from historical emissions and thus partially reverse the efects of sovereignty transfers. Second, our analysis assumes homogeneity within countries. Accounting for income diferences within states would allow a deeper discussion of welfare and inequality consequences of the policy packages. Third, while the above considerations relate exclusively to emissions policy, environmental taxes can also interact with a distortive tax system. In how far the revenues from the emission tax can be used to reduce other distorting taxes in a Pareto improving and progressive reform would add another perspective to the research interest of this study. Fourth, one could calculate the socially optimal transfers and emission tax in this federal set-up with diferences in emission intensity and wealth. Fifth, one could analyze the outcome with a dynamic model in which the technology in a state improves in the long run with available capital or produced output. We plan to address some of these issues in future research.
147 4. Technology Beats Capital
Acknowledgments
Financial support by the German Research Foundation (DFG) grant number LE 782/2-1 is gratefully acknowledged. For their helpful comments and suggestions we thank in alphabetical order Wolfgang Habla, Achim Hagen, Biung-Ghi Ju, Zarko Kalamov, Linus Mattauch, Samuel Okullo, Marco Runkel, Emilie Soysal, Oliver Tietjen, Boyan Yanovski, and participants of the Public Economic Theory 2019 conference. For technical assistance we thank Lavinia Baumstark, Lukas Feldhaus, Anastasis Giannousakis, and David Klein. For proof-reading we thank Andrew McConnell.
148 REFERENCES
References
Acemoglu, D. (2002). Directed Technical Change. The Review of Economic Studies, 69 (4), 781–809. Bergstrom, T., Blume, L., & Varian, H. (1986). On the private provision of public goods. Journal of Public Economics, 29, 25–49. Berlemann, M., & Wesselhöft, J. (2017). Aggregate Capital Stock Estimations for 122 Countries: An Update. Review of Economics, 68 (2), 1158. Berlemann, M., & Wesselhöft, J. (2014). Estimating aggregate capital stocks using perpetual inventory method - new empirical evidence for 103 countries -: Working Paper Series. Review of Economics, 65, 1–34. Böhringer, C., & Lange, A. (2005). Economic Implications of Alternative Allocation Schemes for Emission Allowances*. Scandinavian Journal of Economics, 107 (3), 563–581. Böhringer, C., & Lange, A. (2003). On the Design of Optimal Grandfathering Schemes for Emission Allowances. ZEW Discussion Papers, No. 03-08. Böhringer, C., Rivers, N., Rutherford, T., & Wigle, R. (2015). Sharing the burden for climate change mitigation in the Canadian federation. Canadian Journal of Economics, 48 (4), 1350–1380. Böhringer, C., Rivers, N., & Yonezawa, H. (2016). Vertical fscal externalities and the environment. Journal of Environmental Economics and Management, 77, 51–74. Boyce, J., & Riddle, M. Cap and Dividend: How to Curb Global Warming While Protecting the Incomes Of American Families: Working Papers wp150, Political Economy Research Institute, University of Massachusetts at Amherst. 2007. Burtraw, D., Sweeney, R., & Walls, M. (2009). The Incidence of U.S. Climate Policy: Alternative Uses of Revenues from a Cap-and-Trade Auction. National Tax Journal, 62 (3), 497–518. Cambridge Econometrics. Review of the Energy Taxation Directive: fnal modeling results’. A report for DG TAXUD, European Commission. 2008. Carraro, C., de Cian, E., & Nicita, L. (2011). Modeling biased technical change. In S. Dietz, J. Michie, & C. Oughton (Eds.), Political Economy of the Environment (pp. 167–185). Taylor & Francis. Cazorla, M., & Toman, M. (2001). International Equity and Climate Change Policy. Climate change Economics and Policy, 235–247. Chiroleau-Assouline, M., & Fodha, M. (2014). From Regressive Pollution Taxes to Progressive Environmental Tax Reforms. European Economic Review, 69, 126– 142. Delbeke, J. (2017). A Comment on ”Agreeing on an EU ETS Price Floor to Foster Solidarity, Subsidiarity, and Efciency in the EU”. In I. W. H. Parry, K. Pittel,
149 4. Technology Beats Capital
& H. R. J. Vollebergh (Eds.), Energy tax and regulatory policy in Europe. MIT press. Dorband, I., Jakob, M., Kalkuhl, M., & Steckel, J. (2019). Poverty and distributional efects of carbon pricing in low- and middle-income countries – A global comparative analysis. World Development, 115, 246–257. Fullerton, D. (Ed.). (2016). Distributional Efects of Environmental and Energy Policy. Routledge. Grubb, M., Sebenius, J., Magalhaes, A., & Subak, S. (1992). Sharing the burden. In I. M. Mintzer (Ed.), Confronting Climate Change: Risks, Implications and Responses (pp. 305–322). Cambridge University Press. Habla, W., & Winkler, R. (2018). Capital Tax Competition under a Centralized Emissions Cap: WCERE manuscript. WCERE 2019 conference-paper. Ju, B., Kim, M., Kim, S., & Moreno-Ternero, J. (2019). Climate Change Justice: An Axiomatic Approach. PET 2019 conference paper. Kemfert, C., & Welsch, H. (2000). Energy-Capital-Labor Substitution and the Economic Efects of CO2 Abatement: Evidence for Germany. Journal of Policy Modeling, 22 (6), 641–660. Klenert, D., & Mattauch, L. (2016). How to make a carbon tax reform progressive: The role of subsistence consumption. Economics Letters, 138, 100–103. Klenert, D., Mattauch, L., Combet, E., Edenhofer, O., Hepburn, C., Rafaty, R., & Stern, N. (2018). Making carbon pricing work for citizens. Nature Climate Change, 8 (8), 669–677. Knopf, B., Koch, N., Grosjean, G., Fuss, S., Flachsland, C., Pahle, M., Jakob, M., & Edenhofer, O. The European Emissions Trading System (EU ETS): ex-post analysis, the market stability reserve and options for a comprehensive reform: Nota di Lavoro 79.2014, Milan Italy: Fondazione Eni Enrico Mattei. 2014. Kosonen, K. (2012). Regressivity of environmental taxation: Myth or reality? (Vol. 32). Publ. Ofce of the Europ. Union. Krepps, D. (1990). A Course in Microeconomic Theory. Princeton University Press. Kverndokk, S. (2018). Climate Policies, Distributional Efects and Transfers Between Rich and Poor Countries. International Review of Environmental and Resource Economics, 12 (2-3), 129–176. Kverndokk, S., & Rose, A. (2008). Equity and Justice in Global Warming Policy. International Review of Environmental and Resource Economics(2), 135–176. Lutsey, N., & Sperling, D. (2008). America’s bottom-up climate change mitigation policy. Energy Policy, 36 (2), 673–685. Manne, A., & Richels, R. (1992). Buying Greenhouse Insurance: The Economic Costs of CO2 Emission Limits (1st ed., Vol. 1). The MIT Press. Musgrave, R. (1959). The Theory of Public Finance. McGraw-Hill.
150 REFERENCES
Oates, W. A Reconsideration of Environmental Federalism. Discussion Paper 01–54. 2001. Oates, W. (1972). An Essay on Fiscal Federalism. Journal of Economic Literature, 37, 1120–1149. Oates, W. (2000). Fiscal competition and European Union: contrasting perspectives. Regional science & urban economics, 31, 133–145. Padilla, E., & Roca, J. (2004). The Proposals for a European Tax on CO2 and Their Implications for Intercountry Distribution. Environmental and Resource Economics, 27, 273–295. Rausch, S., Metcalf, G., Reilly, J., & Paltsev, S. (2010). Distributional Implications of Alternative U.S. Greenhouse Gas Control Measures. Roolfs, C., Gaitan, B., & Edenhofer, O. (2018a). Assessing EU minimum prices and revenue distribution from a multilevel policy perspective. WCERE 2019 conference paper. Roolfs, C., Gaitan, B., & Edenhofer, O. (2018b). The good, the bad, and the useless: Multilevel emissions policy and the efectiveness of commonly used federal transfers. initial submission to JEEM in 2018. Sheeran, K. (2006). Who Should Abate Carbon Emissions? A Note. Environmental and Resource Economics, 35 (2), 89–98. Sterner, T., & Robinson, E. (2018). Selection and design of environmental policy instruments. In P. Dasgupta, S. K. Pattanayak, & V. K. Smith (Eds.), Handbook of Environmental Economics (pp. 231–284). Elsevier. van der Werf, E. Production Functions for Climate Policy Modeling: An Empirical Analysis, Nota di Lavoro, NOTA DI LAVORO 47.2007. 2007. Williams, R. [Roberton3Will]# We need to pay more attention to distributional efects. Economists far too often call the most efcient policy ”best”, even when it has problematic distributional efects. And distribution is a key element of political feasibility. [Tweet]. https : / / twitter . com / Roberton3Will / status / 1081957953907974144. Twitter. Date 2020-01-06, 2019. Williams, R. (2012). Growing state–federal conficts in environmental policy: The role of market-based regulation. Journal of Public Economics, 96 (11-12), 1092–1099. Wilson, J. (1986). A theory of Interregional tax Competition. Journal of Urban Economics, 19, 296–315. Zhou, P., & Wang, M. (2016). Carbon dioxide emissions allocation: A review. Ecological Economics, 125, 47–59. Zodrow, G., & Mieszkowski, P. (1986). Pigou, Tiebout, Property Taxation, and the Underprovision of Local Public Goods. Journal of Urban Economics, 19 (3), 356–370.
151
4.A Conditional input demand
Appendix
Appendix 4.A Conditional input demand
Using equation (4.4) the conditional demand function for Ei is
σ σ ( ) 1−σ σ z σ 1−σ zp α χ σ Y = i + i i (4.13) Ei 1−σ 1−σ 1−σ σ τi pi A χi
σ 1 = z Yi (4.14) 1−σ 1−σ . σ τi A χi
The corresponding conditional demand function for Vi is
σ σ ( ) 1−σ z 1−σ σ σ σ ( ) σ α χi Yi α Yi V = (ατ ) + = . (4.15) i i 1−σ 1−σ 1−σ τi pi A pi A
Appendix 4.B Market clearing with immobile capi- tal
Rearranging equation (4.4) yields
σ z 0 σ 1−σ = 1−σ 1−σ ( ) . (4.16) < α pi A − 1−σ τi ≡ ϕi τi σ χi Using the market clearing conditions and , we can describe the rental rate of ϕi capital in state i as
( )βL β B L 1 ( ) = K i 1−σ (4.17) ri τi σ ϕi α 1−σ Ki
Successive replacement of ri (τi) provides us with the other variables, which are exclusively depend on τi. The wage rate in state i equals
( )βK β B K 1 ( ) = L i 1−σ (4.18) wi τi σ ϕi . α 1−σ Li Since = βK βL ,output in state equals Vi BKi Li ≡ V i i
( ) σ ϕ 1−σ Y (τ ) = A1−σ i V (4.19) i i α i
153 4. Technology Beats Capital
State i’s emission level is given by
σ 1 σ 1 1−σ ( ) = z ϕi (4.20) Ei τi 1−σ 1 V i. σ α 1−σ τi χi and aggregate federal emissions are
m E (τ) = ∑ E (4.21) j=1 j
where τ ≡ (τ1..., τm).
Appendix 4.C Market clearing with mobile capital
From market clearing follows that the rental rate of capital is
( 1 )βl ∑ (1−σ)βl i ϕi Li ( ) = Bβk (4.22) r τi σ β α 1−σ K l and the wage rate in state i is
1 ( )βk β B βk l w (τ ) = r β . (4.23) i i σ 1 l 1−σ 1−σ α ϕi Firm i’s frst order conditions imply
σ 1 σ ( ) 1−σ ( ) = z ϕi V i (4.24) Ei τi 1 1−σ . α 1−σ τi χi and capital demand is
1 1 1 β ( ) β l βkB l ϕi 1−σ Ki (τi) = σ Li. (4.25) α 1−σ r
β Substitution of = βK L into equation (4 24) yields V i BKi Li .
1 σ ( )βk βl B βk σ +β z Li ( ) = 1−σ k . (4.26) Ei τi σ ϕi 1−σ σ α 1−σ r σ τi χi
Appendix 4.D Few comparative statics
Now that we have characterized all supply and demand plans and all market equilibria, we can calculate the comparative statics of small emission tax increases in one state.
154 4.E Proof of Lemma 1
Recall that τi(ti,T ) = ti + T and consequently ∂τi/∂ti = ∂τi/∂T = 1, equation (4.16) implies
∂ϕ i < 0. (4.27) ∂τi
From equations (4.27), it follows that ∂ri/∂τi < 0 and ∂wi/∂τi < 0 . This shows that an increase in the emission tax in state i reduces frm i’s remuneration to capital and labor. Similarly as before, using equation (4.27) it follows that ∂Yi/∂τi < 0 implying that an increase in the emission tax in state i reduces the production of frm i. From equations (4.20) and (4.21) follows ∂E ∂E ∂E ∂E i = = i < 0, and < 0. (4.28) ∂τi ∂ti ∂T ∂T A greater emission tax in state i decreases state i’s emissions demand. Further, from equation (4.20) follows that ∂Ei/∂tj = 0 which tells that, at the partial equilibrium, emission demand of state i is is unafected by tax changes of other states j ≠ i. We use this result to solve the state tax level in Section 4.3.3.1.
Appendix 4.E Proof of Lemma 1
Suppose ti = 0. Then τ i(ti,T ) = τ i(0,T ) = T . Further let T > 0 and capital is immobile across states. First, let us consider equation (4.14)
σ 1 = z Yi (4.29) Ei 1−σ 1−σ . σ T A χi
With elYE ≡ TEi/Yi denoting the output elasticity of Yi subject to Ei and rearranging (4.29) yields
1 χiA 1−σ T = (elYE) σ . (4.30) z 1−σ
Since frm i is a price taker it ignores its impacts on ri. The derivative of Ei, from (4.20), w.r.t. and after some algebraic manipulations and using the defnition of 0 from χi ϕi > equation (4.16) we get
σ 1 σ 1 1−σ ( ) = z ϕi Ei τi 1−σ 1 V i. σ α 1−σ τi χi ∂E E ( T )1−σ i = − i F. (4.31) ∂χi χiϕi χi ⏞ ⏟⏟ ⏞ + ( )1−σ with ((1 ) χiA σ). While we can sign all terms except in the equation F ≡ − σ T − z F (4.31) clearly positive, we can only say so far that if F > 0 then ∂Ei/∂χi < 0 and vice versa. Suppose F > 0 and solve for T from F to get
1 χiA 1−σ T < (1 − σ) σ . (4.32) z 1−σ
155 4. Technology Beats Capital
Equate equation (4.30) with the left-hand side of inequality (4.32) to get
1 χiA 1 χiA 1−σ 1−σ T = (elYE) σ < (1 − σ) σ (4.33) z 1−σ z 1−σ indicating that the decisive terms to sign F are elYE and (1 − σ). Thus, as long as
elYE < (1 − σ) then ∂Ei/∂χi < 0.Proceed similarly to show that from elYE > (1 − σ) follows F < 0 and thus ∂Ei/∂χi > 0.
Appendix 4.F Proof of Lemma 2
We provide less explanatory text and refer to the previous one for further explanations.
Suppose ti = 0. Then τ i(ti,T ) = τ i(0,T ) = T . Further let T > 0 and capital is mobile across states.
For the derivative of Ei, from (4.24), w.r.t. χi and after some algebraic manipulations and using the defnition of 0 from equation (4.16) we get ϕi >
∂E E ( T )1−σ i = − i G. (4.34) ∂χi χiβlϕi χi ⏞ ⏟⏟ ⏞ + ( )1−σ with ((1 ) χiA σ). While we can sign all terms except in equation G ≡ − σ βl T − z G (4.34) clearly positive, we can only say so far that if G > 0 then ∂Ei/∂χi < 0 and vice versa. Suppose G > 0 and solve Gfor T to get
1 χiA 1−σ T < ((1 − σ) βl) σ . (4.35) z 1−σ Note that equations (4.29) and (4.30) apply to the mobile capital case, too. Equate equation (4.30) with the left-hand side of inequality (4.35) to get
1 χiA 1 χiA 1−σ 1−σ T = (elYE) σ < ((1 − σ) βl) σ (4.36) z 1−σ z 1−σ indicating that the decisive terms to sign G are elYE and (1 − σ) βl where βl = 1 − βk. Thus, as long as
elYE < (1 − σ) (1 − βk) then ∂Ei/∂χi < 0. Proceed similarly to show that from elYE > (1 − σ) (1 − βk) follows G < 0 and thus ∂Ei/∂χi > 0.
156 4.G Alternative interpretation of Lemma 1 and Lemma 2
Appendix 4.G Alternative interpretation of Lemma 1 and Lemma 2: The relative marginal rate of technical substi- tution
Adjustments of capital and emissions inputs are determined by the marginal rate of technical substitution16 between emissions and capital ( i ). Its slope equals the MRTSEK negative relative marginal product of emissions:
1−σ 1 ( ) β +σβ (1−σ)βl σ Y i z 1 σ K k l L 1 rel. i = Ei = i i (4.37) YE ≡ −MRTSEK i YKi α χi Ei βk ⏞ ⏟⏟ ⏞ ⏞ ⏟⏟ ⏞ (*) (**)
where Y i is the marginal product of capital. Ki Ceteris paribus, we see that the Ei-augmenting factor χi reduces the relative marginal product of Ei, cf. term (*) in (4.37) and the frm will change Ki more strongly than Ei. Acemoglu (2002) refers to this efect as K-bias induced by χi-change. However due to a nested CES function as sketched in Figure 4.2, term (**) in equation
4.37 points at the ambivalence of the role of Ki: Larger Ki can potentially induce an E-bias. Whether χi-change is E or K-biased depends on the relative abundance of and known as the substitution efect17 and determined by the i We Ki Ei MRTSEK . calculate the total elasticity of the i with regard to both variable inputs MRTSEK Ki,Ei stating how output would change if both inputs change in one percent:
(∂rel.Y i E ∂rel.Y i K ) 1 1 elMRTS = − Ei i + Ei i = − β − β . E,K rel. i rel. i k l ∂Ei YEi ∂Ki YEi σ σ As long as 0 1, the slope i is diferent for any combination of < σ < MRTSEK Ki and Ei and thus depends on the actual input share of Ki to Ei to which a χ-change would impose marginal input adjustments. Therefore, we calculate the elasticity of the i with regard to changes in the input share, stating how the i would MRTSEK MRTSEK change if the input share changes in one percent:
∂rel.Y i K /E 1 elMRTS = − Ei i i = . KEshare rel. i ∂Ki/Ei YEi σ The relative elasticity of the marginal substitution rate (or elasticity of the substitution efect) measured in the respective input bundle Ki/Ei is elMRTS E,K = (1 − σ)(1 − β ) elMRTS k KEshare which is the right-hand side of inequality (4.6). Therefore, if the relative elasticity of the marginal substitution rate outweighs the output elasticity of emissions, Ki but also 16Graphically speaking, the i measures the slope of the isoquant along which the frm fnds MRTSEK its optimal input bundles of Ki,Ei. 17 Leading to a downward sloping relative demand curve for Ei if Ki increases.
157 4. Technology Beats Capital
Ei will decrease with a χi-change. Vice versa, Ki decreases (strongly) but Ei increases. Proceed similar for the case of immobile capital.
Appendix 4.H State i’s frst-order conditions
Use equation (4.28) with ∂E/∂ti = ∂Ei/∂ti. It follows that that the frst-order condition of state i’s problem reduces to ∂ui/∂t = ui ∂c /∂t + ui ∂E /∂t = 0. Algebraic i ci i i E i i manipulations allows to rewrite it as
∂ui 1 ∂E ∂E ∂E = t i ui + S T i ui + ui i = 0. (4.38) i ci i ci E ∂ti Li ∂ti ∂ti ∂ti and solving for ti yields the state tax chosen by state i.
Appendix 4.I Technical description of federal solu- tions
With the formulation of the federal problem as in equation (4.11), we make use of a traditional concept: The formulation of a Pareto improvement and targeting at Pareto dominant taxes is equivalent to maximize a social welfare function given specifc weights (cf. Krepps, 1990; Sheeran, 2006). For each minimum level assigned to a consumer in statej, uoj, when maximizing the utility of aconsumer in statei in the Pareto-improvement form as in equation (4 11), there is a set of social welfare weights with ∑ = 1 . λi j λj which produces the same Pareto result when maximizing a social welfare function of all i consumers and with the similar rule, and vice versa and T corresponds to λi = 1.
Appendix 4.J A proxy
The cost share of emissions in production is ωE ≡ τiEi/Yi. We set ωE = z. Solving for A from equation (4.13) we get A = 1/χiYi/Ei.
Appendix 4.K Replication of utility structure
Figure 4.11 replicates the structure of utility curves, the federal tax range and ranking of federal taxes as discussed in Section 4.3.3.2 and Figure 4.3. We report utility levels by plotting the relative change between the multilevel and decentralized equilibrium levels. The diferent curvature of the slopes shows that diferent transfer rules impact the slope and ranking of the utilities’ maxima and thus the ranking of optimal uniform federal taxes.
Appendix 4.L Tax ranking with mobile capital
Figure 4.12 shows the optimal federal emission tax of each country under equal per capita and sovereignty transfers is relatively similar to the immobile capital case, Figure 4.5. However, when comparing the immobile (closed) to the mobile (open economy) capital case, we see that i) optimal federal taxes are generally lower in the mobile than
158 4.L Tax ranking with mobile capital
Figure 4.11: Relative change of utility levels between the multilevel and decentralized equilibrium under equal per capita (EQ) and sovereignty transfers (SO). The points on the i lines correspond to T /Tnorm. Utility levels approaching 0 with increasing T/Tnorm indicate that each consumer in the respective country falls back to or below her decentralized utility level with a further increasing federal tax level. in the immobile capital case, as shown by the fact that the lowest tax under equal per capita transfers is below the dotted line (corresponding to the federal minimum tax under equal per capita transfers in the immobile capital case). ii) Under capital mobility, the total federal tax diferences increases, that is we see a larger diference between the lowest and highest taxes increases.
159 4. Technology Beats Capital
Figure 4.12: Normalized uniform federal tax rates that maximize the utility of the respective country given mobile capital.
160 5 Synthesis and Outlook
161 5. Synthesis and Outlook
By studying multilevel climate policies, this thesis centers around the potential of collective action in climate change mitigation. It departs from conventional climate policy approaches that often focus on global solutions or inefciencies stemming from overlapping multilevel policies. Instead it studies benefcial multilevel design options, translating Ostrom’s notion of polycentric climate governance to public economics. Changing perspectives towards efective and feasible multilevel climate policy becomes particularly relevant for the current international climate diplomacy: For a long time, there was hope that the political and scientifc discourse would fnd a global cooperative solution to combat global warming. The Paris Agreement, instead, calls on countries and regions to fnd individual solutions to rapidly slow down global warming and to bind them together by voluntary commitments . The case of the EU provides an example for multilevel governmental action in federal systems, as climate and energy policies exist at several levels of government. Climate policy at the EU level needs to tackle diferent interests, often with the requirement of ensuring states’ consent to make EU policy feasible. This becomes particularly challenging because EU member states difer from each other in several ways, for instance, in wealth levels, size and CO2 emission intensities. Given these diferences, member states prefer diferent climate policy stringency making it difcult to fnd a common federal policy. The research presented above is thus an attempt to analyze systematically how political impetus for emission reductions coming from just a few countries in a federation can be leveraged on to yield efective climate policy for all member states. Furthermore, policymaking in the EU in general seems dominated by distributional confict. A further motivation for the research presented is the question of whether climate policy can be designed efectively without solving the major general distribution conficts between the member states, but instead creating win-win situations of common climate policy for all member states. This thesis shows how multilevel climate policy in federations can create a win-win situation for all member states. There can be multiple Pareto optima in designing climate policy in a federation. But is there also a uniform federal carbon price and transfer policy that make all states better of compared to their decentralized solution? This thesis fnds that such policy exists, but rich and CO 2 emission intensive states may be the frst to veto federal policy stringency, because they become donors of federal carbon price revenues. There exist, however, federal minimum prices that comply with the political impetus of the largest donor country as they maximize its utility, and at the same time make all other member states better of in relation to their decentralized policy outcome. Such federal minimum price can represent a win-win situation for all member states because the largest donor country benefts and voluntarily gives transfers to the other states.
The main body of the thesis reached the following conclusions about the guiding questions posed in the introduction:
162 How do diferences between member states afect the search for common, uniform federal carbon prices? The greater the diferences, the more complicated it is to fnd a uniform federal carbon price to which all states would agree. In general, member states expecting to face a larger burden from the federal price and transfer policy prefer lower federal prices than states with less federal policy burden (Chapter 3). The gross federal carbon price payment is larger for wealthier states or for those with a large emission intensity, or both (Chapters 3 and 4). While the previous observations may be straightforward, some of the following further fndings may come as a surprise for environmental economists: From both an efciency (Chapter 2) and a unanimity-ensuring perspective (Chapter 3), rich states bear a larger burden under federal policymaking than their counterparts. Rich states, therefore, prefer a lower uniform federal carbon price than poor ones, leading them to indicate a feasible federal minimum price (Chapter 3). Also, diferences in emission
intensity place a larger burden on emission-intensive states. If wealth and CO2 emission intensity are inversely correlated, as it is the case in the EU, a federal carbon price also places a larger burden on poor and emission-intensive states (Chapter 4). Whether an agreeable uniform federal carbon price exists also depends on the distribution of federal carbon price revenue, i.e. who receives and who donates revenue.
How does revenue distribution by commonly used transfers help multi- level policies to function efectively? The choice of transfer rules allows the federal policy burden to be (re)adjusted. Federal egalitarian transfers may limit the efectiveness of a uniform federal carbon price when there is high wealth inequality between states. Sovereignty transfers are feasible irrespective of diferences in wealth. Juste retour transfers render federal policy inefective if states anticipate them. If unanticipated by states, juste retour transfers perform similarly to sovereignty transfers (Chapter 3).
Which member states become donors of federal transfers, and which become recipients? If states only difer in wealth and population size, and if federal egalitarian transfers are used, small rich states are major net per capita donors. Under sovereignty transfers, and if states difer in wealth and population size, large rich states are major net per capita donors. The largest recipient is always a poor state (Chapter 3). If states also difer
in CO2 emission intensity, then poor and emission-intensive states become signifcant net donors under egalitarian transfers. Under sovereignty transfers, large and rich states become major net donors. These fndings depend on the ease of substitution possibilities between the labor-capital input and emission input in production, in contrast to the
cost share of emissions in production. Large CO2 intensity of production results in large emission levels only with a relatively high gross complementarity of inputs, and
163 5. Synthesis and Outlook relatively low cost shares of emissions (which is currently the case in the EU), and vice versa(Chapter 4).
What makes a federal carbon price acceptable and what does this imply for the EU? This thesis considers a federal carbon price acceptable, if it maintains or improves the well-being of any citizen in the federation by delivering Pareto improvements to the case without federal policy. Such an acceptable federal policy is considered a feasible federal policy as all states would beneft from federal policy. Some federal carbon prices, however, do deliver Pareto improvements, but are regressive, resulting in citizens in poor states bearing a larger burden than their counterparts in rich states. In such a case, sovereignty transfers can make the federal price progressive. The EU transfers the bulk of revenue from the ETS on a sovereignty basis, and thus the major EU ETS transfer helps to produce a progressive carbon price impact (cf. Chapter 4).
The remainder of this chapter is organized as follows. In Section 5.1, I synthesize my main fndings. Subsequently, in Section 5.2, I discuss the methods used to obtain my results. In Section 5.3, I critically discuss the novelty of the arguments put forward in favor of multilevel policies. Finally, Section 5.4 suggests directions for future research, and Section 5.5 provides an outlook, especially on relevance of the fndings for climate policy in the EU.
5.1 Federal carbon pricing and transfers
One of the main contributions this thesis makes is to show the potential of multilevel environmental policy in multilateral systems: First, I argue that ensuring the consent of states to federal policies can provide an efective leeway for federal policy implementation. Second, I show that wealthy and emission-intensive states hamper federal policy, but can be taken on board by an appropriate federal minimum price. At this minimum price, the interests of the most critical state manifest themselves as the minimum price maximizes its utility. The utilities of all other states get maximized at larger federal prices, such that the minimum price defnes the least common denominator for a uniform federal price. The overall policy instrument analysis provided by this thesis places considerable emphasis on the burden of multilevel policy and transfer design. While Chapter 2 puts forward normative principles and efciency arguments to assess the EU ETS, Chapter 3 scrutinizes the role of wealthy states as the largest contributors to federal revenues. Chapter 3 focuses on wealth diferences (vertical inequality), while Chapter 4 studies distributional aspects on the vertical and horizontal dimension of inequality. The results of Chapter 3 suggests that consent-ensuring federal policy is automatically progressive;
Chapter 4 shows that this is not the case when production facilities difer in CO2
164 5.2 Models for fnding feasible and efective multilateral and federal policies emission intensity (horizontal inequality). Appropriate federal transfers can compensate for the regressive efect, but, as Chapter 4 shows, these appropriate federal transfers look diferent from transfers considered appropriate when single state economies are considered. Taken together, the policy analysis in this thesis contributes to a better understanding of policy options, bottlenecks, and leeway for feasible environmental policy in multilateral systems governed by a multilevel regime. While there are several climate policy design options, this thesis provides a systematic analysis of price-based policies consisting of a uniform emission price at the top level and individual emission prices at the state level. Both generate revenue that is distributed back to the population. With increased complexity, the model and insights developed in Chapters 3 and 4 can also serve as a basis for a broader research agenda.
5.2 Models for fnding feasible and efective multi- lateral and federal policies
A further signifcant contribution of this work is the development of new models in Chapters 3 and 4, all of which are motivated by the need for sound policy advice in the light of current economic and political concerns of countries or states being members of federal systems striving to fnd feasible and efective climate policies. The results of the existing model of (Chichilnisky and Heal, 1994) applied in Chapter 2, conceptually justifes the development of new models for two reasons: First, the analysis of Chapter 2 shows that an efcient solution is not necessarily a feasible solution as transfer payments be be too high and met the dissent of donor states. Second, the application of a social welfare approach implies a normative criterion that makes an explicit judgment about distribution, in order to determine the socially optimal outcome. The models in Chapters 3 and 4 instead merely use the formulation of a Pareto improvement to capture the feasibility of federal policy. The Pareto approach allows the whole range of solutions to be captured without making an ex-ante statement on distribution, potentially circumventing conficts of distribution in federal climate policy design. The models are developed so as to permit deriving analytical solutions. For example, in Chapter 3, I uncover the driving forces behind the main results by solving the developed static general equilibrium model analytically — a numerical analysis with increased complexity supports the robustness of the analytical results and provides additional insights. The model developed in Chapter 3 can be applied to other contexts when it comes to fnding acceptable burden-sharing for a public or common good where the most relevant country heterogeneity is related to the level of wealth. This is because this model is
165 5. Synthesis and Outlook closely related to the theory of (Olson, 1986; Olson, 1965; Olson and Zeckhauser, 1966), formalized by (Bergstrom et al., 1986). Olson based his analysis on NATO and the disproportionately large contributions made by the United States during the arms race. In the same vein, my model can be contemplated in similar defensive directions, but from the viewpoint of an existing multinational system seeking to ensure that the wealthiest state is on board — be it environmental pollution, war, or a pandemic threat. On a conceptual level, the most important contribution of Chapter 3 turns Olson’s theory (Olson, 1986; Olson, 1965; Olson and Zeckhauser, 1966) on its head. While Olson argues that rich states are willing to create multinational systems, I show how multinational (federal) systems can ensure that the richest state is willing to participate in and contribute to federal policymaking. From a formal perspective, this chapter develops a structural approach to assessing the willingness of individual states to participate in federal environmental policymaking. The model in Chapter 4 is tailored to be applied to the specifcs of EU climate policy due to the heterogeneity of the member states. Therefore, it must refect many interactions simultaneously, i.e. wealth and technological heterogeneity, together with multi-level policies and the impact of mobile capital. A numerical solution method is ideal for this purpose. Nevertheless, I determine the core drivers analytically using a reduced model variant. One important aspect I isolate is that high emission intensity in production causes high emission levels only if the gross complementarity of production inputs is strong, and the cost shares of capital and labor are high. Such a result would not occur with a Cobb-Douglas function. This analytical perspective may also help future research to determine which conditions of constant elasticity of substitution functions determine how emission generation relates to capital and labor demand. Since this thesis is primarily concerned with fnding entry points for more stringent climate policy, it focuses on a short-term perspective. Static models are best suited for this issue and are therefore used throughout this thesis. The only exception is the established long-term energy system investment model LIMES-EU, applied in Chapter 2. The application of the LIMES-EU model (Nahmmacher et al., 2014) is motivated by the comparison of investments in the electricity system of member states under diferent policy scenarios. In this case, policy scenarios are determined exogenously, while all other models employed determine optimal policy solutions. In sum, the two models developed specifcally for this thesis can easily be modifed, extended, or recalibrated to answer diferent research questions relating to aspects such as diferent countries, game structures, policy instruments, and sectors. Sections 3.5 and 4.5 give several examples of possible future research questions that could be addressed with modifed versions of the individual models. Further examples and a discussion of how a combination of the individual models could lead to new insights are given in Section 5.4.
166 5.3 Why study multilevel policies
5.3 Why study multilevel policies and what is the beneft of the chosen model types?
This thesis argues for multilevel policy, even though some previous research has argued that a) global public goods are provided most efciently by a central authority and b) overlapping policies can result in excessive cost without providing additional benefts. It is undeniable that a global or central authority regulates global public goods most efciently. However, current political developments show that this path — at least in time — is not viable. Elinor Ostrom (2009) argued that multilevel governance structures could also improve the provision of public goods. The analysis and better understanding of multilevel governance structures can, therefore, point out an important path to more stringent climate policies. The EU the primary example for this line of thinking. The main contribution of this thesis to studying multilevel forms of governance in public economics is the use of the traditional concept of Pareto improvements to model the federal government. It is, to the best of my knowledge, a new approach which has not been considered in the federalism literature before. The approach ofers the advantages to ensure the consent of member states to federal policy: thereby federal policy can be viable and efective. Both the theoretical analysis and the numerical simulations used in this thesis have limitations. Analytical models can decisively point out individual mechanisms and deliver mathematical truth. However, they only deliver clear insights up to a certain degree of complexity. Combining theoretical and numerical modeling, which can represent more complexity, is an ideal approach to better describe both the individual elements and their interaction. Throughout the main chapters, therefore, I have combined theoretical models with numerical models. I calibrated the numerical model in Chapter 4 to EU data, and fnd similar efects of a large computational general equilibrium (CGE) model in Böhringer et al. (2015). The model I used in Chapter 4, however, difers from such a CGE as follows: A large-scale CGE has a very detailed representation of the production side and considers the impact of a particular policy introduction (policy shock). A policy shock could be, for example, the introduction of a carbon price of 30 EUR/tCO2. Then the result without this carbon price is compared with the result with a carbon price. The model I use has a much less detailed representation of the production side. I contrast the result of a purely decentralized policy to multilevel policy. Here, governments choose their policies endogenously, based on an individual assessment of consumer benefts and emission damage. Thus, my approach has a more detailed view of the strategic interactions of governments and endogenous policy choice, but a less detailed view of economic sectors. I consider it fruitful to compare the medium-sized model results of Chapter 4 to CGE model outcomes since diferent methodological approaches have advantages and
167 5. Synthesis and Outlook disadvantages. Testing the hypotheses from this thesis using diferent methods increases their robustness.
5.4 Next steps
Although this thesis has advanced knowledge of multilevel climate policy, many research questions remain unanswered. The topic of the thesis raises many challenges to extend the current analysis. I name a few of the challenges that I consider the most relevant for constructive future research, both for climate policy (Section 5.4.1) and beyond (Section 5.4.2).
5.4.1 Climate policy In the following part, I frst mention simple model extensions or modifcations, and then proceed to discuss more complex conceptual extension.
Simple model extensions or modifcations There are some simple extensions that can contribute to the research agenda pursued in this thesis that require only simple modifcations of the models developed in Chapters 3 and 4. First, a consideration of other federal transfers could deliver further insights into federal policy design. One possibility is to consider strategic federal transfers that strive to reward states’ ambition, e.g. transfers based on the member states’ mitigation efort or policy stringency. Another possibility is to include the historical accountability of states that have generated large prosperity from historical emissions, partially reversing the efects of sovereignty transfers. This analysis would require the inclusion of diferent transfer rules and, in case of strategic federal transfers, change the frst-order conditions of the states in the model code. Second, analyzing the role of Brexit1 or any other states leaving the EU or the EU ETS, respectively, can deliver insights into the implications of sharing the burden of EU policy and EU policy stringency. This analysis can be pursued by dropping states from the existing model code used in Chapter 4. While this thesis concentrated on multilevel policies and the efects of how they interact, a third modifcation would be to drop states’ climate policies from the models in Chapter 3 or 4. This would greatly simplify the models and give insights into feasible policy design if states do not pursue their own policy choices. Depending on whether insights should be obtained for only horizontal inequality, or for both horizontal and vertical inequality, then either the model of Chapter 3 or that of Chapter 4 should be modifed. Even though this would depart from the consideration of multilevel policies, it
1See for a discussion of Brexit, for example, Hepburn and Teytelboym (2017).
168 5.4 Next steps could be a worthwhile simplifcation to study the role of other heterogeneities such as multiple sectors on policy feasibility. Fourth, the model of Chapter 4 could be calibrated to one of the other many regions and federal systems, such as the US, Canada, Switzerland, or Germany, to explore similarities and diferences across these regions.
Conceptual extensions and model comparisons More complex extensions also promise to be highly relevant to enable further policy analysis. However, these would require more efort than the previously discussed points. First, in the long term, production technologies will become more productive and less emission-intensive. A further analysis building on Chapter 4 could explore the dynamics of technological change or diferently sized CO2 intensive and carbon-neutral sectors for each state to analyze the implications for burden-sharing, unanimity, and the feasibility of federal carbon pricing. One specifc avenue could investigate how carbon-neutral investments could be triggered by a common EU policy. This could have specifc relevance if the EU Green Deal is to be implemented. Second, further comparisons with large-scale CGE model results could be useful for identifying where modeling strategic governmental interactions produce similar or diferent results than a multilevel multinational CGE approach with policy shocks. This could help us to identify appropriate models for specifc policy advice by establishing when medium-sized models with strategically acting governments (as in this thesis) or large-scale models are better or equally suited. In cases where both models perform equally well, medium-sized models are favorable, because it is easier to clearly isolate the main drivers. Third, the insights derived from the power system model (LIMES-EU) in Chapter 2 could be compared with the results of a fow-based power pricing model. Flow-based power pricing allows for an economically more efcient and physically more correct representation of incentive structures of electricity production and transmission capacity. Flow-based pricing in a multinational setting can induce strategic incentives for individual states (cf. Roolfs, 2013). Fourth, the results of Chapter 3 could be compared and contrasted to voting models, which also fnd that pronounced diferences between member states can have signifcant efects for federal policy. In addition, they could also lead to more extreme voting at the federal level than at the state level (Daniele et al., 2020). Such efects could potentially be relevant, and hamper or improve feasible federal policy if state representatives have rather extreme positions. I suspect a hampering efect if representatives have a more nationalist perspective, whereas extreme “green” representatives may be able to pursue more stringent EU climate policy. Fifth, the viability or self-enforcement of federations (cf. Oates, 2005) could be further challenged by applying the set-up of Chapter 3 to an international environmental
169 5. Synthesis and Outlook agreement context such that member states decide whether or not they want to join the federation. I am exploring this further in a working paper (Hagen and Roolfs, 2019). Still further, the models used in Chapters 3 and 4 suppose a game theoretic structure, with the federal government acting as the Stackelberg leader coordinating the member states. One could contrast this to the structure of governmental decisions such that all governments are Nash players. One alternative could be to consider the wealthiest state as the most powerful governmental actor of the federal economy, such that it would act as Stackelberg leader.
5.4.2 Beyond climate policy Several of the insights provided by this thesis contribute to the understanding of the provision of public goods in the context of multilevel policies. In the process, they can be transferred to other contexts involving the search for acceptable burden-sharing for a public or common good where the most relevant country heterogeneity is related to the level of wealth. Here, I sketch three ideas for future research that are not necessarily related to climate change, namely migration, polarization and identity, and pandemics. In the coming decades, populations may become more mobile across EU member states. Also, Europe will face more immigration, e.g. for economic reasons, or due to war and confict, or climate change (e.g. Missirian and Schlenker, 2017). As a result, Europe may need to tackle internal and external migration. Thus, the model developed in Chapter 3 could be extended by considering mobile populations across states or external immigration into the federation. Population mobility could advance our understanding of voluntary burden-sharing among member states when wealth and population size patterns change due to migration. The EU is already, and may become increasingly involved in political and social polarization. The increased appearance of nationally oriented parties within member states can thus shape or hamper common European policymaking. Such analysis would require a diferent modeling approach as, for instance, a voting model capturing multilevel policies and polarization (e.g. Daniele et al., 2020). To the contrary, signifcant eforts are undertaken to expand national identity into a European identity (Ciaglia et al., 2018; EC, 2012). A higher level of EU identifcation may lead to social tipping, rendering feasible a greater willingness for cooperation and transfers and more stringent federal public good policies (Nyborg et al., 2016; Otto et al., 2020). One example of EU regulation that has increased identifcation as an EU citizen (cf. "EU-feeling" Ciaglia et al., 2018) is the abolition of roaming tarifs for mobile communications between EU countries in 2017. In this vein, future research could investigate the scope of policies that support the EU identity and the efects it has on multilevel policymaking in the spirit of polycentric governance (cf. Ostrom, 2010).
170 5.5 Outlook
Another potential use of the model framework from Chapter 3 is to study responses to pandemics and the issuance of multilevel government bonds. This is particularly relevant for the EU, where both member states and the EU as a federal entity are, at the time of writing this thesis, considering to issue bonds for managing an inclusive recovery from the recession caused by COVID 19. Member state governments would seek to maximize welfare at the local level, which depends on consumption as well as on local and EU damage caused by the pandemic. Each state could consider issuing bonds to mitigate damages caused by the pandemic. The federal level could then try to improve the decentralized bond solution by proposing the issuance of federal bonds under a particular distribution scheme for collected funds. The uniform price at the federal level in the current model setting would translate to the amount of federal debt issued, while state policy would translate to the amount of debt issued by the states. In such a setting, the most relevant heterogeneity of the member states would be their ability to borrow, measured in terms of their public debt to gross domestic product ratio, so that states with higher levels of indebtedness would have to bear higher mitigation costs (debt service). Net transfers between states could translate to bailouts. Based on Chapter 3, it is very likely that the richest and least indebted member states would pay to bail out highly indebted member states. An analysis of this issue could help identify acceptable burden-sharing approaches for bail out and deliver insights into the feasibility of federal bonds and underlying incentive structures. The models by Kalamov and Staal (2016) and Baglioni and Cherubini (2012) could provide starting points. Taken together, the issues identifed in Section 5.4 could form a broader research agenda for federal systems. This agenda could, therefore, be devoted to the analysis of policy instruments to address voluntarily fghting common threats such as climate change, environmental pollution and pandemics, taking into account polarization and identity patterns, possibly together with demographic change.
5.5 Outlook
National economies are facing the major challenge of implementing policies that will achieve signifcant decarbonization to limit global warming to below critical levels in this decade. Decarbonization requires considerable eforts on the part of society, engineering, businesses, and politics. This thesis proposes a new perspective to conventional approaches to appropriate policy design from the perspective of multilevel governance: For a long time, there was hope that political and scientifc discourse would fnd a global solution to combat global warming. Instead, the Paris Agreement calls on countries and regions to fnd individual solutions to slow down global warming rapidly. Exploiting the potential of collective action, for example, through efective climate policy at several levels, with an awareness of distributional efects that improve political feasibility, can now make a substantial
171 5. Synthesis and Outlook contribution. The study of the EU and its various member states provides an ideal laboratory for research on climate and energy policy at several levels of government, not least because EU policy must often be decided in consensus of member state to be feasible. This thesis stresses the decisive role of wealthy and emission-intensive states in the feasibility of federal policy. Applied to the EU, Germany and Poland would be the most critical states towards more stringent EU climate policy, since they would have to bear a large burden. This view receives support both from the academic literature, as well as from the actual EU policy arena, where both countries often dominate or hamper EU climate policy progress. One solution proposed by this thesis is to implement a minimum federal carbon price together with appropriate transfers. At this minimum price the interests of the largest net donor manifest themselves, and are thereby included in federal policymaking. A major challenge facing the EU in the future may be the adoption of the EU Green Deal. The framework of the Green Deal sets out a large-scale transition plan towards a sustainable EU economy:
“Climate change and environmental degradation are an existential threat to Europe and the world. To overcome these challenges, Europe needs a new growth strategy that transforms the Union into a modern, resource-efcient and competitive economy [...]” (EC, 2019)
Among other things, the Green Deal aims at EU carbon neutrality in 2050, which comprises increasing the EU’s nationally determined contributions to the Paris Agreement. Such transformation requires enormous eforts on many fronts, e.g. by expanding the EU ETS coverage, and mobilizing huge funds to help adapt entities facing large and maybe unfair burdens due to the transition. This thesis may contribute to laying a foundation for the transition towards a sustainable EU economy by precisely breaching the issues of burden-sharing, donor interests and the consent ability of policy reforms. If the EU and its member states succeed in achieving credible steps for decarbonization, they can also serve as a role model for other regions and federal systems.
172 REFERENCES
References
Baglioni, A., & Cherubini, U. (2012). A Theory of Eurobonds. SSRN working paper. http://ssrn.com/abstract=1928573. Bergstrom, T., Blume, L., & Varian, H. (1986). On the private provision of public goods. Journal of Public Economics, 29, 25–49. Böhringer, C., Rivers, N., Rutherford, T., & Wigle, R. (2015). Sharing the burden for climate change mitigation in the Canadian federation. Canadian Journal of Economics, 48 (4), 1350–1380. Chichilnisky, G., & Heal, G. (1994). Who should abate carbon emissions? Economics Letters, 44, 443–449. Ciaglia, S., Fuest, C., & Heinemann, F. What a feeling?! How to promote ‘European Identity’: EconPol POLICY REPORT. 2018. Daniele, G., Piolatto, A., & Sas, W. Does the Winner Take It All? Redistributive Policies and Political Extremism: Barcelona GSE Working Paper: 1157. https: //www.barcelonagse.eu/research/working-papers/does-winner-take-it-all- redistributive-policies-and-political-extremism. 2020. EC. The development of European identity/identities: unfnished business: a policy review. 2012. EC. (2019). The European Green Deal. [Accessed 6-May-2020]. Hagen, A., & Roolfs, C. (2019). Self-Enforcing Environmental Federations. accepted at SURED conference 2020. Hepburn, C., & Teytelboym, A. (2017). Climate change policy after Brexit. Oxford Review of Economic Policy, 33 (suppl_1), S144–S154. Kalamov, Z., & Staal, K. (2016). Public debt, bailouts, and common bonds. International Tax and Public Finance, 23 (4), 670–692. Missirian, A., & Schlenker, W. (2017). Asylum applications respond to temperature fuctuations. Science, 358 (6370), 1610–1614. %5Curl % 7Bhttps : / / science . sciencemag.org/content/358/6370/1610%7D Nahmmacher, P., Schmid, E., & Knopf, B. Documentation of LIMES-EU - A long-term electricity system model for Europe. https://www.pik-potsdam.de/members/ paulnah/limes-eu-documentation-2014.pdf. 2014. Nyborg, K., Anderies, J. M., Dannenberg, A., Lindahl, T., Schill, C., Schlüter, M., Adger, W. N., Arrow, K. J., Barrett, S., Carpenter, S., Chapin, F. S., Crépin, A.-S., Daily, G., Ehrlich, P., Folke, C., Jager, W., Kautsky, N., Levin, S. A., Madsen, O. J., . . . de Zeeuw, A. (2016). Social norms as solutions. Science (New York, N.Y.), 354 (6308), 42–43. Oates, W. (2005). Toward A Second-Generation Theory of Fiscal Federalism. Interna- tional Tax and Public Finance, 12 (4), 349–373.
173 5. Synthesis and Outlook
Olson, M. (1986). A Theory of the Incentives Facing Political Organizations. Neo- Corporatism and the Hegemonic State. International Political Science Review, 7 (2), 165–189. https://doi.org/https://doi.org/10.1177/019251218600700205 Olson, M. (1965). The Logic of Collective Action: Public Goods and the Theory of Groups (Vol. 20 printing, 2002). Harvard University Press. Olson, M., & Zeckhauser, R. (1966). An Economic Theory of Alliances. The Review of Economics and Statistics, 266–279. Ostrom, E. A Polycentric Approach for Coping with Climate Change: Policy Research Working Paper 5095. 2009. Ostrom, E. (2010). Beyond markets and states: polycentric governance of complex economic systems. American Economic Review, 100 (3), 641–672. Otto, I. M., Donges, J. F., Cremades, R., Bhowmik, A., Hewitt, R. J., Lucht, W., Rockström, J., Allerberger, F., McCafrey, M., Doe, S. S. P., Lenferna, A., Morán, N., van Vuuren, D. P., & Schellnhuber, H. J. (2020). Social tipping dynamics for stabilizing earth’s climate by 2050. Proceedings of the National Academy of Sciences, 117 (5), 2354–2365. Roolfs, C. (2013). Modeling electric network capacity under diferent institutional settings: Master’s thesis. University of Oldenburg, EPF Lausanne.
174 Tools and Resources
This thesis was written and compiled with LyX versions 2.3.2–1 and 2.3.4.3 and TeXnicCenter version 2.0.2. The PDF from Chapter 2 is included via the pdfpages Package (Matthias, 2020)(Thanks to Andreas Matthias for providing me with an extra feature for this package). All literature databases were managed with Citavi version 5.7.1.0. Moreover, additional resources have been used for simulations, data analysis and generating graphical output, as indicated.
Chapter 1: GNU R version 3.6.3 (2020-02-29) "holding the windsock" together with the interface of RStudio version 1.1.463 was used for data-processing and designing the fgures.
Chapter 2: The underlying data of Figures 2.3 and 2.4 was processed in Microsoft Excel 2010. The numerical simulations were performed with LIMES-EU implemented in GAMS as documented in Nahmmacher et al. (2014). Microsoft Excel 2010 was used for post-processing of output and designing the fgures. Schematic graphics were created with Inkscape 0.92.
Chapter 3: The numerical optimization problem was implemented in GAMS using the CONOPT4 solver version 4.02 (Drud, 2019) solving an NLP problem. The GAMS script was controlled with GNU R version 3.6.3 (2020-02-29) "holding the windsock" with the interface of RStudio Team (2016) version 1.1.463. GNU R version and RStudio were also were used for post-processing of output and designing the fgures. Schematic graphics were created with Inkscape 0.92. Numerical experiments (not used in the fnal manuscript) were conducted with Wolfram Mathematica version 10.0.2.0. Analytical results were backed by Wolfram Mathematica.
Chapter 4: Analytical results were backed by Wol fram Mathematica version 10.0.2.0. The numerical optimization problem was implemented in GAMS using the CONOPT4 solver version 4.02 (Drud, 2019) solving an NLP problem. The GAMS script was controlled with GNU R version 3.6.3 (2020-02-29) "holding the windsock" with the interface of RStudio Team (2016) version 1.1.463. GNU R and RStudio were also used for post-processing of output and designing the fgures. Schematic graphics were created with Inkscape 0.92.
175
Bibliography
Drud, A. (2019). Conopt4, Release 4.02, ARKI Consulting and Development A/S. GAMS. (2019). GAMS Development Corporation. General Algebraic Modeling System (GAMS) Release 27.1.0. Matthias, A. The pdfpages Package. 2020. Nahmmacher, P., Schmid, E., & Knopf, B. Documentation of LIMES-EU - A long-term electricity system model for Europe. https://www.pik-potsdam.de/members/ paulnah/limes-eu-documentation-2014.pdf. 2014. RStudio Team. RStudio: Integrated Development Environment for R. http://www. rstudio.com. Boston, MA, 2016.
177
Acknowledgements
Until about ten years ago, I would not have expected that my interest in the nature, the Energiewende and justice would lead me to economics.Several events paved the way. Among them was a public lecture by Ottmar Edenhofer, in his role as co-chair of Working Group III of the "Intergovernmental Panel on Climate Change”, in Oldenburg in 2009. He described the threat of climate change using economic methods, with a decidedly appreciative attitude toward all participants that asked questions that day, regardless of their background. First, I would like to thank Ottmar Edenhofer, who not only inspired me but later hired me and became my doctoral advisor, giving me the beneft of his support, joint discussions and idea development for this PhD project. Next, I thank my second and third referee, Marco Runkel and Matthias Kalkuhl, for reviewing the thesis and for their additional support. Also, I thank Stefan Heiland for serving as chairman of the board of referees. For providing support, challenge, advice and motivation, be it as temporary co-advisors or ofering senior support to this PhD project, I thank Kai Lessmann, Beatriz Gaitan, Michael Pahle, Linus Mattauch, and Brigitte Knopf. I thank my co-authors for collaboration, sharing of expertise, and constructive feedback. I also thank the members and afliates of the Future Lab Public Economics and Climate Finance, the Climate and Energy Policy Group, Economic Growth and Human Development Group, as well as the de.zentral and KLIF project teams for the collegial atmosphere, discussions and inspirational breaks. For organizational and technical support, I thank the teams of the RD3 Research Coordination, IT, and Model Operations Group. My colleagues, former colleagues, and friends at the Potsdam Institute for Climate Impact Research (PIK), the Mercator Research Institute on Global Commons and Climate Change (MCC), and other institutes have supported me intellectually, technically, and not least emotionally, and for this I thank them very much. In particular, I would like to thank, in alphabetical order, Lavinia Baumstark, Max Franks, Achim Hagen, Martin Hänsel, Jerome Hilaire, David Klein, David Klenert, Katrin Kohnert, Sebastian Kraus, Andrew McConnell, Niklas Roming, Eva Schmid, Hendrik Schuldt, Gregor Schwerhof, Jessica Strefer, Susanne Stundner, Ibrahim Tahri, Johanna Wehkamp, Leonie Wenz, and Boyan Yanovski, who often supported me with words and deeds sometimes also BIBLIOGRAPHY during my weekend- and night shifts, and/or made me realize the importance of breaks and team-spirit.2 My friends3, family Köhler, and, last but not least, my family helped me keep up creativity and strength on this dissertation path. I am grateful for their unconditional support and their patience and understanding with me being distracted and missing many events. Marianne, Hermann, Heye, Rita, and Lina supported me for more than three decades and imparted values that encouraged me to pursue my path(s) – it is to them that I dedicate this thesis.
The fnal parts of this thesis were written during the COVID-19 pandemic’s frst and second wave and the related lock-downs. Vaccination started becoming available at the end of 2020 and can be considered "a test of whether the world can work together to confront common threats, such as organised crime and climate change", the Economist concluded.4 During this pandemic many of us experienced frsthand the importance of efective and fair stewardship of the global commons, and the role of multinational cooperation along with the choices made by each individual. When you read these lines sometime in the future after the summer of 20215, I hope that the COVID-19 pandemic will be under control and the necessary climate policies will have been implemented or be on their way to avert a climate catastrophe. I hope "the darker the night, the brighter the stars" will have come true6, so that the darkness of this crisis has led us to a brighter future.
Potsdam, June 2020 and January 2021
2The lunchtime runners, the resulting infamous half marathon team, and the no less infamous traditional kayak tour organizing team may also recognize themselves here. 3Duplications to the previous paragraph are possible. 4The Economist, https://www.economist.com/leaders/2021/01/09/who-should-get-the-jab, pub- lished and accessed January 9th, 2021. 5It is expected and/or hoped that the pandemic could be under control and herd immunity established in the summer of 2021. 6“The darker the night, the brighter the stars” stems from the Russian Poet Apollon Maykov available in Russian at https://45parallel.net/apollon_maykov/stihi/#iz_ispanskoy_antologii, page 1878. The analogy of Maykov’s words to the pandemic was, to the best of my knowl- edge, frst made by Kristalina Georgieva at Opening Press Conference of the Annual Meet- ings of the IMF and the World Bank 2020, https://www.imf.org/en/News/Articles/2020/10/14/ tr101420-transcript-of-imf-md-kristalina-georgieva-opening-press-conference-2020-annual-meetings, ac- cessed January 9th, 2021.
180